Rethinking wood future dimensions of timber assembly 9783035617061, 3035617066

Citation preview

Rethinking Wood

Markus Hudert, Sven Pfeiffer (Eds.)

Rethinking Wood Future Dimensions of Timber Assembly

Birkhäuser Basel

Table of Contents Foreword | Klaus Zwerger p. 6 Preface | Markus Hudert and Sven Pfeiffer p. 12 Concepts and Perspectives p. 16 Think Like the Forest: Maximizing the Environmental Impact and Energetics of Building Timber | Kiel Moe p. 20 Cascading Wood, Material Cycles, and Sustainability | Mark Hughes p. 30 Wood on the Rise: A Speculative Approach to Timber Construction and Joinery in Southeast Asia | Michael Budig p. 46 Joinery Culture p. 56 Designing Through Experimentation: Timber Joints at the Aalto University Wood Program | Pekka Heikkinen and Philip Tidwell p. 60 Reciprocal Timber Structures and Joints | Olga Popovic Larsen p. 88 Press-Fit Timber Building Systems: Developing a Construction System for Flexible Housing Solutions | Hans Drexler p. 100 Glued Connections in Timber Structures | Gerhard Fink and Robert Jockwer p. 116 Digital Processes p. 128 Freeform Timber Structures: Digital Design and Fabrication | Toni Österlund and Markus Wikar p. 132 Bringing Robotic Fabrication into Practice | Léon Spikker p. 150 Concepts for Timber Joints in Robotic Building Processes | Philipp Eversmann p. 164 Joyn Machine: Towards On-Site Digital Fabrication in Bespoke Woodwork | Simon Deeg and Andreas Picker p. 178 New Materials and Applications p. 190 From Pulp to Form: Future Applications of Cellulose | Heidi Turunen and Hannes Orelma p. 196 Wood Foam: A New Wood-Based Material | Frauke Bunzel p. 206 TETHOK: Textile Tectonics for Wood Construction | Steffi Silbermann, Stefan Böhm, Philipp Eversmann, and Heike Klussmann p. 216 Reapproaching Nature p. 232 Designing with Tree Form | Martin Self p. 236 The (D)Efficiencies of Wood | Marcin Wójcik p. 250 Baubotanik: Living Wood and Organic Joints | Ferdinand Ludwig, Wilf Middleton, and Ute Vees p. 262 References p. 276 Biographies p. 286

5

Rethinking Wood

Foreword Klaus Zwerger

“Rethinking Wood” evokes two quite different reflections. The first involves a return to wood as a material characterized by its inhomogeneity. Its optimal use demands respect for the uniqueness of every single piece. A respect that can only grow out of extensive background knowledge and experience with this material—a never-ending process as the best artisans and artists discover throughout their working lives. With every new piece of wood, they face the intrinsic challenge of rethinking how to best bring its unique qualities to light. This is a highly idealized representation. Of all artists, sculptors have perhaps been the most uncompromising in their pursuit to shape their works in alignment with the unique form of natural materials. An endeavor that both tested their attention and demonstrated the breadth of their talent through their flexible response to the raw material at hand. Artisans, such as carpenters and cabinetmakers, on the other hand, were denied such considerations as they were bound by economic concerns. And yet, the responsibility of selecting the most suitable piece of wood for a specific purpose, and particularly when cutting joints to ensure that the reduced cross-section is not carelessly weakened, lay entirely in their hands. Historical objects offer examples of both impressive attentiveness and gross negligence. Sculptors’ concerns and means of expression have profoundly changed. Artists are acutely sensitive to social change. They would be the first to know that only a few people remain who stand to gain something from the material reflection described here that has become foreign to them. But developments are not consistently sequential. Not every path taken leads to completely new fields. Fortunately, there are artisans who still feel connected to, or have newly discovered this idea. The second reflection follows a completely different perception of wood. For the inexperienced, it is highly difficult to estimate the properties of this material, and this in turn has set in motion an effort to control it. In the process of transforming wood to wood plastic composites, many of its positive properties remain intact. The partial elimination of the properties considered negative has reinforced the impression that this was a positive development. Over centuries, architects, engineers, and non-professionals have laid a great deal of groundwork in this area, both theoretical and practical. Roof constructions had to span ever larger distances or surface areas. The practice of simply enlarging the beam cross-sections soon reached its limits in more ways than one. Growing massiveness increased the weight to such a degree that such beams could no longer 6

bear the additional load of the roof constructions that they were meant to carry. On the contrary, there reached a point when they, too required support. A rise in construction led to a noticeably significant reduction in timber resources. Competition between those who could afford the few existing exceptional pieces was predictably limited. If the required dimensions were no longer available, craftspeople had to come up with alternatives. Large timber traders and wood factories have not been in existence for very long. The rise in prices triggered by ever diminishing supplies spurred developments, such as truss frames and strut frames. Other innovative approaches, such as the hammer beam construction, proved unviable. Churches and temples could benefit from gaining an undisturbed view of the alter from all directions. Heavily loaded wagons could be easier to move around in barns if there was no danger of striking any supporting pillars. Bridges had to enable the transport of loads from one bank to another without support. Both interlocking beams and the Chinese “woven” arch bridges extended the reservoir of natural materials. Liu Yan (Liu 2018) has established that similar development approaches had existed in Europe. Log buildings play with the idea of turning a linear material into a two-dimensional one. The observation that two similarly aged trees merge together when they stand in each other’s way was first used by landscape designers as a source of inspiration. Meanwhile, companies are now growing furniture to order. Efforts haven’t stopped there. Scientists are trying to grow branches of different trees into a network. This is a case of re-thinking wood with a view to expanding its scope of application. The desire to overcome material-specific limitations mentioned in the first reflection served as a source of inspiration for developing products that cannot be manufactured with naturally grown wood, at least not in serial production. If we were to place the two reflections in temporal order of appearance, the second reflection initially acquires solely positive attributes: the material no longer dictates the form of the product; now it’s the desired product that determines the development of how the material should be reshaped. The title the editors have chosen is strongly focused on this second consideration. In their collection of contributions, however, there is room for reflection on my first line of thought. If we allow for both reflections, then the title “Rethinking Wood: Future Dimensions of Timber Assembly” encompasses the never-ending story by human standards of one of mankind’s most important building materials. Trees offer a variety of products, wood being the primary usable resource. No one uses bark or bast nowadays. It has become increasingly difficult to find places where leaves are used to cover roofs. However, the moment we say that wood is the primary usable part of a tree, it becomes immediately evident how one-sided and limited this view is. Trees produce oxygen and bind CO2. Is this also not useable? 7

Rethinking Wood

Trees reduce the risk of erosion. Trees are the habitat for many other animals, a significantly valuable function given the world’s looming nutrition problems. Thus, this too is useful. Humans have become increasingly dependent on wood. Numerous analyzes on the manifold and ever intensifying utilization of wood over the centuries have shown this. For a short period, we trusted seemingly more efficient building materials to such a degree that wood was reduced to mere shuttering material in construction. Only in building interiors did it remain a valued material—nothing could replace the warmth and comfort it exudes as it burns in an open or acoustically perceptible fireplace. But wood has experienced an amazing renaissance, making major strides to regain a remarkable status. This would not have been possible had we continued to use wood in the same way for millennia. Other building materials may have had a strong influence on this. But stone, clay, and iron could no longer sufficiently satisfy user needs. The processes of breaking down the material, re-composing it, and adding substances to it have increasingly homogenized an inhomogeneous material. In addition to making the material easier to process by less experienced craftspeople, it was now possible to do so using machines that were no longer restricted by the limitations of human power. Machines work tirelessly, with greater precision and speed. That these developments “affect our built environment and the way we produce architectural space”1 is beyond question. The historic building materials wood, clay, and stone have created a formal canon that has undergone few fundamental changes throughout history. Even though our growing experience in dealing with the material combined with improved tools have enabled us to resolve our current needs more effectively, this has followed previously known patterns. One of the principle thrusts in wood construction was to continually reduce the volume of the material required. Another was to replicate construction designs originally developed in stone and clay building. Wood scarcity drove Philibert de l'Orme and David Gilly to their innovation. Friedrich Zollinger worked in response to the housing shortage after the First World War. Frei Otto was a student fascinated by phenomena that can be observed in nature. He had recognized that there is no better teacher. They all paved the way for curved roof surfaces made of rod-shaped pieces of wood. It was, however, only the development of wood plastic composites that finally made it possible to produce curved beams of any dimension. Cabinetmakers of the Baroque era had demonstrated how to shape thinly cut wood. Flat plywood panels can be easily transformed into curved surfaces. With wood plastic composites, the two-dimensional directionality of wood

1.

8

Quote from the preface of this book’s editors.

is seemingly no longer a restriction. And yet despite this, the task remains the same. Be it flat elements or linear ones, these must not exceed certain dimensions, otherwise they can no longer be transported and manipulated. The individual elements must still be connected to the building. Transporting building materials has always been a logistical challenge and quite often a hurdle. And yet built structures have been mobile for at least a thousand years. Insofar as it was feasible, wooden buildings were constructed and produced on site. The connecting components were laid flat on the ground, trimmed to size, and respectively labeled. At dizzying heights, they were then assembled to roof trusses over churches and palaces rather quickly and requiring little adjustment. It was absolutely common practice to relocate buildings. Many were dismantled for this purpose, a feat only made possible by using detachable wood joints (Zimmermann 2007). Just as this development affects our environment and the production of architectural space, it is clear that the development of new wood construction materials has an “impact [...] on our culture.”2 The way we tackle and solve our current tasks and problems shapes our culture. This is a gradual process. Artificial materials hardly ever emerge from a flash of genius; rather their development is a result of the cooperation and competition of many minds from both economics and science. This is reflected in the wide range and diverse content of the contributions in this book. The human spirit can hardly be reined. In that sense, development can neither be stopped nor questioned. Reflective considerations are always permissible; not only because they are fundamentally interesting, but because they make us aware of the temporal constraints of cultural currents. Hermann Phleps (Phleps 1959) wrote that “(ornamental multiplications of constructive bracing in the framework) spelt the decline of timber construction.” A verdict he may perhaps wish to reconsider in light of screen-printed, ornamental patterns on the many high-rise smooth glass facades or wood-clad concrete walls. In using wood as a raw material for wood-based materials, there is no reason to differentiate between trees that have grown in the rainforest and those on plantations. The selection criteria are cost-efficiency and reproducibility according to demand. This example alone shows the impact both on our culture and our understanding of it. Some structural engineers are aware of the “gap between reality and practice.” Civil engineer, Bruno Ludescher knows that the growth conditions of trees have a significant impact on the quality of their wood. But, as Bruno Ludescher puts it (Ludescher 2015), “how should we correctly comply with this empirical data?” We cannot reliably judge

2.

Quote from the preface of this book’s editors.

9

Rethinking Wood

this development, as captivated as we are by its course. Perhaps this is a task for later generations. I am fascinated by the wealth of ideas expressed in the examples presented here; from the consistency and outcome of intellectual achievement and perseverance to their implementation. We can look forward to seeing which of these ideas will prevail and which proposals will spark transformation or become the starting point of something new. Research, development and testing new ideas are prerequisites for improvement. That many approaches remain unsuccessful should not discourage anyone from bravely daring to taking the leap. But allow me to make a few critical observations that go beyond this material’s current usefulness. We cannot expect everyone to see trees as more than a source of wood. Many don’t even make that association. One consequence of the devastation that Europe suffered in the twentieth century was the questioning of traditional taboos, which brought about enormous improvements. But where there is light, there is shadow. Many a liberation has been ruthlessly carried out on the backs of others. In the euphoria of seemingly limitless growth, we have severely overexploited our natural resources. Many a liberation has left a vacuum. In our pursuit of self-actualization, we have failed to respect our neighbors, leading us to hit limits we can’t quite handle. Disoriented, we seek stability in mindfulness and the like. But where is the mindful understanding of the source of our existence, an awareness of the fact that our survival depends on a nature that makes our lives possible? Repeatedly misusing the term sustainability will not ensure our continued existence. It speaks for the editors that they have also dedicated two contributions on this topic. To err is human. New developments do not arise without mistakes. Yet, regrettably, mistakes are often only detected in the application. But should they not be corrected as soon as they are discovered; and not only when they are acknowledged as such? Should we not examine the consequences of our actions much sooner, in order not to commit as many errors in the first place? The production of wood plastic composites could promote the short-term profitability of monocultures, against better judgement. By stating that the forest grows faster than trees are taken from it in volume may lead to an apparent positive CO2 balance. However, in the longer term, such statements could become meaningless statistical number games. Should we not be critical when all the CO2 bound in the forests is expelled again when we clear them, transport the wood to gigantic factories, refine it, and transport it again over even greater distances to the most remote construction sites? Thonet’s example shows that scarcity sometimes stimulates more innovative behavior than abundance. The company built its factories in the midst of purchased forests to minimize transport routes in at least one direction. We are not proposing this as 10

a conceptual blueprint for today’s major producers of wood-based materials. What we need today is completely different ideas. Innovations cannot be copied. Luckily, there are architects who are not afraid to take issue with and act on the use of non-regional building materials and resources, even if it comes at higher cost for their clients. The supermarket mentality of being able to afford everything because it is cheaply available, rather than sacrificing many other needs for the sake of quality, is understandably a reflection for a minority, reserved for those who are concerned and—it must be said—can afford it. Building ecologically and sustainably does not only depend on the choice of building materials. Is not thinking globally and acting locally also a worthwhile principle in the building industry? Many architects are already investing a great deal of brainpower and design time in developing construction methods that facilitate a clean separation of the various construction materials used after the expiry of the building. Students write diploma theses on this subject. The awareness that wooden components glued with plastics do not simply return residue-free into the natural cycle when they are left to decay has led to a reflection on alternative, residue-free wood composites. This begs the question: Are product developers allowed to be content with the fact that distributors of their new products are using high-budget marketing to hide and lie about product deficits? Are ethical questions not appropriate or even necessary here? Despite our fascination for new developments, should we not investigate the consequences of our actions as intensively as we pursue goals to optimize nature according to our ideas? Business-focused entrepreneurs make no room for such ifs and buts. So, the first responsibility remains with the developers. It must also be taken into consideration in the development process. Because let’s face it: no one – and especially not those working with wood – wants to be accused of being shortsighted.

11

Rethinking Wood

12

Preface The popularity of wood as a building material has been thriving over the past several years. Digital design tools in combination with numerically driven fabrication processes have made many of the outstanding examples of contemporary wood architecture we see today possible. However, developments in material engineering have made an equally important contribution. Whereas traditional timber construction exclusively used components made of solid wood, timber construction, as it is referred to today, relies mainly on wood-derived products. These products are basically reorganized and reassembled forms of previously comminuted natural wood. The application of different strategies of reorganization on the material level, using for example veneer strips, chips, or particles as basic raw material, has led to a large variety of timber-derived products with new mechanical, physical and geometrical properties. These in turn have enabled structures with bigger spans and unlocked new possibilities and challenges regarding their assembly. New materials, such as wood foam, and new processes and applications for material derivates, such as cellulose and nano-cellulose, continue to be developed and explored, opening up new potentials regarding digital fabrication. Consequently, this has led to a number of questions and challenges: What will wood assemblies on different scales look like in the future? How will they affect our built environment and the way we produce architectural space? And how big will the impact be on our culture, and our technological knowledge and progress? This publication seeks to explore and answer these questions by bringing together a collection of recently built projects, seminal research projects, and critical theoretical perspectives, which, among other things, invite the reader to rethink the current rapport between hylomorphic and morphogenetic design approaches. It suggests that digital technology and material driven practices, connected in an intricate web of mutually dependent relationships, can drastically change our ways of thinking about architectural design as a form of cultural expression. The book addresses the above-mentioned topics and questions in five parts: Concepts and Perspectives, Joinery Culture, Digital Processes, New Materials and Applications and Reapproaching Nature. Each part begins with an introduction that briefly summarizes the individual contributions as well as puts them into context with recent and future developments in the respective fields. Whenever applicable, these introductions also highlight crosslinks between contributions and topics of different parts. The first part, Concepts and Perspectives, brings together thought-provoking essays and concepts that address the notion of 13

Rethinking Wood

sustainability and challenge our understanding of the building material wood. The reader is encouraged to see beyond the isolated individual wooden component, and to understand it in the context of the ecosystem from which it emerges, which is the forest. Moreover, this part presents strategies, such as wood cascading and Design for Disassembly, that could lead to a more material-efficient use of wood and help maintain a balanced and sustainable forestry. It offers perspectives and speculations on the future of forestry and wood construction both in Europe and Southeast Asia. Technologically processed wood-based materials have expanded the horizon of possibilities for architects and engineers. The downside of these artificial composite materials is that they are difficult to recycle. Whereas the need for skilled carpenters and manual joint carving is diminishing, there is an increasing demand for experts who could help transfer traditional timber construction knowledge and joint configurations into modern practice. Hence, the projects presented in the second part, Joinery Culture, address the adaptation of traditional joinery knowledge to modern fabrication processes and vice versa. Part three discusses the integration of wood properties into computational design procedures as well as the resulting economic and social implications of Digital Processes. Pioneering research at ETH Zurich, Stuttgart University, and EPFL, as well as by other stakeholders like Design to Production, has paved the way for the next generation of researchers and practitioners. The projects presented here illustrate the current shift from process-specific CNC machines towards more flexible and versatile tools that push the boundaries of what is buildable. Harnessing the power of computation, they not only involve sophisticated design and fabrication processes, but have also started to include the modelling and analysis of the material behavior. The fourth part introduces new wood-based materials as well as new applications. Unlike most wood products available today, the examples featured here do not, at least in principle, rely on adhesives. Based on the general approach behind engineered wood and wood-derived materials as well as some of the featured examples, the introduction to this part also provides an outlook on new concepts of materiality. Moreover, it predicts a future in which architects and designers will have a bigger say in the design of wood-based and bio-materials. It remains to be seen whether this will impact the degree to which materials are designed towards a specific application and the degree to which they are used as design drivers themselves. Finally, but no less important, part five, entitled Reapproaching Nature, presents approaches that explore the potential of using naturally grown and living wood as well as new ways of exploiting the properties and behavior of natural wood as design drivers. Here, the growth parameters of living organisms themselves become integral design drivers for buildings with a novel and unique architectural expression. 14

Dealing with custom-grown building components and the appropriation of natural form and material morphology, the projects and concepts featured in this final part of the book belong to the forerunners of an increasingly dynamic movement. The different contributions of this book address, on various scales, a wide range of aspects regarding the building material wood. Despite their heterogeneity, they all move beyond established ways of understanding and using wood. Enabled and driven by new material developments, fabrication processes, and a renewed interest in the potential of natural wood, these approaches are pushing the boundaries and opening up new dimensions of timber assembly. Each of the contributions highlights different constituents of this new agenda, mapping out future developments for an old and familiar material, while maintaining links to cultural and historical aspects. Despite its familiarity, it seems there remains much to be discovered about this material and its different forms. Currently, many of the featured approaches remain of interest only to academic and specialist circles. But not for much longer. With this publication we aim to make them accessible to a wider audience, including students and professionals in the fields of design, architecture and engineering, as well as to the interested public. The editing and publishing of this book were made possible thanks to the generous support of Aalto University’s Department of Architecture, the Department of Civil Engineering, the Chair of Design of Structures, as well as the Wüstenrot Stiftung. We are indebted to all guest authors and their valuable contributions to this publication. Moreover, we would like to express our gratitude for the advice and consulting from our colleagues and friends. Markus Hudert and Sven Pfeiffer, 2019

15

Rethinking Wood

16

Concepts and Perspectives The steady rise in wood’s popularity as a building material over the past decade puts it well past being a ‘new’ phenomenon. Wood has become a veritable alternative to concrete and steel, at least in many parts of Europe and North America. Reason enough to closely examine some of the causes – as well as assumptions – behind this material’s popularity. In the context of climate change, the main argument for using wood typically is the fact that it is a renewable resource and that it stores CO2. Hence, one could argue that wood deserves the grade “sustainable” more so than other materials.

Tatry forest. Photo: Dominik Slomka (cc)

17

Rethinking Wood

The sustainability of wood, however, is not absolute. What is often overlooked is that all-natural wood is hardly used in contemporary building construction. In this context, and as the fourth part of this book points out, one should bear in mind that the use of engineered wood and wood-based materials, the production of which typically involves adhesives, is not entirely unproblematic. Moreover, the fact that wood is a renewable resource does not mean that its availability is unlimited. Its sustainable production requires a well-balanced and sustainable forestry. In addition, we need to think of more resourceefficient ways of using wood in buildings. In his thought-provoking essay, Kiel Moe shows that wood fulfills many of the criteria that we would define as desirable goals when developing a new building material. More importantly, he encourages us to rethink established notions, and hence question our general understanding of wood and in particular its use as a building material. He suggests that in order to be truly sustainable, we need to unthink or abandon the ubiquitous yet rather narrow understanding of wood in building construction and architecture that typically disregards the ecosystem behind it, namely the forest. In architectural design, the relational and reciprocal dependencies between the components involved, as well as the relationship between the parts and the whole, typically play an important role. Moe argues that we need to expand this systemic way of thinking to materials and energetics as well. As mentioned earlier, wood is a renewable but not an infinite resource. In this context, Mark Hughes highlights the necessity of using wood in a more resource-efficient way. One strategy of achieving this is the re-use or cascading of wooden elements that are embedded in already existing buildings.1 The concept of cascading is comparable to that of down-cycling: a building component is re-used several times, for iteratively less demanding purposes. The approach of Design for Disassembly (DfD) aims at improving the ease of recovery of the constituent elements of future buildings. Similar to Moe, he suggests that sustainability must not be sought in building construction and components alone. It must extend to the forest, which is the ecosystem behind the building material wood. In comparison to central and northern Europe and North America, the renewed interest in wood is only just emerging in large parts of Asia. The region’s booming economy and the high volume of new building construction would make the potential impact of using more wood tremendous. Michael Budig discusses the current situation and future possibilities of building with tropical timber in Southeast Asia. A relevant and promising reference regarding current developments in the area of material re-use in building construction is the project Mine the Scrap, developed by Tobias Nolte and Andrew Witt of Certain Measures. As part of this project they developed a software tool that scans scrap elements from demolished buildings and rearranges them into new architectural envelopes (Nolte and Witt 2016).

1.

18

Moreover, he presents the results of an architectural design study that he conceived and carried out at Singapore University of Technology and Design (SUTD), which aimed to increase the awareness of tomorrow’s architects and designers regarding the qualities and potential of this building material in this region of the world. In the long run, the aim is to cultivate a higher and more widespread appreciation of wood, as well as an increased use, and hence a likewise increased value of this material, thereby helping prevent the clearing or torching of tropical forest areas. Although wood is probably the most sustainable building material currently available, there are still many things we can learn – and some things that we should unlearn – about it. The authors of this part introduce different approaches that address our complex relationships to this resource, including research into forestry and systemic thinking. Instead of focusing on isolated solutions, we should expand our goal of an increased sustainability to a larger scale, to material and energy flows. The available volume of wood for construction is limited as it relies on a sustainable and balanced forestry. With new approaches, like wood cascading and Design for Disassembly, we can help to keep this balance intact.

19

20

Think Like the Forest: Maximizing the Environmental Impact and Energetics of Building Timber Kiel Moe

If we were to design a building material from scratch today, certainly some of its key characteristics would be that it: • captures carbon, rather than emits it, during its production • has production and extraction processes that can augment the cultivation of biodiversity • does not require fossil fuel consumption for its production, extraction, and processing • is renewable at the scale and rate of human consumption • is based on a cellular solid material architecture that can modulate both heat and humidity in a single material • has a thermal diffusivity between that of insulating foams and dense thermal masses • has a thermal emissivity lower than human skin • fosters good health and air qualities, and is certainly not toxic • is fireproof for construction • has a construction ecology readily mapped and quantified • has a knowable and negotiable political ecology • has forms and scales of unequal ecological and economic exchange we can readily address • is capable of enabling small-scale economies • is, as Ivan Illich would say, “convivial” i.e. it is easily workable with simple tools by everyday people • is open source: its core intelligence is non-proprietary • incorporates feedback as it adjusts and adapts to changing conditions The above is a partial list of properties we might associate with wood. No human-derived material/energetic system approaches the nuanced complexity and efficacy of the processes that make “trees” and thus the building material we call “wood”. Wood is a 400-million-yearold material that has evolved and assembled a set of relationships and system properties that we, as human beings, have only begun to understand (Kohn, 2013). Surely there are many more relationships and properties than we have yet to comprehend; or perhaps simply 21

Rethinking Wood

cannot, or may never, comprehend. Thus, more than designing a new material, or even a new material process, what the world most needs today is: • a new way for humans to fully understand a material (like wood) • a new process for acquiring knowledge about a material (how we might research and teach “wood” architecture) • a new way to understand material processes (not merely as engines of commodification and product-making) • and, perhaps most importantly, the way a material relates to the world in the fullest possible sense; not just to some discrete component-object in that world (such as a “building”) Together, this would constitute an “advanced” conception and use of wood. In this regard, though, we must acknowledge that architects have been trained, literally, not to see the forest for the trees, nor even to see the trees when it comes to wood. Even worse, architects typically have little idea of what wood is, or what it does or could do, much less what forest do or could do. Too often construed as the simplest—perhaps most banal of building materials—wood is typically presented as rudimentary in construction courses and textbooks, and thus too often perceived to be in need of new methods and processes. So within the abstract domain of “wood”, it is easy for architects to get lost in the recent proliferation of contemporary claims about timber construction, processes, and possibilities, which are indeed important in themselves and worthy of our interest. However, it pays to recognize that the most important and advanced thinking about “wood” is not, yet, in the building industry. The reality is that there is no such physical thing as “wood,” just as there is no such thing as “a tree.” There is the forest: that amazingly complex and vital set of relationships and processes from which modern western minds vainly attempt to isolate the idea of a tree, or the even more abstract concept of “wood.” We might have a loose, abstract association around a set of materials and properties that we might describe as “wood.” Or we might collectively hold in our minds an abstract, children’s comic book illustration of a tree. These isolating abstractions are violent abstractions that already do epistemic and ecological damage in the very way that we conceive of and name them; to say nothing of the way architects further abstract and externalize wood. We need a conception of wood, like forest, that alludes to the complex and vital set of relationships and processes that are inextricable from the “wood” we use in buildings. To understand wood—indeed to rethink wood today—is often to think about wood for the first time in a serious way. To rethink wood and its assemblies in architecture today, likewise, is to rethink what we think an assembly is and what it can do. Not just building assem22

blies, but all the socio-bio-geophysical assemblages that presuppose those building assemblies (Bennett, 2010). To rethink wood, we need to question our assumptions about assemblies such as trees, as much as the forest, and certainly wood’s complex, sublime relationship to civilization and the planet as a world system. We need to think about building environments in relation to living environments in entirely new ways. This requires us to understand the specificity of what we call wood and all it actually does and can do in the simultaneous assembly of both living environments and in building environments. As one starting point, it probably means dropping the abstractions we use to externalize the consequential dynamics of “wood” in our pedagogical and professional spheres. The degree to which architecture might attain what we once associated with “sustainability” is the degree to which architects will not merely build with wood, but rather finally understand the deep ecological implications of building with living materials. In the case of wood construction, a homeorhetic relationship with living systems will only be understood—and indeed will only be possible—when we understand that building and assembling forests is as important as building and assembling architecture. We need to grasp how related and potentially mutual these acts of building in fact could be. To build with timber we need to foster timber first. For only then might we understand what it actually means to build a world, rather than just consuming our abstractions of the world—like wood and fuel—to build even more abstract buildings. Once we begin to understand the larger set of relationships and processes that encompass what we call “wood,” we will understand that we are merely re-directing the hardened edge of one of the earth’s most evolved and astonishing dissipative structures. And herein lies the ostensible purpose of this contribution to this book: to address the energetics of building with timber. As it turns out, understanding “wood” as a unit of living systems such as trees, forests, and people is an excellent way to learn more about energy and energetics. Wood offers lessons on energy because it is a clear example of our chronic problem with energy: we confuse and conflate fuel for energy. We ubiquitously use abstract nomenclature (like fuel) for much more exacting and rich energetic processes. The fate of the world is trapped in that dangerous abstraction. Given our fuel-centric ontology of energy, we are not trained to see the larger dissipative structures of energy cascading through the universe and our planet. We are epistemically and methodologically misguided in this way about energy in our culture. Rethinking wood can help us relearn what energy is, and what energetics might offer us as a way to act in the world.

23

Rethinking Wood

Energetics Just as we are trained not to see the forest for the trees, or the trees for the wood, we are likewise trained not to see the dissipation of energy through the vast array of nested dissipative structures and systems throughout the universe and in our living planet. We only see and commodify fuels as specific forms of energy and discuss their relative “efficiency”; in a similar way we are trained to see wood beams and panels, and discuss a few of their structural or thermal properties as abstract entities. To learn to see the forest for the trees and its wood is itself a lesson on how to see energetics beyond the abstraction of fuel. To learn that lesson would engender far greater and further-reaching impacts on collective dynamics like energy consumption and biodiversity than any far smaller, instrumental claim about the thermal capacities of wood in buildings, or their carbon sequestration dynamics. These behaviors, properties, and dynamics related to timber building are important, but their relationship to a larger set of energetic and ecological processes are imperative to understand. If we are at all serious about the core concept of sustainability, we need to grasp and hold onto these larger, paramount lessons about energy and forests. In building science, more than ever before, we need to think on a macroscale rather than a micro-scale. We need to get out of the lab and studio and into the field and forest. This contribution is but a sketch of what some of these lessons might be as we collectively continue to think about how we can maximize the environmental impact of building with wood.

Wood “Wood” is the by-product storage media of our planet’s most astonishing and effective energetic process: photosynthesis. As autotrophs, flowering plants like trees have evolved over hundreds of millions of years to absorb diffuse and low-quality, but extremely abundant formations of energy—sunlight—and mix it with diffuse and “simple” compounds like water and carbon dioxide. One of the products of this photosynthetic process is cellulose sugar. These sugars are channeled into the supple solid mass structure of the woody plant-tree; the bulk material that we use as timber building components. As the plant grows, it is effectively a 3D printer (Craig, forthcoming 2019). It deposits cellulose material as it grows up towards sunlight, while simultaneously building a vertical vascular structure that it uses to circulate water and other compounds throughout the dendritic organization of the organism. This vascular structure effectively makes the tree a pump that links the ground to the atmosphere, locking carbon into the soil and either emitting moisture into the atmosphere or into the soil. (In this regard, it is not surprising that wood is an ideal material to not only structure a building, but also to modulate its humidity and moisture content, while also addressing the thermal 24

behavior of the building (Hameury 2005)). The expanding horizontal girth of this vascular mass increases its capacity as a pump, but also its structural capacity, thus permitting the organism to grow taller and thus produce more branches and leaves. As such, this organism steadily increases its intake of more and more carbon, sunlight, and water through this growth, in order to support further growth. The net result is an organism that evolved over 400 million years to maximize the intake of simple compounds and diffuse energy—like water, carbon, and photons—in ways that process and mix them to produce more complex compounds and trophic relationships with adjacent species, including humans, through its modes of feedback reinforcement. It is, in short, a model of thermodynamic optimism and ecological generosity; a stark contrast to the pessimism of our current model of isolating, “self-sustaining” survivalism (Braham 2011). We can marvel at the seeming “simplicity” and sheer effectiveness of this highly evolved organism. Nevertheless, the above description is merely emblematic. It is, itself grossly abstract and inadequate to offer insight on the far more and ingenious evolutionary adaptations and mutualities that particular species co-evolve in particular places. This requires thinking of wood beyond the tree, and the tree beyond our nascent understanding of how it relates to the forest of larger bio-geophysical processes and histories. For instance, if we currently consider wood as only the above-ground portion of the tree, then what are we omitting? Dr. Suzanne Simard from the University of British Columbia points to the intensity of the rhizosphere activity as part of forest communication and metabolism that presupposes “wood” production. As one example, forests metabolize decayed salmon brought into their milieu by bears and wolves using nitrogen from the decayed fish to signal other trees through fungi, trade nutrients, and fix carbon (Simard 2018). Or consider the lifetime study of tree “architecture” in the tropics led by Francis Hallé that helped scientists to understand the tree as a community, rather than as an individual (Hallé 1978). Even our ignorance and abstractions of a particular tree species, such as Gingko biloba can be re-examined. The Gingko is a living fossil that survived both the dinosaurs and the atomic bomb through its structures and intelligence (Del Tredici 1991). If we are in fact motivated by survival, then attention to expert species might be advisable. Such examples and models of study are replete with possibility for design, as elucidated by Rosetta Elkin’s advocacy for bringing the vitality of plant life back into the profession of landscape architecture (Elkin 2107). Erwin Thoma offers some insight for architecture, but models of design are indeed scarce in this regard (Thoma 2014).

25

Rethinking Wood

LEAVES

TRUNK

BEAM

BUILDING

(a)

Solar Emergy Flow:

6E9

6E9

E4j/t

6E9

6E9

6E9

500 100

Solar

500 600,000

20

150

100

0

(b)

6

1000 100

0

100

40

250

2000

Aggregated j/t (c)

Solar

6E9

0.6E7 6E7

0.6E6

0.6E5

0.6E4

6E6

6E5

6E4

Feedback

Solar Emergy = 6E9 Solar Emjoules/ Time

(e)

Solar Transformity, sej/j

(d)

Energy Transfer, j/t

6E9

6E7 6 E6

6E5

6 E4 100,000

10,000

1

100

0

1

1,000 2 Transformation Steps

3

4

Figure 1. “Wood” is the temporary, captured state of a cascade of low-quality but abundant energy inputs and more scarce, high-quality outcomes and feedbacks. Systems ecology diagrams track the flow from the sun to, in this case, the leaves, branches, and trunks to building components and buildings; tracking the material formations that emerge to dissipate this exergy along the way. The diagrams also track the feedback of material, energy, and feedback that flows back through the system and gives it structure. This is called design. The illustration is drawn by the author, based on Howard T. Odum, Environmental Accounting: EMERGY and Environmental Decision Making, New York, John Wiley and Sons, 1996.

26

Energetics of wood We will never understand “wood” unless we begin to grasp the seeming simplicity and effectiveness of the energetics of this organism that produces wood. Woody plants are unmatched as a dissipative structure in how they have evolved to degrade available energy gradients. Consider the cascade of its energetics (Figure 1). Low-quality energy enters the organism, and higher-quality formations and concentrations of that energy emerges as that intake energy cascades through the system again and again as feedback. To cascade and feedback energy in this way, trees have developed mutualities with adjacent species at all trophic levels to further dissipate the available energy. At some point it is sheer hubris to isolate one species from its obligates. For example, humans breathe the “waste” products of these organisms— oxygen—while these organisms transpire our “waste” carbon dioxide. But that carbon dioxide-oxygen cycle is ambiguous without considering the mutualities that presuppose the “tree” in the soil conditions that host the growth of the flowering plant, the tacit agreements with other species to spread seed, or the capacity of the tree trunk to host thousands of species and organisms once the “tree” has died and has fallen to the forest floor. This is where our abstractions of “tree” and “wood” become more pronounced; and where they do the most harm. Why would we fetishize the “tree” its upright living state over its next state of supporting life in other ways on the forest floor or in an apartment building? How can we abstract “wood” from the rich, vital complexity of these long-evolving processes? How can we abstract “biomass pellet fuel” or “R-value of wood” from this complexity? Each of these abstractions has their use and purpose, but we must recognize that we habitually fail to link insight on individual abstractions and studies to the larger context from which we abstract. This failure to link our study of isolated abstractions in architecture to larger systems is at the core of our most “unsustainable” practices and beliefs.

Forests as dissipative structures One simple, but illustrative measure of the energetic effectiveness of the dissipative structure we call a forest comes from landscape ecology. Landscape ecologists Jeffrey Luvall and H. Richard Holbo studied the level of radiation emitted by varied patches of landscape (Luvall and Holbo 1991). The patches were all in the same region, so they received similar levels of solar insolation. The results were that an old growth forest emitted lower temperatures than a quarry, an agricultural field, or a new growth forest (Table 1). The quarry merely exhibits the simple absorption and re-radiation of that solar energy from the stone mass and surface. Very little work is extracted in this landscape patch and it has no mechanisms for further productive dissipation. In contrast, the old growth forest does more work, having evolved a more sophisticated dissipative structure and organization. This is be27

Rethinking Wood

Quarry

Clearcut

Douglas Fir Plantation

Natural Forest

400 year old Douglas Fir Forest

K* (watts/m2)

718

799

854

895

1005

L* (watts/m )

273

281

124

124

95

Rn* (watts/m2)

445

517

730

771

830

T (°C)

50.7

51.8

29.9

29.4

24.7

62

65

85

86

90

2

Rn/K (%)

K* = incoming net solar, L* = net long wave out going, Rn = net radition transformed into nonradiative process at surface, Rn/K* = percent of net incoming solar radiation degraded into nonradiative process. [excerpted from Jeffrey Luvall and H. Richard Holbo, “Thermal Remote Sensing Methods in Landscape Ecology,” in Quantitative Methods in Landscape Ecology, eds. Monica G. Turner and Robert H. Gardner (New York: Springer-Verlag, 1991), 127–52.]

Table 1. Radiation emitted by varied landscape patches. Excerpted from Jeffrey Luvall and H. Richard Holbo, “Thermal Remote Sensing Methods in Landscape Ecology.”

28

cause over time the old growth forest has developed highly integrated and diverse systems that put the same energy input to work in complex ways—such as in the growth of diverse flora and fauna at many trophic levels. In short, an old growth forest reflects a set of highly evolved, complex interrelated systems that cascade and dissipate the same solar energy more effectively through their structure and organization. More work is extracted from the same solar gradient as more living processes tap into that gradient. This is not solar energy as a “fuel.” This is a complex organization of coupled living systems that collectively work towards exuberance, abundance, and vitality from the lowest quality inputs. A directly analogous exuberance and vitality ought to be the purpose and aim of design. What we call “wood” is one of the most complex and vital of building materials, primarily because it is living material that is part of living systems. No single person knows all it is capable of doing or should be doing, not only in architecture, but also in the ecology of our planetary urbanization. However, there seems to be no more important or ambitious task for those who make claims on how to couple environment and architecture than attaining non-abstract, non-externalizing conceptions of “wood.” It is a profound first step towards rethinking wood. As we rethink wood we must finally grasp the hierarchy of energy and other life-sustaining processes inherent to what we call wood, tree, or forest. We can then finally begin to understand how to maximize the environmental impact of specifying wood in contemporary construction. But in doing so, we might finally recognize that we would not be specifying wood per se (thinking like a wood beam), but rather we would always be specifying a vast assembly of nuanced processes and properties that in aggregate could, if adequately understood, meaningfully adopt the adjective “sustainable.” We need to evolve far more sophisticated dissipative structures as the basis of our energetics in architecture. What we once called wood is an important part of that dissipative structure design. But hopefully, as inhabitants of the world, we soon forget the bizarre abstraction of “wood” as a precondition of how we come to borrow the hardened edge of the world’s most sublime energetics—forests—as we build with forest-based cellulose materials. In doing so, we would come to think, and act, like the forest.

29

Rethinking Wood

Gross felling 90 mn m3sk* Net felling 75 mn m3sub**

Imports Logs 6 mn m3 Exports Logs 1 mn m3sub**

10 %

45 %

45 %

Sawlogs 35 mn m3sub**

Bioenergy 5 mn m3sub**

Pulpwood 35 mn m3sub** Bioenergy (bark)

CHP plants

Bioenergy (bark)

Sawmills

Shavings + dry chips 14 %

Pulp and paper mills

Wood chips 31 % Energy

Recovered paper

Heat recovered

Wood products 55 %

Shrinkage, trimming allowance etc 8 %

Electricity and heat

After drying 47 %

30

Pulp and paper

Cascading Wood, Material Cycles, and Sustainability Mark Hughes

Summary Humankind has used wood for millennia, not just as a material for building shelters and other artifacts, but also as fuel for cooking and for warmth; however, wood is so familiar to us that we rarely stop to think about what it really is. With the global challenges of climate change, environmental degradation and population growth, there is an urgent need to reduce our dependence on non-renewable resources and to switch to materials and energy derived from renewable origins. At the same time, we need to ensure that the resources we do use—be they renewable or non-renewable—are used prudently. As a renewable and plentiful, though certainly not inexhaustible, raw material, wood can help us meet some of these challenges. However, if we are to increase our reliance on wood we must do this with care—we must ensure that we make the best use of our finite resources and ensure that the forests from where we obtain our wood remain healthy— we must “live within our means.” We can do this by becoming far more resource-efficient, by recovering and reusing the wood elements and materials that form our buildings and other artifacts. To do this, we need a paradigm shift in how we view buildings as well as how we design, build, use, and dismantle them; we need to think of buildings not just as spaces in which to live and work, but as repositories of valuable materials that can, at least partially, supply our future needs for timber as well as acting as temporal carbon stores.

Introduction, background, and context Humanity, and indeed life on earth itself it seems, may soon be facing an existential crisis. The global population has roughly doubled in the past half century and is set to rise still further from the more than 7.5 billion today, to a predicted 11 billion by 2100. Average global temperatures— and the oceans—are rising, whilst in our quest for raw materials to fuel our continued addiction to economic growth, we are rapidly depleting both renewable and non-renewable resources, creating environmental destruction along the way. We urgently need to change this paradigm. Forests play a vital role in combatting climate change by sequestering carbon, maintaining biodiversity, and providing other ecosystem

Figure 1. Utilization of logs in various wood products in Sweden (* mn m3 sk = million forest cubic meters, ** mn m3 sub = volume in million cubic meters, solid under bark). Source: Swedish Wood

31

Rethinking Wood

services. In addition, they afford the raw material—wood—for an array of products that can be used for construction and a host of other everyday applications as well as the provision of energy. The complex interplay between the natural world and humanity is set to change the way in which we view forests and the dynamics of wood use in the future. In this chapter, we will explore how the utility of this most valuable of building materials can be enhanced and how architects and engineers can play a vital role in this transition. Why wood? A significant part of the environmental impact of our activities arises from building construction. Buildings account not only for about forty percent of primary energy consumption and carbon dioxide emissions (European Commission 2008), but between thirty and fifty percent of material resources are used for building construction (Ecorys 2014), whilst construction and demolition waste accounted for thirty-two percent of total waste in the EU 27 in 2012 (European Environment Agency 2012). Clearly then, to create a more environmentally sustainable​society, we should place considerable focus upon the built environment, and buildings in particular. In this respect, most attention to-date has focused on reducing the energy consumption of buildings during the use phase by, for example, improving their thermal efficiency or providing on-site power generation. However, as the energy efficiency of buildings improves, the embodied impacts associate with their construction become ever more important in determining the overall, whole-of-life impact of the building. For this reason, there is growing interest in the use of low impact, “green” materials in buildings that are often from renewable or recycled sources. For many reasons, timber is arguably our most important green construction material—it is abundant, renewable, possesses good technical characteristics, and can be converted into a host of different “engineered” wood products with relatively minor inputs of energy and other materials. Importantly, wood products not only store carbon,1 but can also substitute2 “energy-intensive” materials like concrete, steel, and plastics. For these reasons, there is a good argument to increase our use of wood in construction and, indeed, it seems that we are now witnessing something of a renaissance in wood building. Nevertheless, if we do see a large increase in the use of wood in construction, can our forests sustain this increase? Currently it seems to be the case—especially if we begin to use species such as European beech (Fagus sylvatica) not usually associated with construction, in building 1. Roughly, 1.75 kg of atmospheric carbon dioxide is required to produce 1 kg wood; so one kilogram of wood can be thought of as “storing” one-and-three-quarter kilograms of carbon dioxide. This stored CO2 is only returned to the atmosphere when the wood is either burned or biodegrades. 2. The arguments for this are complex and highly dependent on how the materials are used.

32

products like laminated veneer lumber (Pollmeier 2018)—but concern has already been expressed by some about plans to increase wood use in the wider bio-economy and the impact on forests (EASAC 2017). The debate about how we use forests is often heated—and with just cause. Deforestation, along with other land use changes, accounts for significant levels of carbon emissions, as well as to habitat destruction that results in biodiversity loss and the collapse of populations. So, whilst the CO2 “neutrality” of wood products and the “substitution” effects gained by replacing energy intensive materials are clearly points in favor of using more wood products, this must be tempered with a sensitive approach to forest use—i.e. a sustainable approach. Forests and wood products The forests that supply the raw materials for wood products, as well as for many other purposes, also provide vital ecosystems services. Forests are important carbon sinks and stores, and influence the climate in other ways as well as maintaining biodiversity (EASAC 2017). This can present a conflict and begs the question: Do we prioritize the use of wood products, with their associated climate change mitigation potential of carbon storage and substitution, or do we prioritize the role of forests in carbon sequestration and the maintenance of biodiversity? Can we have both? Perhaps it is more of a question of balancing these different priorities, as well as taking into account other ecosystem services and societal perspectives, such as the need to maintain woodlands and forests for recreation. Many studies have pointed out that provided timber is harvested within sustainable limits, the combined mitigation effects of carbon sequestration in the forest as well as in long-lived wood products, together with the material and fossil fuel substitution effects, yield the best overall climate change mitigation potential (Taverna et al. 2007). In several parts of the world, such as Fenno-Scandinavia, the supply of timber from forests is plentiful and there is perhaps scope for increasing the sustainable harvest. However, with the ever-increasing interest in the use of forest biomass as a raw material within the bio-economy for a host of products, not just timber for construction, but also for other materials and energy, there are growing concerns about future supply. How do we meet this increased demand sustainably in the future? One way of doing this is to improve the efficiency with which we use wood in the first place. Little of the harvested tree goes to waste in the forest-products industry, though undoubtedly further improvements could be made in materials’ efficiency. If sawn timber, for instance, is manufactured, the co-products that are produced are generally used very effectively for other products or in other processes. As shown in figure 1, a log from a Swedish forest, for example, yields just under fifty percent sawn timber after drying. The bark removed prior to sawing, if not used to produce energy in 33

Rethinking Wood

the mill, can be used in horticultural applications. Wood chips and sawdust from the sawmilling process itself can be used to produce pulp for paper or other fiber products, or may also be used in energy generation. The entire log is thus utilized to produce either material products or energy. In many regions, the forest-based industries are highly integrated, adopting the principles of industrial ecology, where the waste from one process forms the feedstock for another. Particleboard or MDF (medium density fiberboard) manufacturing lines are, for instance, frequently combined with an onsite sawmill, supplying a proportion of the chips needed for production. It is after the production of these primary products that, in some areas, the material efficiency decreases markedly. Wood products and the circular economy Nowadays, paper and paperboard products, based on wood fiber are effectively recycled into new products. There is an efficient recycling infrastructure in place across much of Europe and over seventy percent of the fiber used in paper and board manufacture is recovered and recycled (European Paper Recycling Council 2016). Typically, these products have relatively short life spans—as cartons for liquids or as other packaging materials—and so without effective recycling, the fiber would be quickly returned as carbon dioxide to the environment. In doing so, their carbon storage potential, along with the energy “invested” in their production—the embodied energy—would quickly be lost and additional virgin fiber, from freshly harvested wood with the associated energy inputs, would be required. So clearly, the recycling of fiber from paper and board products is essential to increase resource efficiency and, without doubt, there is even greater importance in recovering, reusing, or reprocessing the non-renewable materials found in everyday products such as cars, mobile phones, televisions, computers etc. Closing the loop on material flows and creating a “circular economy” has recently been the focus of both political and societal debate that has given rise to the introduction of strategies to promote its implementation (e.g. European Commission 2015). To what extent these strategies can be implemented without creating new environmental problems remains to be seen. As recently argued, the circular economy does not really mimic the natural world as is widely thought, rather it is based on the notion that there are no losses in such a “closed loop” system (Skene 2018). With a little reflection, it is clear that this cannot be the case: When we try to recover materials there will never be one hundred percent retrieval, there will always be some losses. Materials will become dissipated in the environment and it will become ever more difficult to recover the last remaining remnants. This problem increases the more complicated our products become (take, for example, the problems of recycling and recovering the materials in electronic devices). We will return to this conundrum 34

later in the chapter, when we consider the design of buildings and wood-based building products. Wood use in construction: buildings as material repositories What is the situation with wood used in construction today? We must first recognize that wood products—sawn wood, engineered wood (laminated veneer lumber, LVL, glue-laminated timber, glulam, cross-laminated timber, CLT) and wood-based panels (plywood, particleboard, MDF, and oriented strand board, OSB) are generally utilized in situations where the product is used for considerable periods of time. In the case of structural timbers and engineered wood products such as LVL, glulam, CLT, plywood, and OSB, this may be for as long as the lifetime of the building, perhaps around fifty years. In this respect, the use of wood products in buildings is efficient, especially in terms of long-term carbon storage and the substitution of other building materials. Likewise, we should recognize that a considerable amount of wood can also be used in somewhat shorter-lived, non-structural parts of a building, such as cladding and interior paneling that are renewed because they are degraded, or because of use change, or aesthetics. If we discount the use of wood for energy, pulp, and other products—which comprise a significant proportion of the harvest—and focus only on those wood products such as sawn wood and wood-based panels that are likely to be used in construction, their total global production in 2016 amounted to around 850 million cubic meters (FAO 2016). If we consider that similar orders of magnitude of wood products are manufactured annually and that a significant proportion of these go into buildings year-on-year, it is easy to imagine that there is an enormous stock of wood tied up in this so-called “urban forest.” In one recent study from Austria, it was estimated that in excess of thirty million cubic meters of wood are tied up in building stock (Kalcher et al. 2017). Throughout much of Europe, very little of this material is recovered and reused in material form. A large amount is recycled into particleboard (which is often used in relatively long-lived products such as furniture), but a very significant proportion is simply chipped and burned for energy-recovery, thus returning the carbon contained within it into the atmosphere. This potentially represents a wasted opportunity; not only to extend the life and utility of the material, but also an opportunity to create economic activity by, for example, re-purposing or re-cycling wood products recovered from buildings in solid wood form. Of course, this over-simplistic analysis does not highlight some of the technical, regulatory, or societal challenges involved in recovering and reusing wood from buildings, nor some of the important details that can significantly influence the sustainability of wood use alluded to in the previous section. Nevertheless, the purpose of this chapter is to explore and discuss how we might rethink the use of 35

Rethinking Wood

wood in building construction and how we might harness the urban forest as a valuable source of material.

Environment, buildings, timber Figure 2 shows the forest carbon cycle. Forests absorb carbon from the atmosphere during tree growth and store this carbon in the biomass. Natural decomposition of the biomass and forest fires lead to the emission of carbon back to the atmosphere, though a proportion of the carbon from decomposed biomass is retained in the soil (this one of the reasons why deforestation is so destructive; there are not only the emissions from the lost biomass but also, because of erosion, a loss of the carbon stored in the soil). Harvested wood products (HWPs) that remain in the technosphere can be thought of as biomass “displaced” from the forest (which makes the term “urban forest” even more appropriate) and residing elsewhere until it is decomposed. As such, forests represent one of the key and most obvious means of helping mitigate climate change through their carbon sequestration potential and other effects and, if managed sustainably, they can foster and maintain biodiversity (EASAC 2017). In Europe alone (EU28), the amount of carbon stored in forests is estimated to be almost 10 billion tonnes (Forests Europe 2015) and since 1990, this amount has been increasing at a rate of about 137 million tonnes per year. The

Atmosphere

Wood products decomposition 87 ± 16

Forest NPP 756 ± 98

Wood harvest 92 ± 16

Fires 7±2

Decomposition 533 ± 105

Fossil fuels 1060 ± 100

Biomass sink 80 ± 15

Wood-based products sink 5±3

Soil carbon sink 29 ± 18

15 ± 6

Fertilizer 0±0

Fluxes in Tg C yr-1

Figure 2. European forest carbon cycle (Source: Luyssaert et al 2010). 36

amount of net amount carbon sequestered by woody biomass (i.e. after accounting for harvesting and losses due to decomposition) is about 100 million tonnes per year and is equivalent to about 10 percent of its fossil fuel emissions (EASAC 2017). It is therefore imperative to protect the continued role that the forests play in combatting climate change and biodiversity loss. On the other hand, trees are, and have always been, a vital source of materials with which to build our villages, towns, and cities and to produce the materials and energy we need in everyday life. In terms of the role that HWP play in storing carbon captured from the atmosphere during photosynthesis, this depends largely on the type of product that is manufactured and the application in which it is used. Likewise, the potential of a HWP to replace another, more “environmentally damaging” material depends on the application in question. Carbon storage in timber products When wood is harvested, only a proportion is converted to timberbased products such as sawn wood, wood-based panels, engineered wood, and the like, the remainder is processed into pulp, paper, board, polymers, composites, or used as energy. Even if timber is harvested exclusively for wood products, only a proportion of the material will end up in those products. This is due to unavoidable losses in conversion. Typically, when manufacturing sawn wood, about forty-five percent of the biomass is converted into product, the remainder, comprising chips and sawdust, is used in other processes to manufacture, for instance, pulp for paper or board products, or to create particleboard. The type of product depends on the local infrastructure and markets. If a particleboard or pulp mill is not nearby, then the co-product may then be used to produce energy. So, when a tree is harvested it is converted into a range of products that typically have different life spans. These different product groups can be assigned different so-called “half-lives” (the time taken for half of the material produced to be returned to CO2) that depend on their typical uses. For IPCC accounting purposes, paper, for example, is considered to have a half-life of two years, whilst sawn wood has a half-life of thirty years (IPCC 2006). This then clearly has an impact on the length of time carbon is stored and explains why, in terms of climate change mitigation, the use of timber in long-lasting products, such as those found in buildings and furniture, is preferable to shorter lived products, albeit in the long-term, all wood-based products can, and are, typically regarded as “carbon neutral.” This is because, ultimately, the tree from which any wood product is produced will regrow (assuming the harvested forest is regenerated). From a climate perspective, however, carbon dioxide emissions are carbon dioxide emissions and the important question is: When is the carbon dioxide held in the wood product emitted relative to the harvesting of the 37

Rethinking Wood

All numbers without units are in M m3 swe (solid wood equivalent)

growing stock, total M m3 o.b. (incl. Stumps) 21.021.0

o.b. 177.7

total wood import

wood outside forest

net annual increment 1277.0 M m3 o.b.

remain in forest

NAWS 546.00 M m3 o.b.

AWS 731.00 M m3 o.b.

132.5

33.4

94.8

forestry biomass 543.7

total wood export

wood resources from trees (WRT) 577.1 M m3 swe

107.8

38.8

208.8

260.6

103.7 34.0

industr. res. 176.3

15.5

pulp prod. 141.8

products in use

paper prod. use 82.0

19.1

cascade factor 1.57

wood prod.

314.9

217.5 97.4

129.4

recycling 144.9

paper 185.2

products in use

wood 169.1

15.5

incineration 13.5

disposal

113.6

15.9

33.2

carbon sequestration

129.4

20.6

energy use 337.2

households

83.1

168.6

85.5

biomass power plants other

Figure 3. Sankey diagram of wood flows in Europe. (Source: Mantau 2012)

38

biomass power plants wood industries

tree from which the product was manufactured? Even though a wood product might, over the long-term, be regarded as “carbon neutral,” when a tree is harvested to manufacture a product, it may take around eighty years for another tree to grow and replace it. If the product created only lasts five years before decomposition, then there will be a net emission of carbon dioxide during the remaining seventy-five years when the replacement tree is growing. If, however, the product lasts eighty years, then the carbon stored in that product will not be returned to the atmosphere until another tree has grown to replace it, thus resulting in no net change in carbon dioxide emissions at any time. This is, of course, a very simplistic analysis, but does serve to explain why long-lasting wood products are better at storing carbon than short-lived products and why we should encourage the use of such long-lasting products for carbon storage (Bellasen and Luyssaert 2014; EASAC 2017; Härtl et al. 2017). Once again, this is a slightly simplistic analysis that might lead to erroneous conclusions about how wood should be used as it neglects the important role that wood products have in substituting “energy intensive” materials that might account for considerable carbon dioxide emissions, but it does serve as a general rule-of-thumb. Substitution of energy intensive materials So, whilst one element in a wood product’s “carbon equation” is its ability to store carbon and thereby simply remove it from the atmosphere until such time the product decomposes, another is the benefit derived from replacing materials that might result in greater net emissions when used in building construction. Concrete and steel are probably the two most cited examples of construction materials that have a significant overall impact on the emissions from buildings. The specific embodied energy of concrete is around 1–2 MJ/kg, which is less than that of wood products (ranging from 8 to 15 MJ/kg depending on the product), whilst that of steel is around 24.4 MJ/kg (Hammond and Jones 2008). Concrete, however, generally constitutes a far greater mass in a building in which it forms the main structure, thus the overall impact is generally higher. Takano and co-workers (Takano et al. 2014), for example, compared different materials in the same construction (same functional unit) and found that concrete had the highest overall life cycle primary energy balance for the structural frame of the hypothetical building. Whilst the embodied energy of steel is greater than that of wood, its specific mechanical properties are comparable and therefore the total mass of steel needed in a building is generally lower. So, although the impact might be greater than that of a comparable wooden structure, it remains less than concrete. The impact of material choice on the energy balance of a building depends largely on details, such as what happens at the end of a building’s life (i.e. how are the materials subsequently used) and so 39

Rethinking Wood

the benefits to be gained from substitution are difficult to predict. Overall though, wood-products offer good opportunities to replace other existing materials as well as acting as a long-term carbon storage “mechanism.” Prolonging the life of wood products increases their capacity to store carbon for longer as well as avoiding the use of material with greater impacts, thus increasing their capacity to mitigate climate change. In addition, there are other good reasons to enhance the life span of wood products. These will be explored in the following section.

Strategies: reduce, reuse, recycle If we consider the flows of wood into the technosphere (Figure 3), approximately 170 Mm3 of wood products (solid wood equivalent) are used annually in Europe. Clearly not all these products will end up as part of buildings, but around seventy percent are used in construction related applications (Hurmekoski 2016). As such, around 120 Mm3 wood products enter the built environment year-on-year, representing not only a significant amount of carbon stored, but potentially emissions avoided due to the substitution of alternative materials. Of course, whether the use of other materials is really avoided is a matter of speculation. Storing carbon for longer By making assumptions about the building types constructed historically, their life spans and renovations, it is possible to predict the volume of wood materials contained within the building stock and the volumes that might become available at some future date for re-use or recycling (Kalcher et al. 2017). We will come back to this point in more detail later, however, what is important to consider for the time being is that by extending the life span of buildings, the materials contained within them, including the wood products, can potentially remain in service for longer. This has direct impact both in terms of the carbon stored (in the wood products) as well as the avoidance of additional materials needed for new constructions and their associated impacts. Extending the life time of the structure might be especially important if we consider the fact that under favorable conditions the life span of a wood product can be many decades or even centuries, so removing it from service prematurely because a building is altered or demolished for reasons other than decay, for instance, represents an inefficient use of that product, particularly if the wood from a demolished building is simply burned for energy and is not reused in another material form beforehand. It might, perhaps, also be important to consider what we really mean by the life span of a building or structure and exactly how it can be extended? Rather like historic vessels, such as the Cutty Sark clipper ship, in which little of the original structure remains due to continual renovation and replacement of materials, buildings of all sort, for one reason or another, often undergo renovation or refurbishment 40

work. In wood buildings, cladding, for example, generally has a limited service life and is replaced regularly to protect the structural fabric. Buildings are also often altered or extended to repurpose them and in doing so, materials will be used, which will clearly have an impact. Perhaps it is time to rethink the design of buildings in order to promote the possibilities for extending the utility of the materials that are contained within them and from which they are constructed. Could, for instance, buildings be designed with far greater adaptability in mind, so that there is no reason, other than the materials reaching the end of their useful lives, for demolition? Could buildings be designed with far greater flexibility so that worn out components could be replaced or the structure be adapted to new uses? This is not really a new concept; in vernacular architecture, there many examples of buildings, such as the ubiquitous log houses of Scandinavia, that can be readily repaired and renovated and in which even the structural fabric of the building, the logs, can be replaced adding considerably to the life span of the building. This would represent a significant shift in thinking about how buildings are viewed. There are already signs of such a shift; the Modern Log City of Pudasjärvi in Finland has been designed to last 150 years! How can this thinking be incentivized so that it becomes mainstream? Wood cascading How can we utilize wood materials that already exist in building that are currently being demolished? Can greater utility be extracted from the materials before they are used to produce energy? The cascade use of wood (Figure 4), that is using wood first in solid wood form before conversion into particle-based products, fiber-based products, and finally as an energy source, has been the subject of research in recent years (Höglmeier 2013). It has been demonstrated that the resource efficiency of wood use can be improved through the implementation of cascading and its environmental impact reduced (Diyamandoglu and Fortuna 2015). Currently in the EU, forest-based industry companies such as the Egger Group utilize clean, uncontaminated, recycled wood in their particleboard manufacture, up to a rate of about thirty percent (Egger 2018). Implementing an additional cascade level prior to particleboard has, however, proved to be more problematic in practice. While the abovementioned benefits to be gained through the introduction of wood cascading are, in theory, even greater if recovered solid wood is cascaded in solid wood form (since the cascaded solid wood will still be available for chipping to produce particleboard at some time in the future) before chipping to manufacture particleboard, its wide-scale practical introduction has so far proved elusive. There are a number of reasons for this. Price and cost effectiveness, industrial scalability, quality, cleanliness, and logistics are all seen as key barriers to implementing solid wood cascading, at least in Finland (Husgafvel et al. 41

Rethinking Wood

2018), though these barriers are likely to be relevant to a greateror-lesser extent elsewhere too. Provided the wood itself is not treated with preservatives, paints and the like, separating ferrous metals such as nails and screws is relatively straightforward if the wood is shredded, or chipped, to manufacture particleboard, or for energy. Retaining the form of solid wood so that it can be processed into another composed solid wood product is, however, limited either by the need to remove nails, screws, and other fixings beforehand or to cut the wood that does not contain any metallic fixings and separate it from the wood that does. In the former case, the costs of removing nails and screws manually would most likely be prohibitive, whilst by cutting away the wood containing the nails and screws, might leave very little intact “clean” wood for reprocessing. In other words, there would be a significant dissipation of material. This would, of course, affect the economic as well as environmental viability of solid wood cascading at large scale and might go a long way to explaining why it has not been introduced before now, except in some very specialized cases such as the recovery and reuse of old floorboards and doors for example, where the very fact that the material is old and has its own “character” is appreciated and the price premium commanded covers the cost of more labor-intensive handling. Time /number of cascade steps

Material life Product life 1

Forest

Round wood

Product life 2

Product life 3

Product life 4

Solid wood-/ Veneer-Products

Particlebased products

Energy wood

Fiberbased products

Chemical products

Decrease of wood unit size and application options Recovery, sorting, processing

Losses during recovery, sorting and processing

Cascading wood

Figure 4. Biomass cascading. (Source: Höglmeier et al. 2015)

42

Incineration for energy Re-use incl. recovery, sorting, processing

The influence of design on the cascading potential of wood In modern wooden construction, the need to improve energy efficiency has led to the introduction of building technologies that can make the recovery of wood materials difficult. In a recent study conducted in Finland, for instance, the demolition of a kindergarten built in the 1970s using lightweight wood framing was studied to better understand the cascading potential of the wood products it contained (Sakaguchi et al. 2016/17). The findings were illuminating. Wood materials which, intuitively might have been thought to have excellent cascading potential, such as the large dimensioned structural timbers, proved to be difficult to separate from other materials like insulation without a great deal of effort and, furthermore, they contained many metallic fixings that were difficult to remove. This is perhaps a good example of how building technology has developed in a way that makes solid wood cascading rather tricky. Another example was the roof trusses. The nail plates held the members rigidly, resulting in the wood invariably being broken during demolition and therefore having poor cascading potential. This is unfortunate, since the wood material is of good quality and generally does not contain contaminants such as preservatives, so is eminently suitable for cascading, if only it could be recovered without damage. In contrast, small-sized roofing battens were more readily removed intact. There was some speculation as to whether the demolition process could be altered to yield a greater amount of usable wood at little or no additional cost, but there were no firm conclusions drawn and further study is required. A more detailed study of the demolition procedure might highlight “hot spots” in the process which could enable a greater proportion of the wood material to be recovered in a form suitable for cascading. In parallel with developments in building technology, there have been advances in wood products too. New structural “engineered wood” products like LVL and CLT have been introduced that have, some might say, revolutionized wood construction and enabled building in this medium to gain new momentum. In fact, it might be argued that these new wood building products have influenced building technologies themselves. Engineered wood products tend to be characterized by the use of high performance structural adhesives that hold the constituent wood elements together. Separating these glue bonded products at end-of-life to recover the wood, if they are to be used in further materials forms, requires new attachment solutions. One of the features of these engineered wood products is the fact that the glues used are extremely durable and long lasting, making the products themselves very robust. This makes replacing the structural adhesives with other bonding agents that ease the separation of the wood components more difficult, as it generally means that those agents are not as durable. This might well point the 43

Rethinking Wood

way toward reusing or repurposing the wood products themselves, rather than recovering the materials and reusing these. There are good examples of long-lasting engineered wood products, such as LVL, which hint at this approach. Mechanical recovery of larger wood elements, such as those found in glulam or CLT might be a viable option simply by sawing away the glue lines. Whether this approach is viable for plywood or similar products, for example, is rather questionable however. Alternatively, purely mechanical fixing as in products such as Brettstapel might be another means of recovering wood material from engineered wood for cascade use. Theoretically, no such problems exist with sawn timber though, of course, fixings bonded to other materials and the use of chemicals as preservatives limit the cascade potential. Is a shift in thinking about the design of buildings and wood products required? Wood construction itself, or at least a greater penetration of wood in construction has the dual benefit of storing carbon and substituting alternative materials. Extending the duration of wood products’ use in construction can be achieved by lengthening the life span of buildings by, for example, making them easier to adapt to new functions, or by recovering the wood products and either re-using them after suitable refurbishment and inspection, or by recovering and reusing the wood they contain. In other words, adopting the cascading principle to the use of wood in construction. Current construction methods and the design of wood-based building products, however, do not necessarily support this aim, so perhaps a shift in how buildings and wood products are designed and used is required. The problem of how to increase the amount of wood cascaded is related to both technical and commercial constraints. Presently, wood recovered from buildings is generally of small dimension, contains organic and inorganic contaminants and is of poor quality. To increase the yield of higher quality material for re-use, new approaches to the design of, for example, timber connections and building systems will be required, in other words (wooden) buildings will need to be designed for disassembly and re-use. By adopting Design for Disassembly (DfD) principles, that is designing buildings or products with ultimate disassembly, for adaptation, refurbishment, or end-of-life in mind at the design stage, a greater proportion of the materials employed in their construction could potentially be reused or recycled. With the use of wood in construction, we could conceivably recover a greater proportion of higher quality wood from buildings (and, potentially, other sources) making its use far more resource-efficient if the buildings were designed to DfD principles. At the present time, there are few real examples of DfD (Rios et al. 2015) applied to construction in general, yet wood buildings, and 44

both research and practice is urgently required to demonstrate the potential. There are probably far more reasons not to instigate DfD in wood than to implement it, including regulatory and policy barriers, but perhaps the most significant barrier at the present time is likely to be that it simply has not yet been done and so there are no, or at best only a few, examples to refer to and there is little incentive to be a pioneer. It is difficult to expect a client wishing to build a new building to agree to design a building that can be easily taken apart a century or more from now! But it will take someone to lead the way and it is likewise important for there to be architects and engineers able to take up the challenge, or lay down the challenge as the case may be. Doubtless changes in policy to help support this kind of shift in building construction are needed, as well as an educated workforce to implement them. Buildings have in the past been built with disassembly in mind and so there is no reason why a modern interpretation of this way of thinking cannot be reintroduced. It is time to rethink how we use wood in construction.

Conclusions Forests are not only vital in terms of providing ecosystem services, they are also an invaluable source of a renewable, though not inexhaustible, raw material for construction. Given the importance of the bio-economy and the urgent need to move away from non-renewable and fossil resources, greater emphasis will need to be made on improving the efficiency with which we currently use wood, to ensure that sustainable forestry practices are maintained in the future. One way of achieving this, which has been quantitatively demonstrated in theory at least, is to cascade wood. Alongside strategies to extend the useful life of buildings, this could ease the pressure from the construction sector on the primary resource and so help ensure the long-term sustainability of both forestry and construction. At present, little wood recovered from buildings is cascaded in solid wood form. True, wood is cascaded in the form of particleboard, but an opportunity is perhaps being missed to introduce a new, higher level, cascade—that of solid wood cascading—before size reduction to particles and the irretrievable loss of potential to form solid wood products. For high value, solid wood, applications, the quality of recovered wood needs to be guaranteed and for this, new approaches to deconstruction or disassembly (rather than demolition) need to be introduced, together with a concerted effort to design buildings with disassembly in mind. Given the constraints on materials’ recovery, quality, and its current use as an energy source, implementing a commercial wood cascading enterprise appears difficult at the present time and no successful business models have yet been developed; nevertheless the concept is sound and should continue to be developed and taken-up by trailblazers. 45

Rethinking Wood

Roof

Roof Plate-Frame

Raft

Raft Supports

46

Wood on the Rise: A Speculative Approach to Timber Construction and Joinery in Southeast Asia Michael Budig

Summary In the rapidly growing Southeast Asian cities, wood has largely disappeared as a construction material. After its recent global revival, it has also gained massive interest in cities like Singapore and is slowly being reconsidered as a serious alternative to concrete and steel. While the industrial knowledge of tropical timber is scarce, vast knowledge in vernacular construction is still accessible. This chapter showcases a series of design speculations that challenge conventional wood construction by translating small-scale physical experiments on wood joinery into parametric design models for micro-towers of up to ten stories high. The various concepts are inspired by vernacular building techniques, but also by less common wood technologies exploring combinations of wood with fibrous and fabric materials to form composite systems. They anticipate suitable tropical solutions and are categorized according to geometric configurations and different types of joinery.

Introduction and background Singapore, the island state on the southern tip of the Malay Peninsula, is surrounded by vast forested areas. It has been regularly covered in haze from man-made forest fires in the neighboring countries of Malaysia and Indonesia. Air pollution reached its highest on record up to date in 2013 and very high levels were recorded again in 2015. At around the same time, various stakeholders discovered an interest in timber construction. Does this represent and absurdity or an opportunity? We believe it is a chance to rethink the relationship between construction and its immediate resources. Timber has entirely disappeared as a building material in cities like Singapore, and it has been marginalized to supplying marine industries with pallets and planks for containers and ships, and as a matter of course to carpentry for furniture, timber decking, and exterior out-fittings.

Figure 1. Axonometric illustration of a Karo Batak house, on the Indonesian island of Sumatra. Illustration: Lai Chee Kien, edited by Ruven Wiegert

47

Rethinking Wood

Forestry and wood manufacturing in tropical Asia-Pacific More than half the world’s population, that is around 4.2 billion people, live in the regions across South, Southeast, and East Asia. This coincides partly with the second largest area in the world by forest coverage and a diverse Eco-system, particularly in Southeast Asia. Here, urban agglomerations have been expanding with tremendous speed and they are predicted to continue their growth in the coming decades. This brings along a huge demand on material resources and causes an immense ecological impact with an alarming rate of deforestation. Several sources claim that forests in some of the regions here could completely vanish before 2050 if the current rate continues. The logging industry has insufficient mechanisms in place for a sustainable forestry, to a large extent activities are even being carried out illegally. The main cause of logging on the Malaysian Peninsula and the Indonesian archipelago is to make place for the palm industries. Over eighty percent of the global palm oil production is sourced here and accounts for the destruction of its bio-diversity. The rapid urbanization in Southeast Asia since the 1950s and 1960s has seen a similar, but there has been a much faster decline and eventual disappearance of timber in the context of the built environment than Western countries have experienced. Like elsewhere, the advancement of the modern construction materials concrete and steel proved to be more suitable due to their efficient production processes and consistent properties. Highly developed cities like Singapore have a cement consumption far above average global rates and pose a challenge to the material supply. Singapore consumes 1,035 kg of cement per capita, not as high as China’s annual cement consumption of around 1,580 kg per capita, but twice as much as the global consumption and around 4 times as much as the USA (Davidson 2014). Most recently, after observation of the success of newer technologies such as Glue Laminated Timber (Glulam) and Cross Laminated Timber (CLT), timber has gained considerable interest in the region. Stakeholders have started recognizing more predictable material properties when studied carefully, potentially lower ecological impact, high degrees of manufacturing and pre-finishing, and lower weight. However, due to building certifications, elements made from European softwoods (and less frequently from Australian sources) are shipped across half the globe for a handful of newly constructed timber buildings.

Timber construction in Southeast Asia Timber industries have a long history in Southeast Asia. Just like in many other regions around the world, wood was once the predominant material used in construction, especially since it has been available as an abundant resource. The preservation of knowledge in wood manufacturing in general is two-fold. On the one hand, tropical 48

hardwoods are currently not as well documented and tested, but they would offer huge opportunities for applications like Mass Engineered Timber (MET): typical species have almost double the density of common timber (650–950kg/m3), but correspondingly almost double the load bearing capacity. For the design studio, we assumed that this could lead to more slender and differently articulated wood construction systems than current technologies. A number of initiatives have started redeveloping construction with locally sourced wood in Malaysia and Indonesia. Currently there is only one company in South Malaysia undertaking serious efforts to get certifications under European codes for their glue-laminated (Glulam) building components. Several projects have already gained wider recognition, such as the construction of a Glulam structure for the World Expo Milan 2015 in collaboration with the Malaysian Timber Industry Board and an Italian industry partner. There is a large diversity in local wood species, and tropical hardwood is just beginning to be researched more as a construction material. The Malaysian Timber Council lists the properties of around 50 “popular” species. The main difference of tropical hardwood to the European and American wood species is its density with up to 1,200kg/m3. The woods that have been examined and are deemed suitable for construction purposes are Keruing (Dipterocarpus spp.), Kempas (Koompassia malaccensis), and Merpauh (Swintonia spp.), which are all classified as medium hardwood and have densities of around 650 to 950kg/m3. On the other hand, there is still a well-conserved culture of vernacular building, both in traditional methods and in combination with modern materials. Particularly the large range of different roof types, but also certain functional details such as mobile facade and roof elements provide an abundant source of inspiration to tropical building. To take up the challenge of revisiting timber construction, a series of design studios at Singapore University of Technology and Design focused on wood as a primary material for the design of micro-towers. The studio investigated alternative construction methods, inspired by the idea of re-inventing timber construction and explored wood as a composite material and in hybrid construction systems to envision new concepts of vertical construction. The aim was to unlock the enormous potential of wood for future applications by recording a material and craft with a long history, re-conceptualizing it and projecting the findings into entirely new concepts.

Physical experiments as design constraints The experiments were focused on the tectonic articulation and formal variation of wood joinery and its assembly processes. Material experimentation initiates the design concepts, setting constraints to the formal exploration. Wood formed the basis of initially abstract geometric studies, and it was further tested in combination with other, 49

Rethinking Wood

primarily textile materials to form composites and hybrid structural systems. Vernacular building techniques in Malaysia and Indonesia became an important source of inspiration to envision new concepts of vertical construction. Examples like the Sumatra Long Houses with lightweight roofs and the tying of timber joints with soft Rattan wood on other Indonesian Islands resemble a fabric of woven wood structure that is treated like textiles as opposed to the rigid connections of post-and-beam structures. Tropical buildings have developed within a very distinct tradition (Figure 1), the Western and modernist concepts of the box as predominant geometry and completely sealed walls are unfamiliar concepts where the climate doesn’t require an absolute and limiting barrier (Tay 1997).

Digital timber topologies The studio focuses on design experiments in physical and computational models. It emphasizes fabrication experiments and the transition of data between the two media by parameterizing constraints from physical models in Rhino 3D modeling software and the Grasshopper plugin. The attention shifts from typologies to topologies (Schumacher 2008)—away from highly specific and mono-functional types (typology) towards topological components that can be parametrically varied and articulated to different conditions. Expressing the structural systems and the construction design of material joints leads to a different understanding of tectonic accentuation in timber architecture. To a certain extent the studio further aimed to embed generative tools, following the notion of the “second digital shift in architecture”: designs are no longer compressed into two-dimensional and simplified line drawings, but instead utilize data representations to bring complex assembly logic to the foreground (Carpo 2017). The studies are framed by predecessors in contemporary research and design. The emphasis on physical experiments and the translation of material properties into design have been demonstrated in numerous full-scale pavilions at architecture schools such as the ICD Stuttgart (Fleischmann et al. 2012). The studio initiates its research with similar strategies of material experiments but tries to anticipate large-scale applications of building components. In this fashion it operates on a smaller than full-scale and with pavilion type projects, although it draws inspiration from their formal strategies. In terms of production, the studio was less stringent on the application of digital fabrication technologies, but used it merely as means of quick design iterations and learning of the actual physical behavior of wood. Hence it did not restrict itself to high precision or the large variation of potential machining processes (Willmann et al. 2016). Similarities here can be seen in the type of base elements, such as plywood sheets on the one hand, and natural timber sections on the other. A more direct reference is the Bartlett School of Architecture’s Research Cluster 4 with its strategies of assembling dis50

crete elements in intricate configurations (Retsin et al. 2017). The aspect that distinguishes the studio mostly from the academic context is the consideration of tropical climate in its design formations.

Formal concepts In the initial design phase, various designs were explored as abstract geometric formations with no intention for any particular functions in mind, but aiming to establish tectonic systems. Physical prototypes were built with veneer, plywood sheets, and timber planks. These experiments started exploring linear load bearing elements (post and beam), surface elements (walls and slabs), and correlating skins. Eventually, the studio produced twelve design studies. They were categorized according to geometric parameters of their basic elements and the topologies of their joints: linear (L), planar (P), curvilinear (C) elements, orthogonal (O), 3-dimensional (3), and transitional (T) connections. Additionally, material combinations of wood with fibers or fabrics (F) were indicated. CTJ—curvilinear elements, fibrous/textile joints This concept is exemplary for the phase of material explorations. Kerfing was used to strategically soften the wood with different cutting patterns, bending it and building connections as transitional joints. The wood elements lose their tension capacity but retain some active bending behavior and most of their compression resistance. In order to regain tensile strength, textile fibers are woven around the bent nodes following the incisions in the surface to guide the fibers and to tie several pieces together (1 + 2). The first test started with a simple cross-wrapping pattern around the kerfing slots and went further into exploring different geometries of individual elements such as wedging, different bending angles, and double curvatures, and continued by experimenting with the connection of several elements into bifurcations and cross-shaped joints. As such, the second phase of geometric explorations started building a catalog of shapes that would turn into the grammar for the building structures. LPT—linear and planar elements, transitional joints In the second phase all design concepts were categorized by evaluating both the formal as well as functional capacities to further work on architectural elements. These elements were adapted to various scales and functions: they were parametrically varied in order to meet different criteria of an architectural program. The catalogs included structural components, roofs, walls, perforated envelopes, circulation systems etc. The exemplary project for this stage, LPT dealt with the lamination of the thinnest possible basic wood component: it explored transitional joints built from veneer sheets. Whereas this first mimicked the lamination process that is either applied in Glue Laminated 51

Rethinking Wood

Figure 2. Lamination of veneers into tectonic formations that enable gradual transitions from vertical to horizontal, and from linear to planar. Illustrations: Nurul Marsya Binte Mohd Shahruddin

Figure 3. The P3A approach combines planar elements with 3-dimensional assembly methods.

52

elements of Laminated Veneer Lumber, it led to a better intuitive understanding of the possibilities beyond the known conventions. The strategy was studied to build transitional nodal elements that could either integrate both linear and planar vertical and horizontal parts into one, or even gradually change from a linear into a planar element (from column to slab and from wall to beam). The actual joining point between elements is shifted from the corner towards the straight sections of elements, which liberates the joint itself from geometric constraints and enables continuous translations of forces in different directions (Figure 2). Micro-Towers Eventually all experiments were translated into architectural programs. Small high-rise buildings of up to ten stories were developed into case studies, with footprints of around six to eight by ten to twelve meters. These Micro-Towers were located behind a row of traditional shop houses in Singapore, allowing the participants to explore the capacity of their formal systems to connect to an urban context. According to the challenge, the systems were further refined to articulate load bearing, circulation, envelope, and partition systems. LOJ—linear elements, orthogonal joints The tower explored formal freedom and complexity with orthogonal joining techniques. The main structural system was based on rectangular frames, but a gradual shift from the main structural system to a secondary layer of frames was introduced to establish a porous envelope, define spaces, and regulate the climate. The architectural program offers a range of spatial configurations and speculates on a differentiation between human accessible spaces, and robot accessible spaces: from small cavities that are limited to the inhabitation by machines and infrastructure, to large spaces for human activities. L3J—linear elements, 3-dimensional joints Inspired by predecessors such as Konrad Wachsmann’s three-dimensional connectors, linear elements were joined in various angles to explore laterally stable structures in this tower proposal. Angles in the joints are varied according to ceiling heights, and individual members have continuously decreasing sectional dimensions from bottom to the top of the tower. COJ—curvilinear elements, orthogonal joints This design was an attempt to directly translate force flow patterns into the articulation of the structural system. The concept started with load simulations on a simplified structure, with vertical and lateral loads applied. The structural software plugin Millipede was used to visualize internal force flows. After this, a rationalization of the sys53

Rethinking Wood

tem was carried out, establishing a hierarchy within the individual members and hence translating the formal concept into a meaningful organizational logic. CCJ—curvilinear elements, curved joints This tower departed from similar initial explorations as the above described “CTJ—Curvilinear elements, fibrous/textile joints.” The curvilinearity in the structural system was also confined to the joints to achieve smooth force transitions throughout nodes. Overall, three systems correlate to achieve a stable structure: firstly, linear vertical columns that are curved at intersections. Secondly, a diagonal grid of beams for the floor slabs, and eventually a tertiary system that establishes the building envelope. P3A—planar elements, 3-dimensional assembly Here, planar elements such as CLT panels were the focus of the conceptual design and were investigated to establish a three-dimensional structure with lateral stability. The eventual hexagon shaped arrangement leads to a structurally strong formation resembling natural cell structures. The joints take advantage of the differently angled seg-

Figure 4. Schematic design for a timber tower based on the P3A approach. Illustration: Benjamin Yong Zhen Hui

54

ment and explore friction-only interlocking mechanisms. The heights of cells are varied according to functional demands and range from infrastructural to inhabitable spaces (Figures 3 and 4). STJ—sheet elements, textile joints The unique character of this tower resulted from its reference to ultra-lightweight structures that were originally conceived for aviation components: a combination of grid structures made of thin sheets that are combined with textile membranes as tension elements and skin layers. CLT panels are conceived as the primary structural elements, with the lightweight wood-textile composite elements as a secondary structure and enclosing system.

Future of tropical timber construction Since wood has not been seriously considered in the construction of tropical cities, the seminar achieved a differentiated recording of the material and accompanying craft to get a sense of its long history. The combination of physical and computational experiments led to a re-conceptualization in varied formal articulations. The attempt to go beyond technologies like Glulam and CLT makes it part of a more critical design discourse that attempts to obtain added conceptual and aesthetic values from the examination and designing with timber. The design studio can hence contribute to achieving a fresh view and spur the discussion on timber construction in a tropical context. This will become exciting, when tropical wood species are further understood, and knowledge starts coinciding with current technologies and vernacular concepts that suit hot and wet climates. There are a great deal of research opportunities for timber construction with tropical hardwoods and for more specific engineering and design for tropical climates. Acknowledgements The design studio was taught by Michael Budig in 2017 at Singapore University of Technology and Design (SUTD), with the assistance of Anna Toh Hui Ping and Zubin Khabazi (design computation). Student projects shown were developed by Benjamin Yong Zhen Hui, Bryan Lim Wei Guo, Caroline, Christopher Michael Wicks, Christyasto Priyonggo Pambudi, Endy Fitri Bin Saifuldin, Goh Wei Hern, Nicole Soh Hui Min, Nurul Marsya Binte Mohd Shahruddin, Pauline Siew Jiting, Tay Jing Zhi, Yue Hui Ying. I am also grateful for the insights on vernacular tropical architecture by my colleague at SUTD Lai Chee Kien, and to the Singapore Timber Association for hosting our field research.

55

Rethinking Wood

56

Joinery Culture The tools and techniques for joining separate pieces of wood to larger compounds are deeply embedded in specific cultural contexts, adapted to local conditions, and available materials. Whereas the foremost function of the joint is to resolve the intersecting forces present in the specific situation, the way in which these intersections are expressed greatly contribute to the architectural language of a timber building. Joints act as concentration points of the physical as well as conceptual dimensions, the motivations and priorities that have driven the creative process.

57

Rethinking Wood

The traditional spectrum of wooden joints consists of many variations of what, in principle, are simple concepts that are mostly based on the traditional uses of hand tools and regionally available wood types and dimensions. A few basic genotypes, such as the dovetail, used for joining two flat members together at right angles, the dowelled joint, in which dowelling is employed to impart mechanical strength; and the mortise and tenon, used to join a horizontal member and a vertical member of a frame, belong to the most common joints, whose phenotypes can be found in many different regions of the world. Climate, availability of raw materials, as well as conditions and constraints of individual societies, such as those of East Asia, North America, Central, and North Europe have shaped different versions of these basic concepts, leading to sophisticated constructive and architectural languages. Whereas in the eastern tradition the master builder, responsible for the choice of materials as well as for construction and design of the appropriate joint, was maintained, in western cultures, a professional distinction developed between the design of a wooden building and its execution. This distinction continues today. Here the exposure to industrial methods and power tools in the nineteenth and twentieth centuries caused traditional joinery techniques to be marginalized. In today’s globalized wood processing industries, economy and utility are the major factors in the development of new timber joints. Connectors from other materials, such as screws and metal splice plates, etc. have replaced traditional methods. Combined with adhesives, they are in many cases stronger than the wooden elements that they join. These advanced technologies have expanded the horizon of possibilities for architects and engineers, but have led to complex composite materials, which are difficult to recycle and reuse. In this part of Rethinking Wood, the presented research projects address the adaptation of traditional joinery knowledge to modern fabrication processes. Whereas the need for a carpenter who is skilled in manual joint carving is diminishing, there is a need for experts in converting traditional timber knowledge and joint configurations in modern practice. In addition, new possibilities of robotic fabrication call for new joint typologies which take into account the kinematic logic of the multi-axial production process.1 Specific skills are required to fully harness the potentials of the higher degrees of freedom available in these new methods. Although wood building became more complex in recent years, it can be argued that principles that have developed over centuries can still be applied today. In their contribution, Pekka Heikkinen and

See also Philipp Eversmann’s contribution on robotic fabrication of glued joints in Part Three, “Digital Processes” 1.

58

Philip Tidwell reassess the use of traditional techniques of joinery in a series of experiments. Their rigorous experimentation process leads to solutions that are singular, and vary widely in function, character, and scale, but share an important set of strategies for prefabrication and dismantling of contemporary wood constructions. As an anisotropic material, the behavior of wood varies in different directions, and material properties, such as the direction of the wood grain, have to be considered early in the design process. Another perspective on how to revive traditional principles of joining wood is given by Olga Popovic Larsen. Reciprocal frames, i.e. assemblies of mutually supporting beams, can create lightweight structures by combining short members in reciprocal configurations that are space forming and architecturally distinct. Here, only two beams are connected at a time, which makes the connections simple. Specially designed and fabricated metal connections needed when joining several timber beams into a single joint can be avoided. The applied research of Hans Drexler is driven by the imperative for an adaptive use of building elements in reconfigurable plan layouts for low-cost housing. Drexler presents an approach for distinct architectural spaces from wood-only elements, in which both constructive elements and fasteners could be made of timber, so the relative strength of all components is uniform. Here, the design of the joint not only leads to a flexible construction system, it also addresses long-term strategies of repair and re-use. Individual parts can be removed and replaced so that the working life of the building can be extended. Engineered wood products, such as glued and cross-laminated timber, can overcome the deficiencies of wooden joints and allow for an industrial use of recycled timber. Due to their positive properties, the use of construction timber in the building sector has become competitive with other building materials. Today, only the use of adhesives enables large span structures and multi-story timber buildings. Gerhard Fink and Robert Jockwer look at new possibilities for timber structures resulting from application of adhesives in joinery. Conceptually and practically, the design of joints should again become a central concern for architects, not only in wood, but for all materials. Many traditional joints are not used because they are not adapted or are inconvenient for modern manufacturing and construction technologies. However, there is a great deal of accumulated knowledge and practice in traditional timber joints. This construction heritage can be harnessed and adjusted for contemporary use.

59

Rethinking Wood

60

Designing Through Experimentation: Timber Joints at the Aalto University Wood Program Pekka Heikkinen and Philip Tidwell

From vernacular traditions to current experiments in computational design, construction in wood usually involves large numbers of building components that have to be joined in one way or another. Regardless of age, most sophisticated wooden structures resolve this challenge with clever solutions that allow for quick and reliable assembly on site. The best examples not only negotiate the demands of construction, but also address long-term imperatives of repair and reuse. In these cases, carefully designed joints enable individual parts to be removed and replaced so that the working life of the building can be extended. For centuries, wooden buildings for industry, agriculture, worship, and living have been designed so that they can be broken down into parts and then moved or recycled in novel ways. However, today many structures seem to take little advantage of the inherent lightness, flexibility, and impermanence of timber construction. Although the scale and complexity of wood building have increased dramatically in recent years, principles that have developed over centuries can still be applied in contemporary practice. Already European legislation promises to place greater importance on the dis-assembly, re-use, and recycling of building components in the future. If the goals of circular construction are to be meaningfully advanced, they will necessitate careful design and detailing of connections all the way from the factory to the building site. Conceptually and practically, the design of joints should again become a central concern for architects, not only in wood, but indeed for all materials. The building projects carried out over the past two decades at Aalto University vary widely in function, character, and scale, but they share an important set of principles: each structure has been designed to be quickly, safely, and reliably assembled on site then disassembled and rebuilt or recycled. The practical challenges of this mode of construction have led to an emphasis on prefabrication as well as careful attention to the ways that wooden components can be fixed

Interior of the Tree House pavilion (2007). Photo: Sarianna Salminen 61

Rethinking Wood

and joined. Here we examine the role these constraints have played in the work of the Wood Program and consider some of the solutions illustrated by specific projects designed and built by students at the university. As a point of departure, the work of the Wood Program does not seek to find universal solutions that could be applied universally. Instead, the constraints imposed by each specific building task are taken up as experiments that can be used to study and explore various materials and methods of building. Every new project defines a unique set of conditions and challenges for the group to consider. As a result, the solutions that emerge from every project are singular, but often raise some of the same principle questions. First and foremost, it is important to remember that compared with other structural materials, wood is relatively soft and delicate. In many historical examples, both building elements and fasteners could be made of timber, so the relative strength of all components was more or less uniform. Today, the adhesives, screws, and metal connectors used in construction are many times stronger than the wooden elements they join. These advanced technologies have expanded the horizon of possibilities for architects and engineers (especially for structures that are dissembled and assembled again), but careful attention must be given to their position and installation to be sure that load is distributed over a large enough area. In most cases, fasteners should be small in size but large in number relative to the wooden elements they join. Although larger hardware can be used, such components tend to concentrate force in the wooden elements of the joint, which are comparatively weak and may become overstressed. In addition to fasteners, grain direction is of principal importance in any timber joint and should be taken into account early in the design process. As an anisotropic material, the behavior of wood varies in different directions, and the softwood species typically used in construction have dramatically different properties along the length of their grain than they do across it. Tensile strength, for example, is about ten to twenty times greater along the grain than perpendicular to it, while shear strength is dramatically less in both directions (only about ten percent of tensile strength measured along the direction of the grain). To overcome these differences, special attention should be paid to the application of force, the position of fasteners, and the dimensions of timber components. The Säie pavilion (page 74), for example, uses slender pine lathes bolted to high-density birch plywood to distribute loads over a wider area and to avoid concentrations of forces at the end of the thin profiles, where they are especially vulnerable to splitting. With multiple veneers layered in opposing directions, the dense plywood elements do not have single grain direction that would create a weak axis and

62

so they can more easily absorb the various forces passing through the joint. In projects with more complex assemblies and numerous layers, such as the Kokoon modular living units (page 78), a clear hierarchy of joints is advisable. In such assemblies, many of the fixings are not visible and will have only minimal influence on architectural character, but technical issues such as insulation and air-tightness play an important role. Only a handful of joints in this type of construction will be taken apart and re-assembled during the life of the building, so attention should be devoted to the most critical connections that will be opened or serviced over the years. In the case of the Kokoon project, the most important solutions are those that bind the modular units together easily and tightly. Effective design of these connections allows other parts of the building to be realized through relatively conventional means, with less concern for repair or replacement. Finally, of all the factors to be considered in any joint design, one of the most important is the effort that must be devoted to producing and assembling each connection relative to the number of times that it occurs. The more joints in any building, the quicker and more straightforward their assembly should be. The Tree House (page 60) for instance, repeats the same connection more than a thousand times, so the joint is fixed simply with a pneumatic nail gun from three sides. However, the geometry and sequence of assembly ensures that all nails are eventually concealed by the slender wooden profiles. In projects with fewer connections, more complex solutions might be used to emphasize the aesthetic qualities of the material or geometry.

Figure 1. (next page) The HDW pavilion installed at the Ateneum Art Museum. Photo: Jussi Tiainen

63

Rethinking Wood

64

65

Rethinking Wood

Figure 2. Interior of the HDW pavilion. Photo: Jussi Tiainen

Figure 3. Section detail of the connection between pyramidal elements. Drawing: Antti Lehto and the HDW team, edited by Philip Tidwell

66

HDW Pavilion (2006) The Helsinki Design Week Pavilion is an information pavilion developed through a fully digital workflow that links computational design and analysis to production and assembly. The geometry of the pavilion was developed using a series of 3D models, and finite element analysis was used to ensure that the surface area and curvature of the vault would be sufficient to perform as a stressed skin structure. The modest scale of the structure allowed it to be transported as a single piece, so disassembly of the joints was not a primary objective. Instead, the most significant challenge in the project centered on the aesthetic and structural concerns of joining two materials with very different properties. At either end of the structure, plywood arches stiffen the building and join to two plywood edge beams running parallel to the ground. Between these elements, 135 truncated pyramids of plywood and glass are fixed together to form the stressed skin. The pyramidal units are linked together by a fiberglass reinforcing strip, which is set into a grove at the end of the plywood. Lateral stability is provided by a series of carbon-fiber cables that are laid along the diagonal seams of the structure and then held in tension to the bottom beams and end arches. The fiberglass reinforcement and carbon-fiber cables are bound to the structure and visually concealed by epoxy that fills a channel at the perimeter of each glass triangle. The dimensions of all glass and plywood components are unique, but each is constructed from a similar set of three, 18-mm birch plywood trapezoids and one triangular piece of laminated glass made from two, 4-mm sheets. The plywood elements are joined together on their short edges using glue and a pair of dowels to create a stiff triangular frame before the laminated glass triangle is fixed to their long edge. The glass is secured to the plywood by acrylic foam tape and then reinforced with another fiberglass strip which is held in the same groove at the end of the plywood. In this way, the plywood frames stiffen the edge of the glass triangles, while also concealing the structural connections between the pyramidal units. A 3D model and specially tailored script were used to program two different NC workstations to precisely cut the plywood and glass components with minimal tolerance. The elements were then assembled by hand and the building was assembled in a workshop before being transported in one piece to its site at the Ateneum Art Museum in central Helsinki. Later, the project was moved to its permanent home at the UPM-Kymmene plywood mill in southern Finland.

67

Rethinking Wood

Figure 1. Shelter assembled and ready for use. Photo: Anne Kinnunen

Figure 2. Detail of the panels, dowel connection, gasket, and cargo straps. Photos: Rebecca Littman-Smith

Figure 3. Assembly sequence of the shelter. Drawing: Liina team

68

Liina Transitional Shelter (2010) The Liina Transitional Shelter was designed as a temporary building that can be shipped flat and assembled quickly for use as an emergency shelter following natural or humanitarian disasters. Because such events often destroy both infrastructure and buildings, a primary ambition for the project was that it be put together on site without the need for sophisticated tools, lifting equipment, or electricity. This simple, but challenging set of constraints led to a design that uses standard nylon cargo straps (liina in Finnish) to assemble, stabilize, and seal the building in a manner similar to a wooden barrel. The design consists of hollow core panels made from frames of 50-mm LVL (laminated veneer lumber) and outer surfaces of 9-mm birch plywood. The panels are filled with cellulose fiber to improve their insulating capacity and manufactured at the width of a standard half-sheet of plywood (600 mm). These dimensions ensure that all building components can be manufactured with minimal waste, carried easily by two people, and packed efficiently on standard pallets for storage and shipping. Assembly of the building begins by laying out panels to form a five-sided frame and fitting them together using the dowel connections along their edges. The frame is encircled by a cargo strap, which is tightened with a ratchet to bind the panels together. As the strap compresses the frame, 16-mm wooden dowels in each panel edge are driven into the holes of the adjoining panel and a gasket along the edge of each joint is compressed. The wooden dowels guide the panels into position and resist lateral forces, while the rubber gaskets form an airtight seal along the outer seam. Once the frame is fully tightened, it is tilted up and joined to another frame to form a series of parallel sections that are then fixed together using the same joint and strap technique. A sleeping loft at the back of the building and walls at either end provide lateral stability for the structure, while a waterproof tarpaulin provides protection from rain, snow, and UV damage. The structural solutions in this system are relatively simple, but not commonly used in construction, so many tests were carried out to guarantee the safety, durability, and practicality of the design. Numerous prototypes of the strap and dowel connection were built and tested in the engineering laboratory to be sure that the capacity of each joint was sufficient. Similarly, a full-scale frame was tested using a shake table to study the behavior of the building during a seismic event. Following these tests, minor revisions to the panel profiles and the size of the connecting dowels were implemented in the final design.

69

Rethinking Wood

Figure 1. Lifting the assembled canopy during construction. Photo: Anne Kinnunen

Figure 2. Exploded axonometry of a roof connection. Drawing: Markus Heinonen and the WDC Pavilion team, edited by Philip Tidwell

70

WDC Pavilion (2012) The WDC Pavilion was built in central Helsinki as a space for gathering, dining, performing, and a variety of other activities during the World Design Capital year in Helsinki. The building consists of a large rectangular roof canopy that extends over a terrace and two enclosed boxes that contain a kitchen and theater. The canopy structure covers more than 500 square meters with a grid of 172 equilateral triangles that are supported by eight hexagonal columns. Within these triangles, a smaller series of beams further subdivides the grid to provide support for the translucent polycarbonate covering. While the grid is mostly uniform, the eight supporting columns are positioned irregularly, so loads and spans vary throughout the structure. To manage this, the perimeter and the transverse diagonals in one direction are treated as continuous elements with simple spans in one direction. Elements in the other direction are treated as shorter spans and hung from consoles on either side of the main beams. All roof beams are fabricated as uniform box-sections of 550 mm x 81 mm with an inner core of structurally graded spruce timber (45 mm) sandwiched between outer surfaces of 18-mm spruce plywood (with a surface veneer of birch). The plates of the supporting columns were fabricated similarly, with overlaps in the plywood layers to provide structural continuity on the outer surfaces. This design allowed the roof and column components to be manufactured as flat panels in a factory and shipped to the site for assembly. The components of the roof were joined together at ground level then lifted as a single element so that the columns could be assembled and installed in a single day. This system reduced the amount of scaffolding needed for the work and eliminated the need for temporary structures to support the roof during assembly. It also improved safety on the site by reducing the height of the work and danger of falls. In conventional construction, the six-point intersections of these elements might be managed using welded steel or extruded aluminum, but here each of the 108 nodal connections is fixed by one upper and one lower plywood star which are cut from 24-mm birch plywood and screwed to the middle layer of each component. Screws compress the elements together, while 16-mm wooden pegs in the lateral arms of each star provide additional shear strength to hold the shorter beams in place. The design eliminates any bolt connections and metal hardware that might disrupt the visual character of the pavilion.

Figure 3. (next page) Interior of the WDC pavilion showing the underside of the roof canopy. Photo: Tuomas Uusheimo

71

Rethinking Wood

72

73

Rethinking Wood

Figure 1. Interior of the pavilion showing the underside of the roof canopy. Photo: Kimmo Räisänen

Figure 2. Roof joint. Photo: Kimmo Räisänen

Figure 3. Exploded axonometry of the plywood connections. Drawing: Hiroko Mori, Laura Zubillaga, and the Säie team, edited by Philip Tidwell

74

Säie Pavilion (2015) The Säie Pavilion was commissioned as an event space for an exhibition on garden design at the Museum of Finnish Architecture in 2015. The Finnish word säie refers to a strand or fiber, which is evoked by the bending and twisting of slender wood elements. The curvilinear forms are meant to reference historical examples of garden pavilions and gazebos, which were often decorated with elaborate woodcarvings or ornamental iron work in vegetative themes. The structural behavior of the building is complex in that all elements are actively bent into their final position and balance the force of adjoining members. In order to study the behavior of this system, the geometry of the pavilion was studied with numerous physical models and full-size prototypes in order to determine the optimal form, dimensions, and system of bending to be used. Only after the final configuration had been developed could the system of joints be completely defined and resolved. To achieve such dramatic bends and twists, the project makes use of high-quality pine lathes measuring 12 mm × 94 mm. The knotless boards bend easily, but their thin cross section makes them difficult to fix with screws and leaves them vulnerable to splitting at their ends. To address these challenges, the joints of the structure are fabricated from birch plywood, which is cut using a five-axis router to the required geometry then through-bolted to the end of each pine lamella. Since the multiple layers of veneer in the plywood are stacked in opposing directions, they do not produce a single grain direction that would create a weak axis. This enables the birch components to resist the various forces passing through the joint that might easily split the pine lamellas. The simple connection not only guides force away from the fragile ends of the pine lathes, it also provides a necessary degree of tolerance during site assembly of the complex structure. The precise configuration of the final geometry is ensured by the accurate cutting of all elements using the NC router. In addition to structural performance, the solution also avoids the practical challenge of machining the long and flexible pine elements, which are difficult to accurately position and cut in most NC machinery. By reducing the scale of the components to be machined, the joint could be easily produced (and re-produced if necessary) using standard equipment that is widely available.

Figure 4. (next page) The Säie pavilion installed between the Museum of Finnish Architecture and the Design Museum. Photo: Kimmo Räisänen

75

Rethinking Wood

76

77

Rethinking Wood

Figure 1. Exterior view of three prototype modules stacked together. Photo: Tuomas Uusheimo

Figure 2. Interior view of a prototype module. Photo: Anne Kinnunen

Figure 3. Section and elevation detail of the module connection and lifting point. Drawing: Léa Pfister and the Kokoon team

78

Kokoon Modular Living Units (2016) Kokoon is a modular system designed as a temporary living solution that can be easily transported and stacked in different configurations to fit various sites in urban and suburban contexts. The units are constructed using large 45-mm LVL (laminated veneer lumber) wall panels which provide a rigid structure and minimal wall thickness in order to optimize the small dimensions of the design. With a footprint of 3.3 m × 4.9 m, each module fits easily on a standard trailer, but provides enough space for one or two people to live. Most of the construction details are relatively simple and were carried out using screw or nail solutions typical for rough carpentry. However, to stack and transport the modules safely, quickly, and reliably, extra challenges were solved through detailed study. Each prefabricated box has to be rigid enough for lifting and transport, while has to also carry the load of two occupied units when stacked. To manage this, the LVL walls act as shear panels that provide strength and stiffness, while also supporting the large cantilever. A glulam column links the assembled modules together at each corner to transfer loads to the ground. To ensure that the units are air and watertight when they are stacked, a ring of felt and linen fiber is fixed to the perimeter of each roof. As each unit is stacked on the next, the fiber gasket becomes compressed by the weight of the module and forms a tight seal. As the seal is fully compressed, the glulam columns come into contact and ensure that all loads are transferred through the columns rather than the fiber seal. Finally, a cover plate is fixed on site using two, 24-mm bolts to bind the columns together. The cover plate is attached to a second steel plate which is fixed permanently to each column at the factory. During disassembly, the cover plates are removed and a rotating lifting point is bolted in their place to lift the units and secure them during transport. As each unit is lifted, its entire weight of 5,000 kg is suspended from these four corners such that all loads on the structure are effectively inverted. In practice, the system has proven remarkably effective as it has been possible to disassemble, move, and reassemble the three prototype units in a matter of hours.

Figure 4. (next page) Stacking the units on one another using a crane. Photo: Juho Haavisto

79

Rethinking Wood

80

81

Rethinking Wood

Figure 1. The pavilion installed at the Suvilahti area of Helsinki. Photo: Ransu Helenius

Figure 2. Exploded axonometry of the upper and lower wall connections. Drawing: Ornella Angeli, based on a 3D model by Antti Hannula, Antti Rantamäki and the Aika-Lava team

82

Aika-Lava (2017) The Aika-Lava Event Stage was commissioned as a space for lectures, exhibitions, and demonstrations at various events during the centennial year of Finnish independence. As a tribute to the nation’s hundredth anniversary, the structure stands on one hundred supports and creates a long rectangular volume 34-m long, 7-m wide and about 5.5-m high. The character of the project comes mainly from the exposed structure, which has a subtle pattern of curvature that supports and stiffens the building. The stage is composed of fifty portal frames that stand parallel to one another and are linked together through bending to produce a rigid lattice. Spacing between the frames is set by the floor structure and the uppermost beams of the roof, which are notched at 600-mm intervals. As with the Säie Pavilion, the system uses a method of active bending to force straight elements into a curved wishbone shape that produces an extremely rigid wall and eliminates any need for lateral bracing. However, while the earlier pavilion uses thin pine, the scale of this event stage demands thicker elements, which are manufactured from large LVL (laminated veneer lumber) panels. The size of the project and the large number of connections required that the bending technique, thickness of the members, and geometry be studied carefully to achieve the desired curvature as efficiently as possible. Numerous physical models and full-size prototypes were used to test these details and special attention was given to the design and dimensions of the fastening hardware so that the simple bolt connections could be executed quickly and easily on site. Components of the vertical wall measure 300 mm × 51 mm and vary in height from 4.9 m to 5.1 m. They are fixed together at top and bottom using specially designed 12-mm sex bolts which are hidden inside the wooden elements. The hardware is long enough to compress the wooden members without temporary clamping or bolting, which would add time and complexity to the construction process. To simplify the assembly further, the vertical LVL components were pre-fabricated in 3-m wide wall panels, so that the number of connections on site is kept to an absolute minimum. At the top of the wall, the vertical components are fixed with four bolts to a short spacer which serves to position the roof beam during construction. After the beam is set in place, it is fixed with another four bolts to create a rigid corner that provides lateral stability. At the bottom of the wall, this configuration is mirrored, but the short spacer is notched to serve as a guide during assembly and to tie the wall elements to the floor slab.

83

Rethinking Wood

Figure 3. The interior of the Aika-Lava pavilion. Photo: Jouni Harala 84

85

Rethinking Wood

Every building project is different and each one poses a new set of conditions and challenges, but one cannot avoid basic material properties or rules of physics. Wood lives, which is to say it continuously swells and shrinks in response to the moisture content of its immediate surroundings. This hygroscopic quality, together with the anisotropic behavior mentioned earlier, define the basic design parameters for any wooden joint. In practice, most common mistakes result from these fundamental material properties being misunderstood or ignored altogether. Rather than an inert substance on which form can be imposed, wood is a complex and heterogeneous material with numerous qualities that must be understood and considered. Dimensional changes and bending from moisture or stress can lead to problems during assembly and deformations in the final building, but such mistakes can be easily avoided by using appropriate tolerances and allowing for variation of individual components. At the same time, environmental conditions in the workshop and on the construction site should be accurately monitored and carefully measured to be sure that the work is carried out according to plan.

Acknowledgements: The research and projects included in this article have been carried out with numerous students, faculty and staff at Aalto University, as well as funding and material support from external partners. The authors would particularly like to note the contributions of Professor Hannu Hirsi and Project Manager Ransu Helenius, who have played a significant role in many of the projects discussed here.

Laboratory test of the plywood connectors used in the WDC Pavilion. Following the test, some screws were replaced with wooden dowels to improve the shear strength of the connection. Photo: Anne Kinnunen

86

HDW team: Antti Lehto, Teemu Seppänen, Antti Autio, Anna Bevz, Uula Kohonen, Sini Meskanen, Aleksi Niemeläinen, Anita Nummi, Eero Puurunen, Ilkka Salminen, Markus Wikar, Elina Voipio and Jussi Ziegler, Terhi Keski-Vinkka (structural design) Instructors: Pekka Heikkinen, Hannu Hirsi, Risto Huttunen, Pekka Pakkanen, Lauri Salokangas and Antti-Matti Siikala Liina team: Signe Äärset Loe, Miguel Castillo Lopez, Aleksandra Cherepnalkoski, Megan Groth, Vladimir Ilic, Rebecca Littman-Smith, Marta Marcos Marono, Maria Mor Pera, Chaeham Oh, Hugo Plagnol, Bernardo Richter, Einari Sutinen and Dmitry Troyanovsky; Kai Nordberg and Ulla Hakulinen (structural design) Instructors: Pekka Heikkinen, Hannu Hirsi, Matti Kuittinen, Pentti Raiski, and Cristina Santamaria Nogueira Project Management: Ransu Helenius and Mikko Merz WDC Pavilion team: Markus Heinonen, Pyry-Pekka Kantonen, Janne Kivelä, Wilhelmiina Kosonen and Inka Saini; Marko Hämäläinen (structural design) Instructors: Pekka Heikkinen, Hannu Hirsi, Risto Huttunen, Mikko Paakkanen, Karola Sahi, and Philip Tidwell Project Management: Ransu Helenius, Miko Merz Construction: Stara Construction Services Säie team: Giulia Archimede, Simen Bie Malde Claire Bouthegourd, Elisabeth Kofler, Kento Manabe, Javier Mera, Hiroko Mori, James Stanier, Ninni Westerholm and Laura Zubillaga; Joni Helminen (structural design) Instructors: Pekka Heikkinen, Hannu Hirsi, Philip Tidwell Project Management: Laura Zubillaga Kokoon team: Alexander Barstad, Akin Cakiroglu, Kristin Ekkerhaugen, Satoshi Iiyama, Nicklas Ivarsson, Stephanie Jazmines, Yuko Konse, Sini Koskinen, Toni Lahti, Maria León, Tomoyo Nakamura, Taeho Noh, Käbi Noodapera Ramel, Léa Pfister, Ivan Segato, Ignacio Traver Lafuente, Tanja Vallaster and Eduardo Wiegand Cruz Instructors: Pekka Heikkinen, Philip Tidwell, Willem van Bolderen and Antti Haikala (structural design) Aika-Lava team: Antti Hannula and Antti Rantamäki Instructors: Pekka Heikkinen, Ransu Helenius and Antti Haikala (structural design)

87

Rethinking Wood

Fig. 6 a-b: ReciPlyDome

88

Reciprocal Timber Structures and Joints Olga Popovic Larsen

Building with wood If we look at history, and more specifically built examples through history—wood stands out both as a material for creating very early, primitive human dwellings, but also, other building typologies of varied complexity and sophistication. The abundance and availability of the material and the ease of creating joints and connections without the need of complex building tools, throughout time has led to a richness of timber architecture worldwide. Trees vary greatly in species and types in the different climatic and ecosystem zones and as a result have very different qualities. Still, all the trees have in common that they provide us with a natural and renewable building material with excellent mechanical properties that are comparable to the strength and stiffness of steel and concrete. At present when sustainable aspects are at the forefront of current thinking, more than ever, wood is being considered for construction as it can contribute greatly to reducing the huge negative environmental impact of the building sector. In the context of promoting wood, this essay presents the opportunities that a special timber structural system—the Reciprocal Frame (RF)—offers. Reciprocal frames (RF) In essence, a reciprocal frame is load-bearing structure: An assembly of mutually supporting beams, where only two beams are connected at a time, making the connections simple and avoiding specially designed/fabricated metal connections needed when joining several timber beams into a single joint. As a result, the system allows for building with short beam members that can be connected in a multitude of ways, creating architecturally interesting, three-dimensional assemblies that are easy to construct, can be taken apart, re-configured, and re-assembled. Although RFs are not limited to the use of wood, timber and engineered timber products are most suited. The inherent qualities of wood such as the fiber structure, ease of working, natural renewable/recyclable properties etc. are suited to RFs. Thus, this text is limited to RFs made from wood. In addition to building with single RFs as self-standing structures, by combining the single units, gridshell-like structures can be formed.

Figure 1. The ReciPlyDome structure is based on polyhedral geometry, plywood beams in active bending and simple reciprocal connections. It is the result of a collaboration between KADK and VUB.

89

Rethinking Wood

Figure 2 a–d. RF single units with a) three, b) four c) six, and d) more beams.

90

They can be configured with combinations of identical or different RF single units. RFs are mainly used for load-bearing building roof structures, but can be also used as facade forming grids, both for temporary and permanent buildings. There are many opportunities, but also some challenges when building with RFs. The main advantages of using RFs are: • simplicity of the connections—as only two members are connected at a time, there is no need to design/fabricate special ball-joint type of connectors that are often expensive • the joints are reversible and as such allow for design for disassembly and reusability • the structure is built of relatively short members that are lightweight and easy to carry • all connections are identical—contributing to ease and speed of fabrication/construction • the grid RFs are more robust structural solutions due to the greater inherent redundancy in the structure—in gridshell-like RFs a local failure of one member and/or connection does not lead to an overall collapse • the system is versatile and distinct—offering great opportunities for different architectural expressions The main challenges or issues an RF design should address are: • the spatial nature of the structure requires three-dimensional geometrical definition • the design requires a high-level of precision, especially when dealing with the gridshell-like RFs as a small mistake in the design and/or fabrication can lead to great assembly difficulties • the complexity of design can require more design time. The geometric definition, structural modeling and precision of fabrication are no longer real challenges, as the current stage of development and availability of design and fabrication tools make it possible to design and fabricate complex, three-dimensional geometries. The design time needed can be seen as an investment as, RF members once designed and fabricated can potentially be re-used multiple times and in different structural configurations. The design time in essence is spent to design a system that can be re-erected many times and in different forms. It is clear that the advantages outweigh the challenges. Thus, it is not surprising that RFs are becoming a mainstream building structure choice that is being used more frequently and in different contexts and ways. The remaining part of the essay presents several built RFs examples in Japan and Denmark. The RFs presented are very different and many important and valuable lessons can be learnt from them.

91

Rethinking Wood

RFs in Japan Looking at the well-known examples of RFs by Kazahiro Ishii (Popovic Larsen 2008) as the Seiwa Bunraku Puppet Theater, built in the early 1990s, the different expression achieved with the auditorium RF structure and the RF in the exhibition space, already in this early RF example, clearly showed the versatility of the RF system and opportunities for achieving a completely different feel in the spaces. The heavy RF beams in the auditorium add to the feeling of oppression, whereas the tall slender columns supporting the slender-beam multiple RF structure roof in the auditorium creates a sense of lightness. This, together with the natural light entering through the large windowpanes of the exhibition hall, contributes to the feeling of hope and positivity. It is in the same building complex that the RF has been used in completely different ways and to a completely different effect. At the time it was designed, the Puppet Theater RF structures posed a challenge to design and build. At the time, this project was pushing the boundaries of what was possible in terms of complex geometry definition and modeling. Also, the craftspeople had to be highly skilled in order for the buildings to be constructed. A wrong cut of the three-dimensional joint meant a wasted timber beam. The geometry definition and three-dimensional structural modeling were on the limit of what computation tools of the time (early 1990s) could achieve. Yet, this design complexity created construction simplicity. All the beams of the RF in both the auditorium and the exhibition structure were identical. Once designed, every beam could be fabricated and inserted into the RF assembly. Another RF example, also in Japan, is the Rokko Mount Observatory built in 2010 by Hiroshi Sambuichi. It was built about twenty years after the puppet theater and it takes complexity to an entirely new

Figure 3a–b. The Puppet Theatre RF structures: a) auditorium b) exhibition hall. 92

level. The building is deeply connected to nature and uses wind, water, sun, and ice, or “moving materials”—as design architect Sambuichi refers to them, to create the building environmental control. In this case, the ice blocks formed in the outdoor shallow pools are stored in the highly insulated basement and used for cooling of the spaces during the summer. The RF creates an open lattice roof structure with the irregular hexagonal tessellations providing a surface for the water drops to freeze in the winter months. Together with ARUP, the engineering consultants, Sambuichi explored the density of the RF grid, so that it can form the correct distances between the members in order for the ice lattice to be formed. The beautiful open RF roof structure is only clad with the ice lattice spanning between RF timber units inserted into the steel RF loadbearing beams. When compared with Ishii’s RFs, the Rokko mount RF is clearly far more complex. The undulating open steel structure has a very complex geometry and the inset timber RFs are also irregular. The structural modeling and design could be performed by the highly sophisticated available design tools that had been developed in the twenty-year period after the construction of Ishii’s. The challenge in this project was the assembly of the structure. The structural RF members were all different, yet visually similar. If for example, by mistake during construction, a member was substituted by another, the mistake could only be revealed quite late in the process and made it impossible to complete the assembly process. To avoid this, the contractor and the engineer developed a construction sequence, marked all the members, and even built a trial construction with a 1:10 model of the structure. This ensured the smooth assembly of the actual structure. Ishii’s and Sambuichi’s RF structures are very different from one another—both in addition to being very distinct in their form and expression—they also present a cutting edge level of technical development of the time with regard to design and construction. Digital tools and methods at the height of what was possible at the time of each project were used.

Two innovative Danish RF projects Since the time of the RF buildings designed by Ishii and the Sambuichi Rokko Mount, project RFs have seen further development. Various, mainly exploratory, projects have investigated what RFs can offer, especially as different morphologies. This development continues. The number of constructed projects has not been extensive. The next section presents two built examples from Denmark. The first project in Denmark is an agricultural building on the small Danish island of Glænø completed in 2017, which was awarded the Architecture Sustainability award in 2018. The client’s requirement for the Glænø project was an optimized timber structure, using local wood (of relatively low quality) that would be easy to construct with93

Rethinking Wood

Figure 4. The Glænø stable.

Figure 5 a–b. The Glænø stable: a) open RF truss roof, and b) detail.

94

out machinery, could be disassembled and, that would be aesthetically pleasing. When building a stable, the usual approach is to use a prefabricated steel or wooden truss. This is a cheap, structurally efficient, but not a sustainable solution because the wood or steel is generally sourced from a long distance away. Also, these structures have no aesthetic sensibility and are purely utilitarian with crudely designed details. Furthermore, they are not designed for disassembly. The author was approached and the proposed stable design was inspired by RFs. By having only two beams connected in one joint, and not more, meant that the design and construction would be simplified. If more beams would come in one point, the connection would be far more complicated to achieve. A specially fabricated metal connection would be needed. Normally due to the geometry, RFs are most suited for roofs over a round or polygonal building in plan. However, as this was not acceptable because the dimensions of the Glænø stable were 16 x 32 meters in plan, forming a rectangular form. The initial idea was a RF arch, but this kind of structure used fifty percent more wood in comparison with the cheap prefabricated steel or timber truss. It also had to carry a secondary roof to form a thirty-degree roof, which added additional load. A number of RF-inspired structures were considered and many optimizations were carried out. The roof that was constructed was a RF-inspired timber truss, with simple connections that were reversible, with just one or two bolts through each connection. All the wood for the building was locally sourced and came from 800 meters away from the farm. It was cut to size by the farmer and left to dry for a year. After one year, all the timber RF trusses were assembled on the ground and lifted into position. Apart from a tractor, no other machinery was used to build the roof. In this project, great effort went into the timber optimization and design of the connections. Through many iterations, the wood quantity for the RF roof was minimized, so that even though the locally sourced wood with lower mechanical properties, the structure could be built to be slender and elegant. This project was about achieving a sustainable resource optimized RF structure that could be self-built, without the need for any heavy machinery. An important requirement was to achieve an elegant RF structure with details that are simple, easy to construct, can be dissembled, yet are also beautiful and contribute to the overall appearance of the structure. The project truly achieved the expectations regarding its simplicity, optimized structure and appearance and has created shelter for the cows, part of the farm’s ecological meat production. The building expression and quality of detailing are such that the building could have also been used as an open plan public building. This project is an 95

Rethinking Wood

Figure 6. 3D drawing of assembled ReciPlySkin.

96

important example case showing that wood, even if it not of highest quality, can be used to create structurally efficient and architecturally distinct sustainable buildings. This message is powerful, especially for Denmark, where sustainable qualities are currently very important, yet where wood is still not used as a mainstream material. It is believed that Danish wood is not of high enough quality, thus concrete and masonry are generally the choice of preference. The RF trusses in the Glænø project convey this important message. Another recent project is the innovative ReciPlyDome structure, developed during 2017 in a collaboration between this contribution’s author from KADK (Denmark), VUB (Belgium), and students from DTU (Denmark). The ReciPlyDome is a bending-active reciprocal dome structure based on polyhedral shapes and simple RF joints. It is a kitof-parts system made from pre-bent, double-layered plywood components. More specifically, the ReciPlyDome is based on a rhombic triacontahedral base shape. A full-scale demonstrator was developed and constructed at KADK during March 2017. It was exhibited outdoors on the grounds of the school for two months, during which time the structure was monitored for deformations. Initially, it was built as an open structure, without cladding. Later, with further development, it was constructed using the same plywood structural members for the Roskilde Music Festival in the summer of 2017. This time, a shading structure was developed as a sun protection. Later and with further development for the Circular Economy Exhibition at KADK in September 2018, the ReciPlyDome was built and was partially clad with a lightweight waterproof membrane, hence the new name ReciPlySkin. The cladding does not fully enclose the space, rather, it suggests that the structure could provide habitable spaces. It is intended to develop the cladding further, in order to achieve a fully watertight insulated enclosure. The full-scale demonstrator developed is about five meters in diameter enclosed by a structural grid of double-layered plywood members with an overall height of four meters. The dome is geometrically symmetrical, with four nearly vertical rhombs and pentagons around them. The rhombs suggest possible entrances, and the symmetry suggests possible strategies for developing a cladding system, as one developed with the ReciPlySkin. The dome was constructed of 12-mm birch plywood plates, providing enough flexibility for bending the beam-members 2.20 meters in length into the required shape with enough flexibility for the assembly and, also sufficient strength and stiffness to fulfill the load-bearing requirements (Brancart et al., Popovic Larsen et al. 2017). A central decision was the reduction of the complexity, with the aim of creating a kit-of-parts that is simple to fabricate and construct. People with no prior construction skills could fabricate and build it themselves. Thus, the ReciPlyDome demonstrator structure was designed 97

Rethinking Wood

to have only two different elements: forty basic beam elements were used together with five shorter members. All elements would have been identical had the form been a full sphere. The choice of plywood was based on having a widely available material that is inexpensive. Plywood is not commonly used for load-bearing purposes, but by introducing pre-stress the plywood is given additional strength and stiffness to offer a structural application. The ReciPlyDome project shows the potential of a system for rapidly assembled structures, demonstrating a fast, low-tech fabrication and assembly process. The kit-of-parts has been developed to create different sizes of domes, with different densities of structural members. All these have reversible joints that allow for disassembly and re-assembly. Furthermore, an universal beam member has been designed so that, once disassembled, other different configurations can be constructed. By virtue of the simplicity in the connection design with only two members meeting in a point, the ReciPlyDome creates opportunities for design for adaptability, assembly, and disassembly— addressing a circular economy approach where structural elements can be reconfigured to create new spatial assemblies providing the basis for sustainable design. To a great degree this is also true for the “common” RF beams; however, especially kit-of-parts designed RFs, like the ReciPlyDome and the ReciPlySkin offer this advantage to an even greater degree.

Conclusion and directions for further development Due to the manner in which they create a loadbearing structure—as an assembly of mutually supporting beams, where only two beams are connected at a time, with simple connections and without the need of specially fabricated metal ball joint-type connectors, reciprocal frames have a great potential. The elements used can be short and lightweight, and by combining them in RF configurations, they can create lightweight, structurally efficient configurations that are space forming and architecturally distinct.

Figure 7 a–b. ReciPlyDome and ReciPlySkin: kit-of-parts and typical connections.

98

The development in digital design/fabrication tools is at a level that deals with all complexities in modeling, designing, and fabricating any complex RF morphologies. If one should invest in research and design development, these efforts should be mainly in new materials and new cladding systems. Possible research and development is needed in: • increasing knowledge about the behavior of plywood used in bending-active elements • experimenting with novel natural fiber-based materials, both for load-bearing and enclosure purposes • developing other cladding systems for RFs that are equally easy to install as the membrane in ReciPlySkin, yet are insulated and thus more permanent. By their further development, RFs will become even more mainstream. Thus, we will make a positive contribution to developing a more sustainable way of building and to the reduction of the negative environmental impact of construction. Furthermore, we will create opportunities for the construction of more examples of functional and aesthetically distinct reciprocal frame architecture.

Acknowledgements The author would like to acknowledge the great teamwork of colleagues from the VUB, professors Niels De Temmerman, Lars De Laet, and PhD student Stijn Brancart from Vrije University Brussels (VUB), Department of Architectural Engineering, Belgium, uring the design, development, fabrication, and construction of the ReciPlyDome and ReciPlySkin. Many thanks also to the Royal Danish Academy of Fine Arts (KADK) research assistant Veronika Petrova for her efforts, especially in fabrication and process documentation of the ReciPlySkin, as well as to students Mikkel Asbjørn Andersen, Niklas Munk-Anderson, and Christian Jespersen from the Danish Technical University (DTU) who were involved in the fabrication and construction of both the ReciPlyDome and ReciPlySkin. This research was funded by Royal Danish Academy of Fine Arts School of Architecture (KADK), the Flemish Institute for Innovation through Science and Technology (IWT, now VLAIO), and the Vrije University Brussels. The collaboration between the Vrije University Brussels (VUB) and Royal Danish Academy of Fine Arts (KADK) was made possible thanks to the support of the VELUX Foundation’s visiting professorship program, which funded Niels De Temmerman's visiting Velux professorship at KADK between 2015 to 2017, and the European COST Action TU1303 on Novel Structural Skins. The images and illustrations for this publication were produced by Veronika Petrova and the author. 99

Rethinking Wood

100

Press-Fit Timber Building Systems: Developing a Construction System for Flexible Housing Solutions Hans Drexler

Press-fit timber buildings: developing a skeleton construction system Before the advent of industrialization, timber was one of the most important construction materials due in particular to its local availability, easy workability, good load-bearing, and overall structural properties. For centuries, carpenters built timber constructions without metal fasteners. Europe and Asia have a wealth of historical examples of timber joints for the widest variety of load cases and construction designs. Multi-story, large, and complex buildings were constructed with purely timber-based methods and only press-fit timber joints. But this tradition fell into disuse in the wake of industrialization. Not only were such joints—made by highly skilled craftsmen in a time-consuming and thus wage-intensive manner—replaced by cheaper metal fasteners such as nails, screws, and bolts, but even timber itself was displaced as a primary construction material by new construction materials such as cast iron, steel, and concrete. To this day, these materials are sold at lower prices because they can be produced in higher volumes and their external costs, especially environmental costs, are left out of the pricing equation. Thanks to advancements in design and production technology, it is now possible with computer-controlled machines to produce perfectly fitting joints economically, regardless of the complexity of the geometry. In a series of case studies and two research projects in cooperation with Pirmin Jung Ingenieure, DGJ Architektur has developed a timber construction system in which the entire load-bearing structure and all joints are press-fit without metal connectors. Part of the research is a complete set of construction details so that the system fulfills today’s regulatory requirements regarding fire protection, acoustic insulation, structural calculation, and industrial construction standards. The construction system is based on readily available materials, technologies, and processes at all levels so that the system can be applied immediately by everybody. Standardized and inexpensive materials that are available in large volumes are used: solid structural timber Figure 1. Yatoi-Hozo-Sashi, three-dimensional joint of the Ashikatame post (main post). Graphic by DGJ Architektur according to Sato, Hideo; Nakahara, Yasua; Nii, Koichi Paul (translator) The Complete Japanese Joinery; Vancouver, 2000.

101

Rethinking Wood

Figure 2. Three possible floor plan configurations (top), physical model of the construction system, scale 1:25. Photograph: Hans Drexler, DGJ Architektur, 2018 (below)

102

(KVH), laminated timber (BSH) and cross-laminated timber (CLT), or glue-laminated ceilings. The structural analysis of the load-bearing structure and the joints are based on the current Eurocode and in principle any engineer can calculate and dimension them. Thus, the innovative part is not newly developed materials, but the more intelligent use of existing methods, technologies, and materials and combining them into a system. As part of our research project, DGJ developed a design tool that enables an application of our system to be pre-dimensioned as well as costs and performance to be evaluated. The joints are designed using modern, three-dimensional CAD-systems widely used by architects and engineers, and so the machine data can be generated in the design phase with little extra effort. The system including press-fit joints can be produced with widely available timber CNC milling machinery. In this way, traditional knowledge is translated into contemporary building technologies and materials.

Global structure: timber skeleton versus solid timber In developing the load-bearing structure, several options for construction were explored: a skeletal construction, a hybrid construction, or a solid timber construction. An essential part of the research and development process work concerned the description and evaluation of the advantages and disadvantages of the various construction methods. A key parameter is the amount of timber required, which is compared for the global load-bearing systems. Below this, comparison is made on the basis of an 84-square-meter housing unit. Other parameters that were examined included the usability and flexibility of the building, but also fire protection, acoustic insulation, and the thermal construction physics. Timber—produced from tree trunks—is initially a linear, rodshaped construction material, which lends itself to building skeletal constructions (Merz 2009). Taking Steven Groák’s definition of flexible housing (Groák 1992, Schneider and Till, 2007 1), i.e. a structure that is physically and structurally changeable, the skeleton construction offers the best starting point from which to create flexible housing. The advantage of a skeleton construction is that it separates the functions of space creation and load-bearing. As a result, the load-bearing skeleton allows both external and internal walls to be moved and altered. If one considers the rapid changes happening in the world today in terms of lifestyles, housing forms, and demographics (e.g. the declin-

Tatjana Schneider and Jeremy Till take up Steven Groák’s definition: flexible housing as a living environment that can be physically changed. In this sense, our designs should be classified as flexible housing because occupants have the option to physically reshape the surroundings in which they live. Adaptive housing is described as adaptability with regard to various social uses without making physical changes.

1.

103

Rethinking Wood

Figure 3. Diagram of a possible global load-bearing skeleton structure, designed by DGJ Architektur and Pirmin Jung Ingenieure. Illustration: DGJ Architektur

Figure 4. Diagram of global load-bearing structure, “hybrid” variant. Load-bearing structure designed by DGJ Architektur and Pirmin Jung Ingenieure. Illustration: DGJ Architektur

Figure 5. Global load-bearing structure variant “Solid Timber Construction”. Designed by DGJ Architektur and Pirmin Jung Ingenieure. Illustration: DGJ Architektur

104

ing the number families with the typical, traditional family structure, shrinking household sizes), then it is clear that inflexible housing layouts are not sustainable in the long term. In Case Study 1 (Heidelberg International Building Exhibition/IBA Heidelberg) with a load-bearing timber skeleton, a bracing core, and ceilings of cross-laminated timber (CLT), as well as partition walls and facades as a timber-panel construction, the amount of timber used totaled 24.3 cubic-meter per housing unit of 84 square-meter. For finishing (non-loadbearing walls) and facades, timber-panel and timber-frame constructions are used, which can be produced inexpensively.

Hybrid solutions: combining skeleton and solid construction In the mid-1990s a solid timber construction known as cross-laminated timber (CLT) came onto the market, which offered key advantages especially for taller buildings, because solid wall panels perform better in terms of vertical load transfer and cross bracing of the global load-bearing structure. When developing this press-fit construction system, a combination of the advantages of the two construction methods was explored: parts of the skeleton structure are replaced by load-bearing and bracing cross-laminated timber elements, which allows for an efficient transfer of the horizontal loads. Thus, a part of the walls enclosing the bathrooms, which are located in the same vertical line on all floors, and a part of the party walls are designed to replace posts as load-bearing wall segments. The horizontal bracing is achieved by means of bracing walls of solid timber (cross-laminated timber) that are connected to the posts by means of tenons. For the hybrid construction, consisting in part of solid walls, the amount of timber used in Case-Study 1 (IBA Heidelberg) totals 28.2 cubic-meter or 14 percent more than with the skeleton construction.

Solid timber construction In the course of the development of the system, it emerged that the introduction of cross-laminated timber resulted in a construction that combined aspects of a skeleton construction (i.e. only linear load-bearing elements) with aspects of a solid construction for ceilings and bracing walls. This led to redundancies at the intersections. The bracing wall could take all loads vertically and thus replace not only the column, but the beam as well. It is thus conceivable that this system, based on the grid, could be implemented as a solid timber construction. The disadvantage of solid interior walls is their low adjustability. The solid construction for ceilings and bracing walls is more material-intensive, although the strength of the individual walls can be reduced by the uniform load transfer. For Case Study 1 (IBA Heidelberg), the amount of timber used for this variant totals 31.9 cubic-meter per 105

Rethinking Wood

Figure 6. DGJ Architektur, 3D model of the construction system, beam to post joint. DGJ, 2016

Figure 7. DGJ Architektur, 3D model of the construction system’s beam to post joints. DGJ, 2016

106

housing unit, or 24 percent more than the amount of timber required for the variant combining skeleton structure and timber tiles.

Local structure and joints Given the wide variety of legacy timber joinery connections such as tenon and mortise, lap joints, dovetails, birdsmouth joints, etc., we examined which joints can be applied to modern constructions easily in real-life practice. Here we examined and compared two different approaches: integrated or differentiated joints. The integrated joints are cut so that the elements fit into the geographic space of the adjacent elements and guarantee a press-fit joint via the interlocking of the elements’ geometries (geometrically interlocked and press joints). The initial idea of our construction approach was to produce geometrically locking and press-fit joints (press-fit, or friction-fit joints) when cutting and sizing the elements, thus making fasteners unnecessary. Our reference point was a Japanese “YatoiHozo-Sashi” joint, which fits two main beams into the geometric space of the post to form beam-to-post joints, but also forms a lapped end-to-end connection from beam to beam by means of rod tenons, which are secured in place with wedges or cross tenons. Here the load of the beams is transferred not only via the rod tenons, but above all via a parapet from beam to post, into which the beam is precisely fit. The principle of this joint is that the geometric spaces overlap and the members are joined with comparably complex geometries. Differentiated joints separate (i.e. differentiate) the geometric spaces of the components, which are then connected via secondary elements. The joining of the solid ceiling elements to the post is achieved with dovetail joints. This type of joint is complicated and costly to produce but allows for the transfer of shearing, compressive, and tensile forces. The assembly of the construction components is accelerated in that the bond between the elements is created directly when they are inserted into the structure without the need for secondary elements such as wood dowels or dovetail keys. The system is also accordingly easy to dismantle later on. The disadvantage is that putting it together requires a high degree of precision and leaves little room for dimensional tolerance. In this context it is interesting to discuss the weakening of the supporting beams by such geometric joints. At first glance the sections in beams and columns seem to increase the volume of timber necessary for the structural performance. But it became apparent during the development that there is a system synergy between fire protection and the structural performance. When measuring the load-bearing capacity of the beams under fire conditions, the entire timber volume is not applied to the load transfer, but it is assumed in the case of fire that the outer layer of the beam burns away slowly, which ensures the stability of the remaining beam for the prescribed amount of time. In 107

Rethinking Wood

Figure 8. IBA Heidelberg, Self-Governed Student Housing, Collegium Academicum. DGJ Architektur, Model scale 1/200. Photograph: Hans Drexler

Figure 9. IBA Heidelberg, construction details and prototypical housing unit. Photograph: Hans Drexler

108

the case of the press-fit joints, this layer is also the part of the structural elements used for the joints, which serves as integrated fireresistant cladding. In the first case study (see Case Study 1: IBA Heidelberg), the connections between post and beams and beams and ceiling were increasingly simplified in the course of the design process, i.e. separated into their own geometric spaces. Vertically, the forces are achieved by lapping the elements, which are secured in position with beech wood dowels. The stable plate effect of the ceiling elements is achieved by lap-joining ceiling elements with dovetail joints. This differentiated construction has two advantages: first, it makes the timber trimming process for the individual elements easier because the elements are lying above and next to each other. The geometry of the components is simpler and the production cheaper. Second, the differentiated construction of the components allows the dimensional tolerance of the individual parts to be compensated across the load-bearing structure, as the horizontal accuracy of the fit is initially not decisive for the load transfer. The vertical inaccuracy can be compensated locally. The disadvantage of joining in this way is that the connections with beech wood dowels can only be released by drilling out the dowels, which makes dismantling more difficult.

Case studies The system was developed in 2015 through various case studies. In 2019, the first two prototype buildings will be completed. During the research, the development of the system (planning and design methodology) was combined with analysis (evaluation and optimization). This has brought to light an essential aspect of explorative applied construction research: Design has both creative and analytical aspects, which interact with each other in iterations and recursions. In the day-to-day designing and planning process, iterations, parameters, criteria, and decision paths are seldom explicit, and as such they are not transparent. They take place in numerous sketches, drawings, models, and discussions.2 In the research and development project, the analytical parts are made explicit and operative, thus complementing design. From these analyses, tools were developed with which the construction system can be adapted to different locations and functions. In developing this system, the levels of research and parameters for a given object are evaluated and optimized. This led to the creation of an Excel-based design tool that allows the designer to estimate the

2. Edward T H Liu (Liu 2012) states: “In a climate where parametric design and generative components are regarded as the cutting edge of architectural design, it is worth noting that all buildings are parametric, in the most literal sense of the word. Through the architectural design process, form inevitably emerges from a series of overlapping and, often, contradictory constraints—whether physical, environmental, cultural, or aesthetic.”

109

Rethinking Wood

Figure 10. WBS2017—IBA Thüringen, DGJ Architektur, 2017.

Figure 11. Arrival City 4.0, Model of construction system, scale 1:50, DGJ Architektur. Photograph: Hans Drexler 110

inputs (timber volume, costs) and performance (living space, efficiency (gross floor area/living space), lighting, ventilation) of an implementation of a construction task in the system. Case study 1 IBA Heidelberg: The Collegium Academicum is a self-governed student residence with 176 units located in Heidelberg. A new housing form was developed which enables residents to change and adjust their living space on various spatial configurations. The residents can easily move and reposition partition walls within the apartments. This means the size of the rooms can be changed. The individual rooms consist of two sections: a core area of seven square meters and a flexible area of seven square meters, which cannot be spatially separated from the common area of the apartment. Depending on individual preferences, the flexible area can remain completely open, partially separated by dividing elements (table, shelving), or partitioned off completely by placing the wall between the core area and the flexible zone. The spatial and construction ideas of this housing design with its timber-to-timber joints were tested in a one-room prototype building consisting of a core area and a flexible area, each seven square meters in size. Case study 2 The housing group Gemeinsam Suffizient Leben (Living together sufficiently), Frankfurt: This project started with a competition for a vacant plot in Frankfurt, in which the best housing concept and not the highest price wins. Our skeleton construction proved to be a suitable solution. The square grid permits different apartment layouts to be combined flexibly. At the same time, the individual layouts of the apartments remain flexible/adjustable throughout the lifecycle of the building. Case study 3 WBS2017—IBA Thüringen: In the competition entry for the IBA Thüringen (International Building Exhibition) we used this construction system for an eight-story building. The main focus of our work in this project was the flexibility and adjustability of the apartment layouts, which permit a number of adjustments to be made. Case study 4 Arrival City 4.0 was designed as a response to the so-called refugee crisis in Germany at the end of 2015. The project was intended to demonstrate how spatial and social structures can contribute to integration. The task was to design suitable housing for people to give them not only a place to live in the short term, but also to give them a long-term perspective. Arrival City 4.0 is an expandable design that represents a low-investment approach to housing shortages in gen111

Rethinking Wood

eral. For a minimal investment, simple housing units can be built in a very short space of time. In contrast to normal emergency shelters (tents, containers), Arrival City 4.0 provides the prospect of being consolidated into a lasting building that represents a valuable part of the city. This design also features the possibility for the new arrivals to help with the construction of their new home. The case studies are a vehicle and a research object at the same time. The transferability and usability of the construction system are identified. Through the systematic research of the application parameters, load cases and characteristics, an estimate of the dimensions of the structure can be calculated for each application. To study the possibilities in terms of building typology, we first defined a series of design parameters, the permutations, which can be expressed as a matrix. The matrix is a parametric design tool for the performance and dimensions of the system depending on the building-typological characteristics of the respective permutation. In the matrix, all three analysis levels—design, structure, and performance—are interlinked and influence each other.

Advantages and opportunities of the timber skeleton construction system Due to the use of non-renewable construction materials and given that most constructions are never dismantled and separated into reusable and non-reusable waste, the construction industry is responsible for a large share of waste that is generated today. This situation also results in a high use of resources. Sustainable buildings must be built differently: The individual construction elements should be joined in such a way that they can be taken apart again. The shearing layers of a construction should be joined in such a way that they can be maintained, repaired, or replaced individually, since they are all subject to different degrees of stress and strain and have different useful lives.3 Pure timber constructions can be reused, recycled, or down-cycled to engineered wood. Such constructions promote strict sorting, which is an important prerequisite for recycling materials or using them efficiently for energy generation. For all of the above, timber-to-timber joints offer an ideal starting point. Our construction system was developed so that it can be modified, dismantled, and reused with as little destruction as possible. Theoretically, the construction elements can be put together and taken apart like puzzle pieces. Elements and materials can be managed

Frank Duffy, as cited by Stewart Brand (Brand 1994) Michael Braungart and William McDonough have developed the concept of a cradle-to-cradle economy in which products and materials are kept in closed-loop cycles (Braungart et al. 2002)

3.

4.

112

in closed-loop cycles, thus eliminating waste.4 In practice, the ability to dismantle a primary construction is constrained by the addition of such things as interior cladding, facades, and building technology. Thus, for now, we are focusing our research on the load-bearing structure, although the secondary and tertiary systems are built in such a way that they can be separated as easily as possible from the primary structure. As a domestic, renewable resource, timber is the most sustainable construction material available, since it is the only construction material that can be used for all construction tasks in large volumes and is not based on limited resources (fossil fuels, sand, iron ore …). Thanks to Germany’s established, sustainable forestry industry, timber can serve indefinitely as a renewable raw material. Model computations indicate that the timber needed for the country’s total construction activity can be covered by just one-third of its annual timber output (Kaufmann et al. 2017). By using timber as a renewable resource, our construction system has ecological advantages; but improving the reusability of the construction elements can increase these advantages even further. The production and processing of timber requires significantly less primary energy. Wood binds carbon dioxide from the atmosphere and thus stores it over the lifetime of the building. In the hybrid construction variant, the reduction of CO2 emissions by wood is partially offset by high emissions related to secondary materials, like cement or steel. With regard to commercial viability, timber-only joints may have advantages over constructions using metal fasteners. The fasteners by themselves are high-quality and fairly expensive construction elements. Moreover, mounting the fasteners represents a considerable delay to production and assembly at the construction site. When processing times are reduced, costs are also reduced. The more complex timber-trimming process means longer processing times at the trimming machine. Eliminating fasteners and faster assembly at the construction site saves costs. Timber constructions in general and our specific construction system in particular also have disadvantages compared to cement and masonry structures, the most common construction forms for residential buildings. Due to their light weight, timber constructions absorb less sound energy than heavy constructions made of masonry and reinforced concrete. This results in a conflict of objectives in the case of constructions using press-fit joints. Such joints generally represent an acoustic coupling through which sound is transmitted through partition walls and floors. Two strategies can be used to minimize the problem: decoupling and stiffening. Where the construction is decoupled, the sound transmission is minimal. The partition walls between apartments were doubled and completely decoupled. The floors are equipped with sound insulation, which decouples them from the primary structure and the rooms below. Beams and 113

Rethinking Wood

load bearing walls need to be assessed individually as secondary noise transmission channels. Here the strategy was to stiffen the connection between the column and the beams, as well as between the walls and the ceilings respectively, to a point where all components acted as a unified system, which due to the increased mass absorbs more sound energy, will therefore not be transmitted. Given the combustibility of the primary construction material, special care must be given to fire protection. In the simplest scenario, timber can protect itself from combustion. Timber forms a relatively stable charcoal coating in a fire, which in turn delays full destruction by fire. This effect is included in the dimensioning of the construction elements for fire conditions, in that the beams are sized larger than what they need to be to meet their load capacity requirements, so that in case of fire the remaining beams assure the stability of the building for the prescribed amount of time. Press-fit joints can as a rule be considered homogenous construction elements. In practice, however, care should be given that the joints are so tight that fire cannot penetrate them. Supporting and bracing elements are generally not reached by a fire for five to ten minutes due to the low combustibility of a closed timber surface. In most cases, fires are fueled initially by furniture and fixtures, so that houses have been evacuated by the time the fire reaches the supporting and bracing elements. This, in turn, means that the primary fire protection goal would be met. If added fire protection of the supporting and bracing elements is required, e.g. because it is prescribed by regulations and/or due to increased evacuation risks (higher buildings, more complex layout), the primary timber load-bearing structure can be additionally protected from fire by encasing it with cladding so that the combustible elements can be protected from the fire for a defined period of time (30 min., 60 min., 90 min.). Press-fit joints require a high degree of precision in production and assembly. Especially in the case of larger buildings, where imprecision in construction elements and their fitting can add up, joining with press-fit elements is a challenge. When developing this system, two methods were developed to deal with the tolerance problems that will inevitably emerge at the construction site. One method that can be used with integrated joints (see the above section, “Local load-bearing structure”) is to introduce tolerances in the joint details, manufacturing the partner elements of the joints with a gap 5–10 mm depending on the joint design. In this way, the individual construction elements can be joined more easily and repositioned within the planned tolerances. Dimensional deviations can be offset when the individual elements are

6. Kaufmann, Hermann; Krötsch, Stefan; Winter, Stefan; Atlas Mehrgeschossiger Holzbau; München, 2017, p. 26 et seq.

114

being fixed to one another. The disadvantage of this method is that for the press connection between the elements to work properly, the gaps must be friction-locked with a filling material. Materials suitable for this purpose would be expanding cement or epoxy resin. The disadvantage of this solution is that it is an additional processing step and drying times will delay the building process. Closing gaps with filling material also makes it more difficult to dismantle the construction later on. The goal of developing this system must be to reduce tolerances to the extent that the system can be stabilized without filling materials. Another method would be to separate the geometric spaces of the individual layers, which would allow the individual elements and layers to be positioned independently. Key data This construction system is being developed within the aforementioned model projects and is being accompanied by two research projects, in which fundamental aspects are being researched and, for the purposes of a construction system, developed and generalized. In one research project at the German Federal Institute for Research on Building, Urban Affairs and Spatial Development/BBSR (SWD10.08.18.7-17.28 “Development of a timber-only construction system with press-fit geometric joints”), the global and local support structures (joints) are being developed. In the complementary project at the German Federal Environmental Foundation/DBU (“Press-fit timber construction system—the interplay of support structure, fire protection, noise insulation, and building physics”) the synergies and obstacles to interaction of the support structure with the construction-physical requirements are being examined. Project partners: Research: DGJ Architektur GmbH PIRMIN JUNG Deutschland GmbH Industry partners: Brüggemann Holzbau, Neuenkirchen Brüninghoff GmbH & Co. KG

115

Rethinking Wood

116

Glued Connections in Timber Structures Gerhard Fink and Robert Jockwer

Introduction As a naturally grown material, the dimensions of structural timber exhibit certain limitations. In order also to use timber for larger structures and wider spans, different techniques have been developed to overcome these limitations. Initially, carpentry connections have been used to combine individual timber elements to frame or blockhouse structures. As a further development mechanical fastener and shear plugs have been used to combine individual timber boards arranged in parallel to larger structural members, so-called “Bohlenbinder,” a precursor of glue-laminated timber. The development of adhesive offered new possibilities for timber structures. One of the first, and still most important, applications of adhesives in timber engineering is the production of engineered wood products. Well-known examples are finger joint connections between timber boards that are needed for overcoming the limitations in length of timber boards and for the creation of lamellas of arbitrary length, as well as surface gluing between timber lamellas or veneers that are needed for creating elements of greater thickness or height and that are applied for almost all types of engineered wood products such as glue-laminated timber, laminated-veneer lumber, and cross-laminated timber. Engineered wood products allow for an industrial use of the recourse timber and enables building large span structures and multistory timber buildings. The use of timber in the building sector has become competitive with other building materials. With the increasing number of timber constructions, new challenges are being faced and new solutions are being developed. Among the most relevant glued connections that have been developed are universal finger joint connections, glued in rods, shear connectors for timber concrete composite systems, and butt-joint bonding. Some of them are still under further development. The excellent performance in existing applications, as well as promising results from ongoing research activities indicate that these modern glued connections will replace traditional, mechanical steel connections in many applications. The main advantages of glued joints are higher rigidity and higher load bearing capacities compared to metal type fasteners and the potential of automation as it has been already shown, for example, for finger joint connections. On the other hand, the fabrication of glued connections is more sensitive to ambient conditions and production quality. Therefore, high requirements are set to the qualification of the 117

Rethinking Wood

manufactures and on-site manufacturing is restricted. Furthermore, the majority of glued connections show brittle failure mechanism, which might result in fewer redundant systems and requires special attention in design. In this contribution, the structural behavior of glued connections is introduced briefly. Different types of glued connections currently used in practice are summarized and discussed. Their potential with respect to future applications will be evaluated and new and innovative solutions will be presented. The contribution is limited to structural adhesive joints used to transfer loads in a defined area usually between two members.

Structural behavior Glued connections in structural applications have to fulfill certain requirements with regard to strength, stiffness, and ductility as well as fire resistance, durability, and health protection. The performance of glued connections can be benchmarked and discussed on these requirements. The most essential structural application in order to evaluate the performance of the adhesive is the connection of two axial loaded timber members assembled in tension. This will be discussed on the basis of a finger joint connection. In continuation, the discussion will be extended to other types of connections. Example: axially loaded finger joint connection The connection of two, individual timber boards to a continuous lamella is one of the basic challenges in timber engineering, in order to overcome the limitations of the natural material wood. In figure 6, the typology of a finger joint connection, a standardized solution to connect two timber boards, is illustrated. In case of pure axial tension, which is the relevant load situation of individual lamellas in a GLT beam, the axial tensile load has to be transferred from one timber element to the other through the glued connection. The finger joint geometry is optimized in commercial applications and can be described

Figure 1. Schematic illustration of a finger joint connection.

118

with the finger length l, the pitch p, the tip width wt and the tip length lt. The tip length lt results from the other geometries as well as from the fabrication process, and is commonly lt/l≈0.03 (Aicher 2003). The mechanical behavior of the glue line is rather complex and thus different approaches have been developed. Examples are: linear elastic strength of materials, plasticity theory, linear fracture mechanics, and nonlinear fracture mechanics. Comprehensive summaries are presented in Serrano and Gustafsson (1999) and Serrano et al. (2001). For an axially loaded example, the stresses along the glue line can be subdivided into stresses acting in shear and in normal direction with regard to the contact surface, both depend on the joint angle α. For small angles and, hence, very long finger length, the normal stresses in the contact surface are comparably small, or even negligible. The shear forces also decrease for decreasing α, mainly due to the increasing contact area, however, the stresses will not become negligible. Due to the high shear strength of common adhesives compared to their relatively low tensile strength, such situations were preferred where shear governs the design of the glued connection. From a technical perspective, a scarf joint (only one inclined glue line) would be the most effective solution. However, due to economic reasons, such as the smaller offcuts, finger joint connections have been established in most applications, having both economic efficiency and good structural performance. When loading a finger joint connection in tension, the following failure scenarios (and combinations of them) are possible: 1. Failure of the glue line itself 2. Failure of the interaction between the glue line and the timber member 3. Failure of one of the adjacent timber members The first two failure modes can be avoided by carefully choosing an adhesive that shows sufficient strength and that is suitable for the timber species combined with a sufficient production quality. The third failure mode is not influenced by the adhesive and sets the upper limit of the load-carrying capacity of a finger joint. In the area of the finger base, the cross section that is able to carry load is slightly reduced, thus the timber failures often occur in this zone. Furthermore, different growth irregularities such as knots and the resulting grain orientation might affect the capacity of the connection. As a result, the European standard EN 14080 (2013) defines a minimal distance between knots and the finger joint connection (Figure 1). As the lamella consisting of two timber members and a finger joint connection is loaded in axial tension, it will fail as soon as its weakest part reaches the load-bearing capacity. Accordingly, it is not possible to make the lamella stronger than the members that are connected together. Even if the tensile capacity of the finger joint connection 119

Rethinking Wood

(failure modes 1 and 2) is much larger than the tensile capacity of the timber members, the lamella will not be able to resist a higher load. This general limitation due to the failure of the weakest link has to be considered for every kind of connection. Types of adhesives used in structural applications In a built structure, the connections usually have to transfer a variety of different loads, not only loads in axial direction. Hence, typical connections are designed to resist different types of loads, including shear loads, compression loads, bending loads, and torsional loads. The loads may act in different directions to the grain orientation, in the worst case also perpendicular to grain direction, which is the direction with the lowest strength and stiffness of the timber. In addition to mechanical loads, structural members, and thus also the glued connections, have to resist environmental exposures such as different temperatures, changes in humidity (variations of moisture content), UV radiation etc. Accordingly, the requirements for adhesives vary significantly between applications and different types of adhesives have been developed for specific applications. Examples are Polyurethane adhesive, Phenol formaldehyde adhesive, Epoxy resins, and Melamine formaldehyde adhesive. For more information about available adhesives, refer to standard literature (e.g. Dunky and Niemz 2002, Frihart and Hunt 2010). In order to guaranty optimal (mechanical) behavior, each different type of adhesive requires a specific procedure for application; typical varying parameters are the amount of adhesive, the curing time, the curing pressure, the time between the application of the glue and the connection of the timber parts, but also the quality (e.g. the roughness) of timber surface. For different wood species and different wood products, different procedures and different adhesives are also necessary. Strength, stiffness, and ductility Deformation behavior is one of the most relevant parameter for connections in a structural application, in particular under high load exposure. Deformation behavior can be characterized by the strength, stiffness, and ductility and the resulting failure mode. It can be distinguished between brittle failures (no or almost no plastic deformation) and ductile failures (showing considerable plastic deformation before failure). In figure 2, two typical stress-strain curves of a connection are schematically illustrated. One line illustrates a linear elastic brittle behavior and the other line a ductile behavior with non-linear plastic deformations before failure. Both lines have the same initial stiffness and show an ideal elastic behavior at the beginning; i.e. if the load is be released, the connection will deform back to its individual initial state. The main advantage of a ductile connection is that due to the 120

larger deformation capacity, the static system can adapt: stresses can be redistributed from the highest loaded elements to other parts of the structure, thus the stresses in the connection decreases and the collapse of the entire structure may be avoided. Mechanical connections with metal dowel-type fasteners can show a failure behavior with considerable ductility. Glued connections are commonly very rigid connections, which do not show considerable ductility in the glue line. Rigid glued connections are usually connections with very thin glue lines that show a brittle failure mechanism. In applications with small load eccentricities, rigid glued connections lead to stress peaks in the edges of the glue lines, whereas the inner part is almost free of stresses. In such cases an increase of the length of the glue line does not increase the load bearing capacity of the connection significantly. By choosing a lager glue line thickness, structural behavior of the glued connections can be enhanced. The stress distribution in the glue line is more equal and by ductile deformations in the glue line, local stress peaks can be reduced and force redistributed. Nowadays mainly thin and rigid glued connections are used in timber engineering and thus glued connections are often characterized by predominantly brittle failures. However, several applications for ductile-glued connections are currently under development. In cases of glued metal fasteners, such as glued-in rods, the connections should be designed in order to achieve ductile failure in the metal fastening elements. This can be achieved when overdesigning all the other failure modes, including timber failure and failure of the glue lines.

Figure 2. Schematic illustration of a brittle and the ductile behavior.

121

Rethinking Wood

Fire resistance In timber engineering, the fire resistance of elements and structures plays an important role in general, and especially for connections, behavior under fire exposure is of even higher importance. Under fire exposure timber has a rather predictable behavior: Due to the relatively constant charring rate combined with the high thermal isolation of the charred wood, the majority of the remaining cross section of the timber member can carry the load even after a long fire exposure. Critical points in a structure under fire exposure are the connection points. In case of a connection with metal type fasteners, these conduct a great amount of heat into the inner part of the timber members and cause additional charring from the inside (e.g. Palma 2016 for more detailed description). Glued connections do not show this problem, however, the thermal stability of the adhesive might be relevant for glued connections under fire exposure. However, due to the good thermal isolation of the wood, the majority of the connection(s) are commonly not affected by the fire. Nevertheless, careful design of glued connections is necessary especially in the event of fire. As an example, cross-laminated timber elements may be mentioned, where the strength loss in the glue line and the associated fall-off of individual layers are a challenge in case of fire. Area of application The applications of glued connections can be subdivided in two main groups, namely connections for engineered wood products and connections between structural members. Engineered wood products Engineered wood products can be categorized according to their basic material, starting from timber boards up to fibers. In general, it can be stated that the ratio between wood and glue increases with decreasing size of the basic material; i.e. engineered wood products from timber boards need a smaller amount of glue compared to those products made from wood fibers. Typical amounts are 3–5 percent for glued laminated timber, >5 percent for veneer products, and 7–14 percent for span products (Dunky and Niemz 2002). Within the framework of this chapter, the focus will be concentrated only on selected engineered wood products that are commonly used for structural purposes: glued laminated timber, cross laminated timber, and laminated veneer lumber. For both glued laminated timber and cross laminated timber, two different types of connections exist: finger joint connections, in order to connect individual timber boards in axial direction to lamellas of arbitrary and surface bonding to combine the individual lamellas on their side surface to beams of the required height or plates of the required thickness, respectively. As already mentioned, finger joint connections 122

are mainly loaded in axial direction either in tension or compression. In contrast, the surface bonding has to transfer mainly shear loads as well as torsional loads in case that cross laminated timber plates are exposed to shear forces. In addition, some cross-laminated timber producers also use narrow face bonding between the timber boards of the same layer. However, this bonding has mainly practical reasons, in order to optimize the fabrication process and is not considered to carry any loads. For laminated veneer lumber only one type of glued connection exists, the surface bonding between the individual veneers. In axial directions, the veneers are placed next to each other without connection. In order to guarantee the load transfer over the gaps between the veneers, the layers are located alternatively above each other. In figure 3, the typical cross sections of laminated veneer lumber are illustrated. Due to production reasons, engineered wood products are still limited in size and shape. However, individual parts can be combined to new members, by using large finger joints (universal finger joints) or surface bonding (block glulam). For both, the production follows similar principles as for glued laminated timber and cross-laminated timber. In addition to traditional engineered wood products also combined elements such as I-beams or box girders are available on the market. Combined beam type elements are usually fabricated to optimize the structural performance. A large number of different standardized solutions as well as several solutions for specific applications exist. The fabrication procedure and the structural performance of the glued connections follows similar principles as described above.

Figure 3. Typical cross sections of laminated veneer lumber. Photo: Metsä Wood

123

Rethinking Wood

Figure 4. Free-floating stairs at the University of British Columbia in Vancouver.

124

Connections between structural members Individual structural components can be connected to larger structural systems. The connection can be created without using any additional material (carpentry connections), with fasteners (usually metal dowel type fasteners, such as bolts or screws) or with adhesives as glued connections. In many cases the connections contains a combination of these. So far, the use of glued connections in practice is limited, the most relevant ones are glued metal fasteners such as glued-in steel rods or glued-in perforated steel plates. An example of the latter are the free-floating stairs at the University of British Columbia in Vancouver, Canada, (Figure 4) or the Timber Tower in Hannover (Harms et al. 2012). An example of a structure using glued-in rods as connecting elements is the dome structure for salt storage Saldome 2 in Rheinfelden, Switzerland, where each node has six individual beams connected, forming a spherical structure (Bogusch, 2012). For glued connections using a metal fastening member such as the glued-in rods or the perforated plates, the load has to be transferred from the timber element via the adhesive to the metal fastener. The adhesive is commonly optimized to show a good bond with the timber. In order to guarantee the load transfer between the adhesive and the steel element, a certain texture, such as the thread or ribs of steel rods or the holes in the perforated steel plates is usually required. For these types of connections, additional failure modes are possible compared to timber-to-timber connections, namely the failure in the interaction between the glue line and the steel member itself. As for a finger joint connection, the load-carrying capacity of the connection is limited to the capacity of the weakest link. The failure modes in the timber and in the bondline are brittle, thus these failures occur suddenly without any warnings. However, if the failure occurs in the steel elements, ductile failure modes can be achieved (e.g. Gehri 2016). Glued metal connections are to a large extend prefabricated, since the gluing process demands high quality. This remains rather challenging.

Potential for the use of adhesives in timber structures The ongoing development of digital fabrication processes shows great potential to be implemented more extensively in the construction process in the future. Even though the prototypes are becoming larger and larger, the dimensions of the individual structures remain small compared to existing structures in practice. This is amongst others a result of the limited production capabilities of e.g. the robots. For larger structures and assembles made of several small substructures, the quality and efficiency of the connections is of even greater important. In figure 5, possible examples of digital fabricated elements are illustrated. From both examples shown in figure 5, it become obvious that each individual connection in the structure is different with regard to geom125

Rethinking Wood

Figure 5. Complex Timber Structures, elective course, © Gramazio Kohler Research, ETH Zurich, 2013, Collaborators: Luka Piskorec, Michael Knauss1

Figure 6. Butt joint bonding offers potential for new ways of construction, © TS3 AG

1. Students: Lukas Ballo, Nishtha Banker, Tom Doan, Jacob Fink, Dominik Ganghofer, Pierre-Jean Holl, Rossitza Kotelova, Renuka Makwana, Daniel Michel, Unnati Mistry, Takashi Owadat, Irene Prieler, Micha Ringger, Pascal Ruckstuhl, Enrique Ruiz Durazo, Mari Saetre, Grau Sara, David Schildberger, Nishita Shah, Abigail Stoner, Taku Sugimoto, Andreas Thoma, Achilleas Xydis

126

etry and fabrication as well as with regard to the required structural performance. Highly standardized connections by means of mechanical fasteners can hardly satisfy the highly individualistic requirements on connections in such structures. Compared to mechanical fasteners, the application of adhesive is a rather simple process and requires a significantly lower workload. This shows the great potential of glued connections in digitally fabricated structures. Nevertheless, currently the application of such connections and structures is still limited and under further research and development. One of the main challenges is to guarantee the quality of the glued connection when being glued in-situ under suboptimal and changing environmental conditions, such as humidity and temperature but also due to the specific requirements of the highly unusual geometry and configuration of connections and glues. The production parameters, such as pressure, curing time or surface treatment have to correspond with developments in production processes. However, for standardized applications, efficient fabrication procedures exist and the application on glued connections, in particular in the fabrication process of engineered wood products indicating the potential regarding automation. A promising alternative to mechanical connections in larger structures is butt joint bonding of individual structural timber members (Figure 6) (Zöllig et al. 2017). Using butt joints structural elements of larger dimensions could be fabricated also on site and new building processes can be established. As mentioned before one of the main challenge is the quality of the in-situ gluing. However, since this type of connections having a regular shape some of the difficulties might be solved easier.

127

Rethinking Wood

128

Digital Processes Digital processing techniques for wooden materials have undergone extraordinary advances which maximized efficiency throughout the process from a digital description to a finished wood product. When combined with computer-aided design and simulation techniques, new possibilities are opened up beyond pure efficiency. For wood as a material in architecture, the boundaries of the buildable are pushed by increasing the material’s performance itself for highly adaptive, efficient, and differentiated structures.

129

Rethinking Wood

Whereas the wealth of formal and constructive complexity in wooden architecture of pre-industrialized times resulted from the availability of inexpensive artisan labor, our current automated wood processing techniques have resulted in higher efficiencies, mostly at the expense of expressive qualities and variability. In recent decades the development towards automation of a linear sequence of previously known human processing steps for greater efficiency has continued. Most industrial timber machines are optimized for a floating production in which a sequence of milling, cutting, or drilling operations outputs many simple wood connections, bound to mostly closed software environments. As the necessary simulation techniques, machinery, and infrastructure are becoming more easily available and software protocols for generating machine code are becoming more accessible, the experimentation with design and fabrication of mass-customized wooden elements according to digital instructions generated by the designer is becoming increasingly popular. Here a shift can be observed, which moves the intellectual activity from the design of architectural objects towards the design of a process of interconnection between digital information, machine and the tool. The malleability and availability of wood further promotes experimentation of new concepts in the manufacturing process which are influenced by the production lines found in automotive and other consumer product industries. Highly flexible and versatile industrial robots are being tested in university surroundings and among young practices. Robots offer more applications, since here specifics lie only in the control software and the tool used. By adding six degrees of freedom, the robotic workflows act beyond the scope of current CNC machines. More recently, a systemic and contextual understanding has initiated a change of focus towards previously underrepresented aspects, such as the material’s anisotropic properties or the use of inexpensive but highly individual pieces of scrap wood which are excluded from industrial processing, due to irregularities or inhomogeneous material properties.1 Harnessing computation, we do not only develop sophisticated design and fabrication processes, but also start to be able to model and analyze the behaviors and relationships of the inner structures of wood. In this chapter, new approaches to intelligently use industrial manufacturing processes and the material’s properties are presented, which explore and expand the possibilities of the timber construction industry’s current capabilities. These new and flexible workflows give rise to innovative designs for highly specific contexts, in some cases

1. See also Part Four, “New Materials and Applications, Designing with Tree Form,” Martin Self

130

by applying traditional wood processing methods to a technical and precise fabrication machinery. Working with several industry partners in the context of today’s wood construction industry in Finland, Markus Wikar and Toni Österlund of Geometria present complex and differentiated timber structures, designed with algorithmic tools and produced by CNC machines from wood elements in regular dimensions. Here, the importance of clear design methodologies which incorporate material and fabrication restrictions and transfer protocols from design to production environments becomes apparent. Utilizing state of the art computational tools, Philipp Eversmann’s research takes advantage of the increasing work space of the industrial robots to produce complex glued wooden structures. Here the loop between design and fabrication leads not only to greater efficiency in timber construction, but the interplay between computational form-finding, joint geometry, material properties, and assembly techniques generates a specific and unprecedented form of expression, produced with robotic fabrication. The versatility of industrial robots is also used by Studio RAP, an interdisciplinary group of architects, designers, and engineers. The multiple degrees of robotic manufacturing are used here to create context-specific solutions from pieces from scrap wood. Again, not only a novel architectural language is created, but a new type interdisciplinary, manufacturing, and material-oriented practice, in which architects collaborate with robotics engineers and material specialists. In recent years, a number of initiatives have been addressing the lack of awareness in the end-user about building with wood. By reviewing and redesigning steps in the value chain of architecture, processes are opened to actors traditionally not involved, such as the customer. These DIY developments are additionally stimulated by the increasing accessibility to hardware and software environments, as well as by a higher demand for sustainable and aesthetic outcomes. Advancements in technology are coupled here with innovative approaches of organizational and economic aspects of design. The sharing of knowledge and expertise enables the development of customizable products, and fosters community empowerment. To take the DIY production of wooden structures to the next level, complex processes have to be simplified. The project JOYN MACHINE by Simon Deeg and Andreas Picker of Studio Milz presents an interactive tool that enables the design of wooden framework constructions and their semi-automated production. The focus here lies not on the final designs, but on the machine itself, as well as the software front and backend, on aspects of usability.

131

Rethinking Wood

132

Freeform Timber Structures: Digital Design and Fabrication Toni Österlund and Markus Wikar

Introduction Digitalization has swept through the timber construction industry, currently encompassing all facets of it from tree harvesting and lumber production to the fabrication of building components. The industry is well equipped when it comes to the hi-tech machines used to fabricate timber and other wood-based products, such as plywood, CLT (Cross Laminated Timber), and LVL (Laminated Veneer Lumber). The reasons for this digitalization are speed and efficiency. The timber construction industry is a highly competitive market and faster and accurate fabrication of products gives companies a competitive edge. However, advances in machinery have not led to significant innovations in the fields of architectural or structural design, which rely on traditional design paradigms. This is partly due to the fact that generally design and fabrication are completely different fields with their dedicated design software and professionals. These fields rarely interact about the possibilities modern fabrication methodologies could offer for the design of timber buildings. Also, as the carpentry profession has been digitalized and automated by machines, the distinction between designers and operators is more and more distinct. In order for innovations to emerge, we need multidisciplinary teams and designers that embrace the new technological capabilities of fabrication. We, as architects heavily involved in the field of digital design, computation, and digital fabrication have sought to find new ways to take advantage of industry’s fabrication machinery. We have conducted explorations in this field through a series of small-scale, freeform timber structures. Some of these have been realized through universities and some through our commercial practice Geometria Architecture Ltd. The unifying aspect in these structures is the use of commercial companies as manufacturers. The absence of available machinery in Finnish university infrastructure has prompted the turn to industry fabrication, but in the long run it has also provided us with valuable lessons and connections. The industry machines are rigidly tailored to the needs of their core businesses. This means that “hacking” of the machines or Figure 1. The straight plywood strips of the Pauhu Pavilion create the appearance of a smoothly curved surface.

133

Rethinking Wood

altering their setups is out of the question, as we have to find solutions that fit to their specific processes and software (Figure 2). The limitations of the industry in what is possible to construct, restricts our explorations in formal and material solutions, but on the other hand makes them readily available for commercial purposes. The explorations, though small in size here, can be scaled up to larger constructions and the possibility to use industry scale fabrication and processes make the transition to larger scale relatively easy from the perspective of construction.

Design methodology The utilization of computation plays a critical role in the design of freeform timber structures, as they mostly rely on non-standard pieces unique in their geometries. The structure is defined through building components, which vary according to the possibilities and restrictions defined by the design itself, structural aspects, as well as the fabrication methods. As a single structure can contain hundreds of individual pieces, their manual modeling is not practical and prone to errors. This means that the component geometry has to be designed as a system, based on local modeling rules and driven by an automated process. Local rules define the geometry of an individual component based on its location and function as part of the global system, i.e. the overall structure or form (Figure 3). Designing a timber structure and describing it as a process through algorithms is more time consuming in the beginning of the design process than traditional manual modeling. The algorithmic process needs to be flexible enough so it can be used in various formations, but also very accurate in generating the 3D model, while simultaneously taking the fabrication restrictions into account. The benefits of defining the design as a process are in the ease of design explorations,

Figure 2. The Hundegger K2, a five-axis milling machine, Figure 3. Freeform timber structures often consist of is a versatile machine for cutting and milling joints for tim- several hundreds of unique pieces. Pictured here are some ber beams and logs. It is used extensively by the Finnish of the pieces for the Sotkanmuna shell structure. log house industry.

134

where different forms and compositions can be studied with minimum effort—accurate and fabrication-ready. Computational design methods and file-to-factory connections are extending the boundaries of what we can design and fabricate. It allows us to manage complex solutions, analyze faster and more accurately, and fabricate details in a mass, customized way. Working within the digital design and fabrication environment, we can directly connect our designs to the computerized fabrication machinery, their possibilities and limitations. Architects become digital craftspeople, developing new forms, structures, and details that are tailored for computerized fabrication.

Design process The design of free-form timber structures begins with the mapping of the available fabrication methods and especially their possibilities and limitations, such as allowed material dimensions and what kind of operations the machines are capable of. When working with wood, we tend to use readily available materials, such as beams and sheets in their standard dimensions. The dimensions of the raw material and the bounds of the production machinery together form the constraints for our design explorations. Within these constraints we are seeking novel solutions for the use of wood by utilizing the capabilities of production machinery to their full potential. The modeling algorithms that define the geometry of the building components are often highly project specific. The material and fabrication restrictions are coded into the design algorithm as geometry generation rules, or they are revealed through analysis of the created geometry. It is a delicate balance on how many rules you want to incorporate in to the algorithm itself and how much is revealed through analysis. The more rules you apply to the design algorithm as restrictions, the more you limit the freedom of the exploration. Material, fabrication methodology, and fabrication limitations become the guiding factors in the design, form creation, and definition of the structure. In that sense, using algorithms in the aid of the design does not differ from the more traditional design processes. But through the inclusion of material and fabrication properties to the design algorithm, they are more explicitly guiding design explorations.

Rationalization and assembly From a structural point-of-view, free forms provide greater challenges and often need optimization in order for the design to be rational, functional, as well as cost-effective. Rationalizing the geometry, distribution, or details can have a significant influence on the cost or viability of fabrication. Sometimes it is an intuitive design decision stemming from experience in the field, and sometimes it is the result of complex rationalization processes. When we discuss rationalization, we do not 135

Rethinking Wood

just mean the optimization of the structure, form, or any other singular aspect of design, but the simplification of the overall process from global to local geometry, and from fabrication to assembly. The rationalization of the global form and local components heavily influence the construction and assembly procedures of the structure. The use of prefabricated and unique components mean that the assembly resembles a puzzle of hundreds of pieces that need to fit each other in a specific way and are often installed in a specific order. The joint often dictates the assembly order. In all of this, clear and comprehensible assembly instructions become vital in construction as they are the document that enables a worker previously unaware of the structure to be able to assemble it. The assembly instructions differ from construction documents as they do not only indicate what the end result should look like, but they are a step-by-step instruction guide on the process. The creation of these instructions is no minor task and they are made with as much precision and thought as the design model itself. Input from the constructor is often needed in order to suit the drawings to their needs, capabilities, and assembly processes.

Challenges As there is no unified approach in industry, the fabrication machinery and their software as well as the machine operators heavily affect the workflow from design to manufacturing. In some cases, the journey from the 3D model to translated machine operations can be quite complex. The way that the model is prepared for fabrication, how data can be transferred and in which file format, vary from case to case. Furthermore, there is also the impact of human factor, as experience and knowledge of the machine operators can vary drastically. Based on our experience, the biggest challenges in the field do not lie in the elements of the design or fabrication, as usually both fields are well knowledgeable in their respective parts. The challenges lie in the information and knowledge transfer between them. Correspondingly, there exists a broad spectrum of machines in the industry and software. Generally speaking, the geometry transfer from design to fabrication is fairly simple. It is the additional meta-information, such as piece code, that still causes issues between software. Through our experience in working with several industry partners, we have gathered different best-practice methodologies to solve information transfer processes. Some methods are direct exports from the design software, as others require interpretation to match machine or software-specific guidelines.

136

Case studies Presented here are four small-scale projects that illustrate the possibilities and challenges that lie in the design and fabrication processes of freeform timber geometries. The projects represent distinct structures and forms, providing insight into many aspects of timber construction. Hila Pavilion, a wooden lattice structure The Hila Pavilion was part of the summer workshop series at the University of Oulu, Faculty of Architecture. The workshop was held in collaboration with the DigiWoodLab project, which aimed to research and develop methods for designing wood architecture using computation and computerized manufacture. The goal of the workshop was to test the computational design, digital fabrication, and assembly processes in their entirety through the design of a one-to-one prototype structure. Hila was constructed on Kiikeli Island, located close to the city center during the summer of 2014. (Figure 6) The prototype pavilion was created in a very short time; taking six weeks from the start of the workshop to the finished structure. The first week was used to develop the overall concept and structure, then four weeks was reserved for the model creation, structural analysis, material acquisition, and fabrication. The sixth week was reserved for assembly on site. Students were given no specific tools, restrictions, or guidelines in the beginning, so the form and structure developed completely from the innovations of the students and through interaction with the teachers. The result was a three-dimensional, timber lattice structure bound to the specific solid shape design. The lattice is connected by simple cross lap joints. The tight slots in joining members together with the bolt connection brace the structure, so no diagonal bracing is needed. The homogenous and orthogonal structure was fairly simple to define and model and had a limited set of design possibilities. This meant that the majority of design explorations were the refinement of the solid shape that defined the bounds of the structure. The challenge was in finding a balance between form and structural integrity. The entire design of the pavilion was defined through a single Grasshopper script, which provided a very flexible exploration process. Structural and fabrication restrictions were revealed through real-time analysis of the generated model, instead of them being part of the modeling algorithm. By keeping the analysis separated, we were able to keep the algorithm and its control parameters fairly simple and less time consuming to construct. All parts of the pavilion were fabricated using a five-axis CNC milling machine Hundegger model K2, intended for milling of timber beams and logs. The Hundegger K2 is generally a versatile machine, used extensively by the log house industry in Finland. Its greatest restrictions lie mainly in the 50 x 50 x 500 mm minimum material 137

Rethinking Wood

Figure 4. The solid shape of the pavilion defines the boundaries for the three-dimensional lattice structure.

Figure 6. View of the finished pavilion.

138

Figure 5. The simple cross lap joint of the lattice structure.

dimensions the machine can fabricate. The restrictions are defined by the “grippers” used to hold and move wood material through the production line. For fabrication purposes, the 3D model needs no special preparations; beams can stay in their original location, no re-orientation or ordering is needed. The 3D model was sent to fabrication in two files; one file defining the solid piece geometries as an ACIS file (.sat) and one DWG file defining the drill holes as lines. Both files can be directly exported from Rhino3D software. The machine operator used HsbCAD to import both files and export corresponding operations to the Hundegger machine. HsbCAD generally understands basic solid geometry as beams and joints. Only the drill holes needed to be manually defined by the operator by using the lines in the DWG file. The main shortcoming of this workflow is that the ACIS file does not support object names, so any predefined piece code does not transfer with the file. During the export process, all pieces receive a consecutive numbering by HsbCAD. The challenge was to translate this numbering back to the assembly drawings. There was no automated way to export this information from HsbCAD into Rhino, so we needed to match the coding manually. The process was tolerable with the relatively small number of pieces in this pavilion. With larger structures, this shortcoming would become a bottleneck in the workflow. However, recent HsbCAD update has provided the possibility to define the solid geometry together with the object name through IFC files. The size of the finished pavilion is 5 x 5 x 4 meters and it was assembled by students in four days. The lap connection was easy to assemble on site and the fabrication accuracy made it possible to erect the entire structure without the use of any power tools. (Figure 8)

Figure 7. False-color analysis of the structure; cardinal directions, beam lengths, and connections.

139

Rethinking Wood

Figure 8. The three-dimensional lattice gives a light, lace-like appearance.

140

Figure 9. Interior view of the Sotkanmuna shell structure.

141

Rethinking Wood

Figure 10. The asymmetric shell structure resembles the shape of a bird’s egg.

Figure 11. The Zollinger-type grid is filled with milled plywood panels.

Figure 12. The 3D model of the entire structure was modeled in Rhino3D using RhinoScript.

Figure 13. Model and coding of the plywood panels.

Figure 14. Individual pieces of the structure with piece code written on the end of the beam. The code makes it possible to find the location in the overall structure.

Figure 15. The pieces were secured during assembly with long screws.

142

Sotkanmuna, a wooden shell structure Sotkanmuna is a gallery space for a video installation in Haltia Nature Center which is located in Nuuksio National Park. The design of Haltia Nature Center was inspired by Finnish mythology. Sotkanmuna, The Egg of the Diving Duck, is part of the mythology that inspired Haltia’s overall design. There was a need to design an asymmetrical dome construction for Haltia’s main exhibition. Computational design methods made it possible to design a cost-effective structure that could be assembled from prefabricated, easy to manufacture, yet unique parts. The structure was built in a relatively short time, along with the rest of the exhibition. (Figures 9 and 10) Sotkanmuna’s asymmetric shell is assembled from around 600 straight wooden beams configured as a so-called Zollinger structure. The 150 x 94 mm glulam beams vary in length between 0.7 and 1.4 meters. The beams are connected with mortise and tenon joints milled into the beams. The grid is filled with CNC-milled plywood panels. (Figure 11) The form was modeled in Rhinoceros as a single surface, where the center lines for the structural grid were mapped onto the surface based on its UV-coordinates. Purpose-built Rhino scripts were used to define the geometries of the construction elements using the guiding surface and structural grid (as beam center lines) as inputs for geometry creation. The structural analysis was done using Lusas software. Analysis was needed in order to make sure that the structure could suspend hanging glass panels for the exhibition and to find the optimal suspension points for those panels. There is no additional ring beam for the suspension points, as the structure above the points also works as a stiffening structure for the suspensions. (Figures 12 and 13) A solid model of the wooden beam model was exported as an ACIS file to HsbCAD. Mortise and tenon joints were defined within the HsbCAD, which then directly exports them as machine operations for the Hundegger K2. (Figure 14) The shell is clad with diamond-shaped plywood panels. The space between beams narrows from outside to inside, as the overall shape is convex. The panels are mounted against wooden plugs inserted into the sides of the glulam beams. Hole positions for the plugs were defined with RhinoScript. There is a 2 mm offset between the plywood panel and wooden beam in the install position. To make installation easier, the panels were designed with narrow pointed tips protruding beyond the offset in the top and bottom corners. This allows for small deviations from the model geometry and also helps to position the panels and maintain the 2 mm offset on all sides. Constructing the shell structure from straight beams enabled a cost-effective production of building components on a CNC-unit that is typically used for milling parts for log buildings. All parts of the six143

Rethinking Wood

Figure 16. There is a contrast between the rough exterior and smooth interior surface.

Figure 17. Pauhu Pavilion was designed and constructed to host various small events and to promote interaction between the citizens of the city of Tampere.

144

meter-high structure could be easily transported and assembled on site. The structure was assembled from ground up, one ring of beams at a time. The mortise and tenon joints are secured with long screws. The opening for the CNC-milled entrance arc was cut after the dome had already been assembled. No temporary support structures were needed during the assembly. (Figure 15) Pauhu Pavilion: a cost-effective ruled surface Pauhu Pavilion was constructed as part of the annual Tampere Architecture week in 2015. The theme of the event that year was interaction, initiating discussion between architects and other citizens of Tampere in Finland. The pavilion functioned as an open-stage for performances and public debates and also aimed to promote forward-thinking ideas about the innovative use of wood in architecture. (Figure 1) The city provided a site for the pavilion. The location was part of a temporary experiment by the city to limit the flow of cars the in the city center by extending the sidewalks. The pavilion was located on this narrow extension strip, wedged between trees and street light fixtures. (Figure 17) The construction budget was minimal and all materials and fabrication services were sponsored. The tight budget and selection of sponsors limited the use of digital fabrication. Plenty of timber and plywood was readily available, but we could only use three-axis plywood CNC-milling. These limitations set the ground rules for the design. The concept of the pavilion is a ruled surface that is carved out of a solid box. The surface is defined by rotating a line along an arc by almost 180-degrees. The resulting curved surface looks complex, but can be defined with straight lines. This is called a “ruled surface.” (Figure 18) With the ruled surface design, it was possible to construct the most complex part of the form from straight pieces of plywood. The parallel plywood strips visually appear as a smooth surface. This innovation made it possible to construct the pavilion on time and with the material and service selection at our disposal. The only parts that needed to be fabricated with CNC were the curved support arched for the plywood strips. The rest of the construction was simple and straightforward timber balloon frame. The dark exterior surface of the pavilion was provided as a ready product, so its fabrication was not part of the process. (Figures 16 and 19) The entire pavilion was modeled in Rhino and the interior part was defined through a Grasshopper script. With the script, it was easy to explore the dimensions of the box shape and the space defined by the ruled surface. A separate script was used to define the locations and dimensions of the plywood strips and their support arches. Due to the shape of the surface, each plywood strip overlaps with its neighbors differently. Some pieces need to be 300-mm wide in 145

Rethinking Wood

Figure 18. The concept of the pavilion is a ruled surface generated by rotation and then cut out of a box shape.

Figure 19. The different structural layers of the pavilion. Figure 20. The widths of the plywood strips were The ruled surface is not part of the load bearing structure. optimized in order to reduce material.

146

order to fill the gap, but in some areas a lot less is needed. In order to reduce the use of material, we optimized the widths of the plywood. Instead of using unique widths per piece, in total eight different widths were selected. We had access to automated plywood cutting machine, which made it simple to fabricate these eight variations. (Figure 20) For the support arches, we created a simple DWG file, where the arches were nested on plywood sheets as 2D polylines. Piece code was written as polylines inside the perimeter of each arc and via its layer it was defined to be engraved on to the piece itself. Some arcs were too large to fit onto a single sheet of plywood, so we created a simple jigsaw puzzle joint, in order to divide the long pieces into several sheets. The separation also provided for tighter nesting, thus reducing material use. (Figures 21 and 22) The construction of the pavilion took eight days from a group of architecture students. It was built off-site in a specialized construction facility and brought to its final location on a truck. After the urban experiment ended, the extended pavement was removed and Pauhu Pavilion was purchased by the neighboring Lempäälä City and moved to its center.

Figure 21. Because the support arches were too big to fit onto single plywood sheet, a jigsaw puzzle joint was used to connect the pieces together.

Figure 22. The straight plywood strips were connected to the support arches with small nails.

147

Rethinking Wood

Figure 23. Joint detail and assembly order of the exhibition wall structure.

Figure 24. Detail of the lap joint. The overlap was 10 mm per joint.

Figure 25. A bench was integrated as part of the structure.

Figure 26. The finished structure.

Figure 27. The arched gateway leading to the exhibition area. The arch was supported by an integrated pillar and a CLT beam.

148

Exhibition wall: an interlocking freeform surface The wall was designed for our fabrication partners Woodpolis and TimberBros, to be part of their joint exhibition stand for the FinnBuild 2016 construction expo. FinnBuild exhibition is the largest expo event in the construction and building services industry in Finland. The joined exhibition stage demonstrates the capabilities and know-how of the companies working in timber and CLT construction. (Figures 26 and 27) The design is meant to be something that catches the eye of the expo visitor in the myriad of bulk construction companies. It should distinctively point out the innovative attitude of the companies involved and the possibilities of modern woodworking machinery. However, the design should not cost much and should be fast and simple to assemble. The FinnBuild expo is a two-day event, with only two days given to the exhibitors to construct their exhibition stands. Our challenge was to design a simple system that could be easily fabricated and assembled, but which would provide enough design freedom to form an eye-catching sculptural piece. (Figure 23) The design is based on two interweaving ruled surfaces, which are defined and controlled by two edge curves. A third set of curves defines auxiliary surfaces that help in functional parts such as the bench. The surfaces are divided to 60 x 60 mm untreated pine beams with 10 mm of interlocking lap joint. With a screw connection with the floorboards, the beams form a rigid triangular shape. (Figures 24 and 25) The system is easy to assemble as long as the pieces are installed in consecutive order. The pieces and joints are similar to those of the Hila Pavilion, so the fabrication process was simple. And the beam geometry was simple to define based on the ruled surface. The model was sent for fabrication as three separate ACIS files; one file representing the components of a single surface. The files were imported in HsbCAD, which automatically numbers the unnamed pieces. The order of the automatic numbering matches the order in which the pieces are modeled, but because the piece geometry was generated in Grasshopper, it subsequently matched the piece’s location in the surface itself. In this way, we could circumvent the challenges in piece coding and assembly instructions. The beams were fabricated with the Hundegger K2.

149

Rethinking Wood

150

Bringing Robotic Fabrication into Practice Léon Spikker

Studio RAP is an architectural design and fabrication firm based in Rotterdam’s Innovation Dock. It was founded in 2015 by Wessel van Beerendonk, Lucas ter Hall, and Léon Spikker, three alumni from Delft Technical University, as a continuation of their graduation research into the application of robotics in architecture. As architects, they believe in the power to deliver a substantial contribution to society by transforming the fragmented analog building process into a digitally-integrated process. An industry where expressive, tailor-made buildings are produced efficiently. Digital fabrication is at the heart of their studio’s approach to architecture. They realize that technology in and of itself is not the goal: it is the application of these methods and technologies to achieve a more expressive and efficient architecture that is important. Their focus on applying these tools in commercial projects has allowed them to test various robotic manufacturing methods in practice: hotwire-cutting, five-axis robotic milling, concrete and clay printing, and timber-assembly. This focus is reflected in their workshop which features 700 m2 of fabrication space with three robotic arms. RAP has a special interest in the way in which computational design and robotic fabrication enable architects to reclaim control over the building process and enforce a higher quality end result. In 2016, Studio RAP launched a spin-off to achieve this goal even faster. RAP Technologies develops robotic fabrication software for manufacturers in the construction industry. Slicing, paneling, or robotic control algorithms that were developed for Studio RAP’s projects are transformed into industry-grade software for techniques like 3D concrete and clay printing. In this way, RAP wants to increase the access other architects and designers have to digital fabrication techniques. In this article, their workflow in relation to two specific timber projects will be described.

Figure 1. Close-up view of Circular Experience wood assembly.

151

Rethinking Wood

Figure 2. Photo of Studio RAP’s workshop and office.

152

Skilled-In Office Context For their first building, Studio RAP was commissioned by the Port of Rotterdam to design and build an indoor office for one of their tenants in the RDM Innovation Dock: a training company for industrial personnel called Skilled-In. The Port of Rotterdam has been pursuing the redevelopment of former industrial port areas into either working or living areas since many larger ships will no longer sail deep into the harbor. At the same time, the city of Rotterdam has been growing fast and is in great need of space for housing and commercial activities. The Innovation Dock, where the Skilled-In Office is situated, is a business complex at the center of the redevelopment of the RDM Campus, a large former ship-building facility turned into a hub for educational institutions and innovative startups.

Design brief Within this business complex, the new Skilled-In Office provides space for four trainers and twelve students for instruction and practical training. Two of the main requirements for the project were a flexible open floor plan and transparency for natural light and connection to the industrial shipbuilding warehouse. The addition of two or three partitioning walls in a later phase was to be made possible.

Design Starting from a flat, rectangular plot of 13 × 9 m, the optimal functional shape for the floor plan was rectangular. To accommodate the program and allow for partitioning later on and to use the least amount of material, the design concept featured a glass wall surrounding the perimeter with a central column that merges into the roof. This column was placed slightly off center to allow for a large central instruction space and two smaller offices on the side. The rectangular outside edge of the roof rests on thin beams that function as a tension ring because of the addition of four tension cables. Since the load transfer was radial from the column’s base, the corners of the office do not contain any columns and are completely transparent. The vaulted ceiling is made from cross-bonded laminated veneer lumber (37 mm Metsäwood Kerto Q) because of its low weight, dimensional stability, and good machinability. The overall shape and curvature of the compression-only vault were derived from digital form-finding calculations using Rhino Vault. The double-curved shape was then divided into triangular elements that fit within the range of the robotic arm and within the size of the Kerto stock.

153

Rethinking Wood

Figure 3. Detail of finger joint including screws.

Figure 4. CNC milling with ABB robot arm.

154

Detail Assembling these triangles into the rectangular panels was based on the grain direction. Because Kerto has a higher tensile and compressive strength parallel to the grain than perpendicular to the grain, the grain of the panel was aligned with the stress lines in the structure to reduce the thickness of the panels. The connection between the triangular panels was designed to reduce in-plane shearing and to simplify the precise placement of panels during assembly. The finger joints were held in place with screws to resist point-loads on the structure during construction or maintenance.

Fabrication process The first step in the fabrication process was the engraving of panel outlines and numbering on large 3 × 1,8 m Kerto stock with a large cncrouter. The panels were then cut out roughly with a circular saw. These semi-finished panels were then transferred to the ABB 6400 robot where the finger joints were milled out using RAP’s own robotic control software. All of the 230 unique panels were engraved, numbered, and milled directly from a central parametric model.

Assembly After assembling the steel structure, the first panels of the column were fixed to a steel base. From there, the vault was assembled panel by panel outward toward the perimeter while being supported by shoring

Figure 5. Section of design, showing the tension ring.

155

Rethinking Wood

Figure 6. Assembly of wood plates.

Figure 7. Photo of Circl – the sustainable pavilion in its context.

156

posts. After the vault was completed, the tension cables connecting the perimeter beams were tensioned until the facade was vertical and the vault reached its correct shape. Wiring, glass, and a climate system were then placed to finish the office. The end-result is a spectacular, thin vault rising from a central column in which the grain pattern enhances the curvature of the vault and visualizes the forces within the structure. The office embodies the ethos of the Innovation Dock, where transparency, connection, and innovation are paramount. In the end, the office is solid proof that digital fabrication technology has reached the stage where it’s accessible enough to enable three recent graduates to design, engineer, and produce something as efficient and attractive as the Skilled-In Office.

Future The Port of Rotterdam has returned to Studio RAP for another project: the redevelopment of the Rotterdam Quarantine Area. Studio RAP will re-use the algorithms and production processes of the Skilled-In project to produce five, 400 m2 open floor plan office buildings. This shows that the potential of computational design and fabrication is not just viable for a single indoor project, but is scalable and repeatable in larger outdoor projects. Recognition and awards Arc’16 Innovation Award Dutch Construction Award for Building Process

157

Rethinking Wood

Figure 8. Production of four seperate panels of 1,200 mm.

Figure 9. Detail of glue-free dowel joint connection.

158

Circular Experience For decades, the large Dutch bank ABN AMRO has been planning to build a pavilion outside its main office, a tall sleek office tower in Amsterdam’s financial district. As one of the largest real estate banks in the Netherlands, it felt the responsibility to take the lead in the transition towards a more sustainable, even circular, future of the construction industry. As part of this vision, they decided to build Circl, a multi-purpose, public pavilion (designed by Architecten Cie) built following the circular economy principles. These principles dictate that through better design, maintenance, repair, reuse, and recycling products, buildings can last longer. This is a fundamental shift away from the traditional linear economy where products are disposed of after use.

Brief ABN AMRO asked RAP to rethink the aesthetics of sustainability (which had quite a dull reputation) and improve a specific part of the project: the underground entrance staircase leading into the pavilion from both the parking-garage and the large main office tower. This subterranean level contains an art exhibition space and meeting rooms for employees. After closer inspection, RAP noticed that the secondary entrance was frequently used by employees but only featured a bare concrete staircase, which didn’t communicate the importance and sustainable spirit of Circl. At the same time, the discovery was made that a lot of waste material remained at the factory where the laminated loadbearing structure had been made. The circular design paradigm dictates that these “material loops” should be closed and that waste should be reused. In close discussion with the bank, RAP decided to design wall and balustrade elements that would improve the acoustic performance of the space, would close the loop on this specific waste-stream, but would also enhance the employees’ experience of the paradigm shift from linear to circular economy.

Material As the main starting point of the design, the waste material was analyzed. There were several thousand larch and pine strips in a ratio of two to one. These strips had been discarded because they were too warped or bent, or because they contained cracks, and couldn’t be processed into glue-laminated trusses. As a first step, to reduce the absolute warping and bending per element, the strips were sorted, transported, dried, and sawn into 30 × 40 mm battens of 300 and 500 mm length to within a tolerance of +/- 0.5 mm. 159

Rethinking Wood

Figure 10. Image sequence of the robotic fabrication process.

160

Design With these material dimensions and quantity as a starting point, a design was made that would adapt to the floor, walls, staircase, and handrail, but would try to be as expressive as possible where it had space to do so. The design contains about 4,000 vertical pine components and 6,000 horizontal larch components. Variation in multiple design variables generated a complex ornament that can only originate from the merging of parametric design and robotic manufacturing. For this purpose, a two-directional stacked component system was imagined, where components would become flush near the handrail of the staircase, but would fan out the further away they got from this constraint. The pattern would fade out along the curved section of the wall in order to end up completely flat alongside the entrance. With a recently acquired second hand ABB6400 robotic arm on a five-meter track and some pneumatic grippers, a first proof of concept was assembled. The battens were tacked together manually in between layers, greatly limiting the production speed and safety of the setup. To maintain the required accuracy and remain within the production process size limits, the design was subdivided into 1,200 mm wide segments.

Detailing After experimenting with tacking and gluing, which would require little peripheral machinery, the final detail was a bit more complicated. To adhere strictly to the principles of circular design, the battens are connected using glue-free dowel joints to guarantee its structural safety as a balustrade. Threaded metal rods are used to post-tension the separate wall and balustrade elements after robotically stacking them. This increased the stability and prevents anyone being able to pull off battens on the exposed sides of the elements. An added advantage was that, by only using glue-free joints and mechanical fasteners, the entire project can be taken apart after its functional life span into two separate material streams: around 10,000 wooden battens, 50,000 dowels, and 150 metal rods.

Testing Although made entirely out of solid wood, which passes the building code’s fire regulations, the Circular Experience had to be tested both structurally as a balustrade as well as for its fire-resistance because of its inherent complexity. It had to comply with the highest fire-resistance class because of its location along an escape route. Since it consists of thousands of unique wooden battens, both its structural performance and fire resistance couldn’t be validated from a computational model. After many trial tests, the pine and larch needed to be

161

Rethinking Wood

Figure 11. Final installation at entrance staircase of ABN AMRO.

162

both impregnated and coated with bio-based fire-retardant to receive a permit. The load testing proved much easier, with the thousands of dowels joints providing sufficient resistance to impact.

Fabrication A small manufacturing set-up was built to perform the following routine thousands of times: A pneumatic feeding-mechanism supplied a pine or larch batten of either 500 mm or 300 mm length (depending on whether the desired batten length was under or over 300 mm long). The robot drills six, carefully positioned dowel pockets into the top of the batten with a flange-mounted spindle. The robot picks up the batten with a flange-mounted gripper and pushes it down onto a fixed spindle for an additional six pockets in the bottom of the batten. The robot then cuts the batten to the desired width on a fixed circular saw. One-by-one a pneumatic dowel-feeding mechanism provides dowels to a fixed position where the robotic arm subsequently pushes the batten onto each of six dowels. With the sub-assembly complete, the robot moves over the track to push the batten into the previously positioned ones. This entire set-up was run from instruction files generated from a single parametric model, generating all of the input and output signals as well as every robot movement. After the robotic assembly process, metal rods were threaded through the stacks manually and tightened to a preset tension. These segments were then checked, packed, and transported from the production site in Rotterdam to Amsterdam for assembly by a local subcontractor.

Conclusion The Circular Experience shows that, even though being a relatively small addition to the Circl Pavilion, sustainability doesn’t have to be dull or boring and can thrive with regard to new technological advancements. It demonstrates the versatility of computational design and robotic fabrication to accommodate the imperfections of reclaimed natural timber to achieve a richness and expression that is uncommon in today’s construction industry.

163

Rethinking Wood

164

Concepts for Timber Joints in Robotic Building Processes Philipp Eversmann

Abstract The ease of access to sophisticated computational technologies normally used for engineering purposes have thoroughly transformed the way architects can design and generate structures. In a first step of digitalization, traditional timber joining methods have been adapted for CNC-fabrication and manual assembly, enabling intricate joinery methods. The next step integrates the digital workflow to the extent of a fully automated assembly using robotic building processes. The potential of these processes calls for a rethinking of traditional connection geometry and requires the creation of specifically adapted joints to accommodate the constraints and precision of kinematic movement. This article gives an overview of recent developments in joining techniques for robotic assembly and investigates two prototypical timber structures, which were constructed in the framework of the NCCR Digital Fabrication at the ETH Zurich. The first structure consists of connections of multiple members with complex intersection geometries, which were joined through an ultrafast-curing resin. The second prototype consists of a double-story timber structure of more than 4,000 individual members with singular face to face connections with a maximum of two members, connected by carbon steel screws. The results are compared in terms of the relationship between computational form finding, joint geometry, connection system, robotic fabrication process, and structural stability. We discuss applications in the building industry, spatial potentials, structural challenges, and fabrication developments and conclude by showing the potential development in future research.

Introduction: computational development of timber structures and joining techniques Computational generation of timber structures The spread and availability of sophisticated computational technology featuring easily usable, adaptable, and extendable software plugins with visual programming interfaces has fundamentally transformed the way architects are now able to conceive and generate their designs. Even advanced engineering optimization methods, such as Topology Optimization (TO) (Bendsøe 1989), Dynamic Relaxation (Day 1965) and Finite Element Modelling are now easily accessible (Aage et 165

Rethinking Wood

al. 2015, Piker 2013). These computational form-finding methods allow architects and researchers to explore completely new geometrical and structural forms, resulting also in novel constraints and requirements for joining techniques. In comparison to other construction techniques, modern timber construction is already highly digitally integrated (Jeska and Pascha 2014). The majority of wood structures are nowadays fabricated using CNC-fabrication machinery, producing joints designed for manual assembly. These types of joints typically only allow for a single possible assembly trajectory (one Degree of Freedom: DoF), limiting assembly errors on construction sites. Robotic building processes allow to further integrate the digital workflow up to a fully automated assembly. Since industrial robots can reach positions and orientations in space at extremely high precision, joints with multiple DoF in assembly can be realized. During robotic assembly, ideally, joints are not completely fixed in one position, in order to leave space for the robot to position and maneuver. This allows for more design freedom and simple cutting procedures but requires a rethinking of traditional joint geometry and the creation of new concepts for structural joining in automated fabrication processes. Research in robotic building processes in architecture has also led to a renewed interest on basic material properties (Krieg 2016), with the idea to also integrate simple timber products such as split panels of random sizes (Eversmann 2018), natural tree branches (Monier et al. 2013), or randomly-sized recycled plates for new envelope structures.1 In the following section, we give an overview of digital joining techniques such as nailing, screwing, gluing, milling, and connecting through spatial elements in relation to typically employed timber products and describe their level of automation.

Current automated joining techniques Nailing Nailing can be very efficiently automated for standard construction applications like plates on timber frame structures (Bock 2007), the fabrication of wood pallets, or variable frame structures (Bolmsjö et al. 2008). In 2015, a highly unusual roof structure was realized at the ETH Zurich (Apolinarska et al. 2016, Apolinarska et al. 2017). Here, a robotic portal system automatically cut thin, solid timber beams to length and assembled them in their specific positions through an automatic nail gun mounted on the robot end effector. On the case-study project referred to in this article, an automated staple fixation system was used to fix variable timber panels on the facade of the structure (Eversmann 2018).

Mine the scrap project, certain measures, online resource: http://certainmeasures.com/mts_installation.html accessed 14.01.2018. 1.

166

Screwing Automated fastening of screws is more complex than nailing due to the loading and pressure that needs to be applied during the process. It is nevertheless already a standard process in product manufacturing (Saito 1987) (Dhayagude et al. 1996). Recent schemes allow to continuously control the screw aliment stages to avoid procedural failures through visual feedback techniques (Pitipong et al. 2010). It is now even possible to fasten screws in moving objects (Berger et al. 2017) through adaptive control strategies. In the timber construction industry though, those techniques are still rarely used, but could potentially be employed for a large range of timber products and architectural applications. Gluing In industrial manufacturing as in the automobile industry, gluing is nowadays a widely automated process (Wagner 2010). Fully automated adhesive dispensing systems have already existed since the early eighties (Westermann 1984, Geisel et al. 1987). In architecture, research on the assembly of glued wooden structures was undertaken for a combined automated assembly and adhesive application of custom shaped walls (Oesterle 2009) and roof structures (Helm et al. 2017) using solid timber slats. Glued-in steel rods first studied in the 1980s (Möhler and Hemmer 1981), now also used for the stiffening of laminated timber beams (Steiger et al. 2015). In research for automated assemblies at the ETH Zurich (Helm et al. 2017), this was evaluated but not attempted for integration due to its automation complexity. Milling CNC milling of joints is widely used for making timber roof structures. Nowadays though, automated production lines are mainly designed for producing joint types for manual assembly. Robots have also been used for the repair and reproduction of old joints (González Böhme et al. 2017), and also for recombining and connecting natural branches through robotic milling (Mollica and Self 2016). CNC-milled tightly fit joints can be also assembled by robots, but require complex and extremely stiff setups for holding the connecting pieces during assembly (Robeller et al. 2017). In the production of timber frame walls, robots have recently been integrated in industrial production lines.2

Weinmann Fertigungslinie, online resource https://www.youtube.com/watch?v=a6pYd0_F5ns, accessed 13.1.2018.

2.

167

Rethinking Wood

Spatially connecting Spatial connectors using manually formed timber pieces are historic joining elements in timber construction. Here they are described in relation to nails or screws to larger connecting pieces such as plates, wedges, and other volumetric joining elements (Gerner 2000). Sophisticated examples can be found in ancient Chinese (Fang et al. 2001) and Japanese (Sumiyoshi and Matsui 1991) timber architecture. For the exhibition pavilion Futuropolis by Daniel Libeskind, dovetail connector pieces made of aluminum were used in an automated production process to connect plates at various angles (Scheurer et al. 2005). Research projects also show the possibility of using CLT-plates as connectors in sawn joints (Fischer et al. 2012). Pre-dried timber connectors that expand when assembled can also be considered, but have the disadvantage of requiring a relatively long time until they achieve final stability. Scope of research This article describes research on robotic assembly and joining techniques in relation to computational form finding of timber frame structures constructed at the ETH Zurich in the framework of the NCCR Digital Fabrication. In section two, we describe two collaborative research projects, both involving structural form-finding and optimization techniques as well as gluing and screwing methods adapted for robotic assembly processes. In section three, we analyze and compare these projects in terms of the relationship between joint geometry, connection system, robotic fabrication processes, and structural stability. We conclude, in section four, by discussing applications in the building industry, spatial potentials, structural challenges, and fabrication developments, and show future research potentials.

Figure 1 a. Design space with given set of loads, boundary conditions, nodal points, and resulting topology optimized structure as connection diagram.

168

Integrated computational form-finding and robotic fabrication of timber structures Robotically assembled joints for topology optimized structures This project was the result of collaborative research between the ETH Zurich (NCCR Digital Fabrication, Gramazio Kohler Research), Aarhus School of Architecture (Asbjorn Søndergaard), and Israel Institute of Technology (Oded Amir) (Søndergaard et al. 2016). The aim of the project was to investigate an integrated computation and robotic fabrication workflow using Topology Optimization (TO) techniques for spatial timber structures. Computational Approach TO (Bendsøe 1989) is a mathematical method that is used for optimizing material distribution within a given set of loads, boundary conditions, and constraints. The resulting geometric models are usually complex and continuous free-form shapes, which are difficult to apply to most architectural construction projects. Therefore, TO based on linear elements with a predefined number of possible elements and connection types within ranges of cross-sections was investigated. We used a “ground structure” approach (Dorn 1964), which discretizes the design space using a fixed set of nodal points, which are connected by potential linear elements. The TO algorithm calculates the force-flow of all possible connections in this three-dimensional grid of nodes with additional forces and foundations. In multiple iterations, over-stressed cross-section areas are gradually increased, while the under-stressed ones are decreased. Unnecessary elements are eliminated by assigning them a zero-cross-section area. The TO Model for our prototype resulted in thirty-four timber slats with three different cross-sections (Figures 1a and b).

Figure 1 b. 3D print with corresponding cross sections (right). The result was reached by TO of 1,711 possible connections of a 5 × 5 × 5 3D grid with 3 supported nodes and an eccentric 5 kN single point load.

169

Rethinking Wood

Figure 2. Principle of a multi-member joint cast procedure: a) Timber elements with perforated faces are assembled using distance holders and sealing tape. b) Ultra-fast curing adhesive is injected through predrilled holes. c) This results in a customcast epoxy connection, eliminating small assembly tolerances.

Figure 3. Prototype of a seven-member joint and generated cutting planes.

Figure 4. Final prototypical structure (image courtesy of Michael Lyrenmann)

170

Joints For the connections we used an ultra-fast curing (5–10 secs.), two-component adhesive (Zock et al. 2014), developed during a research project on reciprocal timber structures with singular face connections (Sigrist and Zock 2016). If more than two members are connected, a geometrically complex intersection geometry results. Usually these types of joints are realized through custom-made steel parts. Our idea was to robotically assemble the elements, fix them temporarily into their position, seal the joints outer borders, and then cast the entire joint through injecting the adhesive through a predrilled hole (Figure 2). Initially, each of the joining faces was additionally perforated for a stronger adhesion. Fabrication Process For fabrication, we used a standard industrial robot arm with a parallel gripping device and a construction circular saw. The robotic process consists of the following steps: First, a timber slat was gripped. Then, the robot performs all three-dimensional movements for cutting, while the saw can remain in a fixed position. After the robot reaches the final assembly position, multiple beams are connected following the joining technique explained in figure 2. A prototype of a seven-member joint and its cutting planes can be seen in figure 3. The final prototype (Figure 4) was tested under point loading with a numerically controlled tension valve fixated to the employed eccentric load of the TO procedure (Figure 1), to analyze the precision and validate the computational approach. We recorded maximum displacements of around 15 mm within 5 kN loading. Under a 13.6 kN load, the structure still held due to rupture of one of the foot metal braces. This nevertheless indicates that the joints performed overall within the expected range of the computational model.

171

Rethinking Wood

Spatial logic

Assembly logic

Figure 5. Left: Geometry generation: a cuboid serves as basic element multiple geometrical connections and permutations are possible within its faces. Right: Spatial module design: Size, form, and sequential order of the spatial modules correspond to functional, fabrication, and assembly constraints.

Figure 6. Structural design of double-story structure of dimensions of around 8.5 × 5 × 7m (h) Left: FE- analysis showing resulting maximum utilization of members in relation to wall openings. Right: Orientation of bracings following stress lines.

Figure 7. Left: Types of full-threaded carbon steel screws used for the joints. Right: Calculation of screw type, length, and angle depending on corresponding geometry and material thickness.

172

Robotically assembled joints for multi-story timber structures This case-study project consists of the first ever double-story robotically assembled timber structure, which was realized in the course of a master’s program in Digital Fabrication at the ETH Zurich (Eversmann et al. 2017). A complete digital design and fabrication workflow of large spatially fabricated modules of more than 4,000 different beams with an integrated timber envelope was realized. Computational Approach Our computational model was directly connected to finite element analysis software (karamba3d)3 using standard values from Swiss building code for loads and safety factors. We integrated custom routines for optimization for the defining load cases and cross sections. The geometry is based on an adjustable cuboid, which can be multiplied to translate seamlessly between different functions as wall, slab, staircase, balustrade, etc. (Figure 5). Similar to the topology optimization procedure described in the previous section, the nodes of each cuboid can be connected in multiple ways. This was used to optimize the diagonal bracings predominantly for compression forces following the principal stress lines (Figure 6 right). Additionally, we performed a custom scripted cross-section optimization, which was able to reduce material by more than thirty percent. Joints Each of the joints had a slightly different orientation and intersecting geometry. We developed a custom arrangement routine which computationally aligned the joints between the members so that only joints between two members occur. Since screws are most efficient when aligned perpendicular to the fiber direction of the beams, we calculated the length and orientation of the screw for each joint, considering the beam’s thickness and joint angle (Figure 7). Through deconstructing the vectors of the forces that act on the screws in both the shear and axial components, we could verify whether the load capacity of the screws was not overreached.

Fabrication process For fabrication, we decomposed the whole structure into a series of spatial modules, responding to functional, fabrication, and assembly constraints (Figure 5 right). Therefore, a custom analysis of robotic reachability and kinematic motion planning was performed and opti-

3.

www.karamba3d.com

173

Rethinking Wood

mized for each of the modules. Even though each of the robotic movements is unique, we were able to create a repetitive and extremely robust protocol of robotic procedures. This was enabled through coding all information needed for fabrication (as cutting, drilling) in each of the computational beam elements. In our robotic setup, we integrated a fully-automated custom CNC circular saw (Figure 8). For assembly, we stored information on the geometric orientation through a classification into families for each element, in order to automatically deduct the best assembly approach direction in the structure. The resulting building modules were assembled manually on site with steel bolts. The process resulted in a densely branched foam-like timber structure (Figure 9), in which structural and fabrication constraints were negotiated in optimizing the overall geometry, length, and orientation of bracings. Even with thirty people on the top floor, the structure was able to remain within calculated maximal deflections.

Comparison of computational systems and performance Both case-study projects were based on a nodal geometric generation system. While the TO-project described in section 2.1 allowed for generating spatial connections in a full 3D grid, the double-story structure described in section 2.2 used a far more restrained procedure through an outer and inner driving surface which the cuboid nodes and their spatial connections had to follow. This obviously generated a far more regular and controllable structure. In terms of fabrication, this enabled a robust robotic assembly procedure through generating a familybased assembly computation approach through the similarity of geometrical orientations. In the TO-project, assembly sequences of each

Figure 8. Robotic Fabrication Process: A timber slat is CNC cut and robotically assembled in the truss structure.

174

member had to be separately computed. Regarding the structure, the TO-structure was only calculated for final stability, which meant that some temporary fixation elements had to be used during fabrication. The cuboids of the double-story structure were, in contrast, always able to support themselves also during each fabrication step without any temporary holding structures. In the TO-structure, gluing could accommodate larger fabrication tolerances through the casting procedure. Screwed connections need a high fabrication precision of the connecting faces and their performance is directly related to the angle between the screw and the fiber direction of the connected members. Consequently, in an irregular structure, not all joints are able to transfer the same loads and one must assume a certain structural redundancy. In general, screwed connections have been extensively studied, which makes it possible to rely on well-established strength values, while especially long-term behavior of glued connections. The behavior of multiple interrelated joints still need to be studied further.

Conclusion The research presented in this article shows methods for creating techniques of joining geometrically, highly variable connections in robotically assembled spatial timber structures. For a high concentration of bars at individual nodes, we developed a custom casting technique of an ultra-fast curing adhesive. For the connections of the double-story structure, only one screw with custom calculated angle was needed for each element and clean and simple singular face-to-face connections were enabled through de-spacing the intersection geometry. Both structures demonstrate a consistency between the predicted

Figure 9. Interior views of the upper floor. Image courtesy of Kasia Jackowska.

175

Rethinking Wood

Figure 10. Exterior view of the double-story structure at the ETH Zurich.

176

structural performance and physical testing. Results show that the geometric variability necessitates adaptive feedback processes to cope with material behavior during the assembly steps. These adaptive robotic processes can be used in combination with relatively unprocessed material of variable sizes to limit material waste. Also, both joining techniques were not yet fully automated. An industrial automation for these types of joints requires a robotic setup with multiple arms which can hold, seal, drill, or fasten elements with synchronized kinematic movements and operations. This requires further research in the connection technology to eliminate structural redundancies and safety factors for the structural detailing. In the future, this could enable an approach to timber construction similar to 3D printing, in which material is dispensed and assembled only where it is needed, regardless of its geometric orientation or traditional type of connection, leading to extremely material efficient, lightweight structures. Acknowledgements The TO-research project was performed within a research exchange between the ETH Zurich and the Aarhus School of Architecture in collaboration with the NCCR Digital Fabrication, Gramazio Kohler Research, and the Israel Institute of Technology. The contribution of the Aarhus School of Architecture was enabled through the generous financial support of the Danish Ministry of Higher Education and Science under the Elite Research Travel Grant program. The doublestory case study project was realized in the framework of a Master of Advanced Studies class on digital fabrication with the students Jay Chenault, Alessandro Dell’Endice, Matthias Helmreich, Nicholas Hoban, Jesus Medina, Pietro Odaglia, Federico Salvalaio, and Stavroula Tsafou. Both research projects were supported by the NCCR Digital Fabrication, funded by the Swiss National Science Foundation SNSF (Agreement # 51NF40-141853), and build directly on research findings and developments from the NRP-66/ SNSF research project “Additive Robotic Fabrication of Complex Timber Structures,” established in collaboration between the ETH Zurich, Bern University of Applied Science, and Nolax AG. We would like to thank the companies Schilliger Holz AG, Rothoblaas, Krinner Ag, ABB, and BAWO Befestigungstechnik AG for their generous support. We would also like to thank Philippe Fleischmann and Mike Lyrenmann for their advice and continuous efforts for the robotic setup and Dr. Jan Schikora for reviewing this article.

177

Rethinking Wood

178

Joyn Machine: Towards On-Site Digital Fabrication in Bespoke Woodwork Simon Deeg and Andreas Picker

Introduction We live in a world that has changed rapidly over the last centuries and decades. Hiring a carpenter for individual and personalized solutions for all sorts of woodworking tasks was a common scenario only one hundred years ago. Today, woodwork is, more than ever, dominated by industrialized mass-production. With digitalization though, things are once again changing. In this essay, we would like to introduce our project Joyn Machine, which we see as being part of this change. Joyn Machine is an interactive tool that enables the design of wooden framework constructions and their semi-automated production. The underlying principle is smart simplification of otherwise complex processes, thereby taking mobile and fast production processes of wooden structures to the next level. This, in turn, leads to a new dimension of scalability; from designing to erecting wooden structures on site. Joyn Machine achieves this by the computer-aided milling of wooden slats to create joints that can be assembled by hand. This makes it possible to create simple furniture and smaller wooden structures as well as architectural concepts on a larger scale. A design that has been created and tested can then be shared online and is easily reproduced with any Joyn Machine worldwide. In a possible future, designs and construction plans could be selected from a large online database or created individually, thereby encouraging the evolution of the consumer to “prosumer.”

Where we come from Starting out as designers, we already ventured into the areas of rapid manufacturing and FabLabs (fabrication laboratories) during our studies at Folkwang University of the Arts in Essen, Germany. Back then, we slowly but steadily moved into a multidisciplinary field between interaction design, industrial design and architecture. On the way, we gained experience in 3D printing and other technologies connected to the “maker movement.” This eventually led to a collaborative diploma project that incorporated all these aspects. Being early adopters of these technologies, our thinking and approach to the relationship between design and the production of goods was lastingly shaped by these experiences. All along, we were fascinated by the maker movement’s notion of enabling people. Our vision with Joyn Machine is focused on this idea: Building a tool that enables a Futurium Pavillion, Closeup of Joint 179

Rethinking Wood

wide-range of people, regardless of their woodworking skills, to design and assemble large-scale wooden structures. Additionally, our idea was clearly inspired by the concept of desktop 3D printers. We had witnessed the technology opening up completely new opportunities for the individual and changing its role in a thoroughly industrialized economy. This notion is closely linked to the concept of the “prosumer,” first introduced by Alvin Toffler in 1984 (Toffler 1984). Toffler defined a prosumer as someone who blurs the distinction between the “consumer” and the “producer” (Toffler 1987). It means democratizing design, production, and construction of many goods. We are fascinated by the idea that FabLabs could have a significant impact on our future. We support the idea that the concept could eventually “re-unite” the virtual and the physical world by enabling many people to produce many goods through sharing designs digitally and producing locally (Gershenfeld 2007). For us, this means making industrial methods of production available to a wide base of people who are interested in creating whatever comes to mind. In the case of Joyn Machine, this means using an existing technology in woodworking and combining it with a purpose-built software solution. The “Abbundmaschine” (joinery machine) has been used on an industrial level since the 1980s. These large, very expensive machines use CNC-milling and require experts’ knowledge (see for example Bock et al, 2015). Ideally, Joyn Machine will lead to collaboration between humans and machines. Regarding this aspect, we are influenced by Richard Sennett’s idea of encouraging “die denkenden Handwerker” (thinking craftspeople). He claims that they should not see machines as competitors, but rather as “suggestions” that can lead to results that would otherwise not be possible (Sennett 2009). To sum up, the people and ideas that inspired us led to the following goals we want to achieve with Joyn Machine: • Transferring experience and traditional knowledge from the area of crafts into the future and therefore preserving it • Pushing the boundaries of how a highly sustainable material such as wood can be used as a construction material for the future— merging high and low tech • Enabling easily feasible solutions for individual construction needs • Enabling quick, cost-efficient and eco-friendly production • Reducing the value and cost of the individual piece of woodwork since our digital approach enables the production of copies anywhere, anytime • Creating a digital format or platform that allows people to share designs, constructions, and architectural concepts 180

Description of project How it came about Over the course of the year 2013, while completing our studies at Folkwang University of the Arts, we started thinking about the idea of a tool that would be able to create wooden joints easily. A while later, we teamed up with Bau Kunst Erfinden, (Built Art Invention) a trans-disciplinary research platform at the University of Kassel, BZT Maschinenbau GmbH, a producer of industrial milling machines near Bielefeld, and Feld, a studio for digital crafts. Together, we applied for project funding from “ZIM-Zentrales Innovationsprogramm Mittelstand,” an initiative by the German Federal Ministry of Economics and Energy. ZIM provides significant funding for R&D-projects that have small and medium-sized businesses (SMBs) among their project-partners. Our ZIM-project ran from July 2014 to July 2016 and aimed to explore and conceptualize ways to combine traditional engineering for woodwork with new digital tools. After the ZIM-project had ended, we, at Studio Milz, decided to continue working on the topic on a fulltime basis and make Joyn Machine a reality. Together with our interdisciplinary team of an mechanical engineer, a software developer, an architect, an industrial designer, an interaction, and a parametric designer, we re-developed software, hardware, and the process behind Joyn Machine during the past two years. We have now produced a fully functional prototype that is able to demonstrate the full potential behind the concept. How it works Joyn Machine follows an “all-in-one concept” covering the entire production chain from designing a wooden structure, to processing the raw material (wooden slats), to manufacturing the wooden parts that can then be assembled into the designed structure. The idea of providing one tool that combines all these skills is achieved by closely linking software and hardware components, but also by the construction of the hardware itself. With a weight of about one hundred kilograms, Joyn Machine is also portable. At the core, it is a three-axis milling machine with a rolling conveyor for the y-axis positioning the wooden slats, including mechatronic, automatic calibration. In practice, this means that the user will take a wooden slat and “feed” it to the machine which then completes the task. The milling-engine has a power-output of 1 kw and can use milling-tools with an operating width from 6 to 12 mm. The machine is completely cased-in to protect the user from noise and it has a built-in solution for the extraction of waste using an external heavy-duty vacuum cleaner. The machine uses a bespoke motion controller circuit board and generally features fully integrated electronics. On the software side, the focus is clearly on user-centricity and “app-like” interaction and 181

Rethinking Wood

Description of the production process.

182

the possibility to control the machine via a built-in tablet computer, which includes a live video of the milling process from the inside of the machine. How it can be used The user flow, i.e. how the user can work with the machine, was the crucial factor for our design of the machine. This process has four main steps: • The design-process for the wooden structure to be built • Giving Joyn Machine detailed parameters on how to mill the joints • Providing Joyn Machine with information about the raw material (wooden slats) • Putting Joyn Machine to work Essentially, this process is pre-defined within rather narrow boundaries. But despite output or, in other words, the wooden constructions being limited in terms of creative possibilities, the clearly prescribed approach leads to almost guaranteed success: Even a person who has little or no experience with woodwork can create functioning structures that serve their purpose. Step 1. The design-process Joyn Machine is able to cater to the specific needs of various types of users. We see three main scenarios for step number one, the design process. Depending on the scenario, this will vary: a) User without computer-aided design (CAD) knowledge or technical craft knowledge For the layperson, there are two different options to determine a design for his or her wooden structure. Firstly, it will be possible to download a ready-to-use design from an online-database. Secondly, he or she could use a ready-made design that allows for parametric design variations, e.g. a bench where the user can choose between different numbers of seats, but cannot change the basic design. b) User with CAD knowledge as well as construction skills The advanced user, on the other hand, can make full use of his skills while at the same time making use of Joyn Machine’s potential to drastically simplify the workflow. Using CAD-software, he or she can create individual designs. Of course these have to consider the flexibility that is technically achievable with the machine and they have to remain within the construction logic. For this approach, the construction’s skeleton needs to be created in any CAD-software, e.g. SolidWorks, Rhinoceros, SketchUp, or Fusion 360. This model is then analyzed with a custom-made software 183

Rethinking Wood

which automatically translates the skeleton’s intersections into wooden joints that Joyn Machine is able to mill. This is undertaken in a 3D environment, enabling a first look at the planned wooden construction. At this point, it is also possible to simulate and define assembly sequences. This allows for marking the parts that will be milled later on, in order to simplify assembly. All the joints suggested by the software can be edited or a different type of joint can be chosen by the user.1 Once this is completed, the design including the joints can be exported as a JOYN-construction file (*.joyn) that is computable by Joyn Machine. It is also possible to export the design to other 3D formats (*.step, *.iges etc.) for further tasks such as load analysis or renderings. c) Craftspeople skilled in constructing, but without CAD knowledge For users who prefer to use Joyn Machine in a more manual mode, we have created the simple interface “part maker” that allows for single parts and joints to be created outside of a complex CAD-environment. Part maker enables the user to arrange cuts and the various possible types of joints freely on the wooden slat, effectively allowing for the construction of complex wooden structures without CAD-skills. One could say that in this scenario Joyn Machine can become a sophisticated tool for the traditional carpenter who is able to mentally calculate what others do in CAD-software. Also, “part maker” is a useful tool for quickly producing replacement parts. For either of the three scenarios, Joyn Machine will automatically calculate the following parameters: approximate time for milling, raw material needed (= cost of wooden slats), total weight of construction, and number of joints. Step 2. Giving Joyn Machine detailed parameters After transferring the *.joyn-file to Joyn Machine, the user is still able to choose between options that are specific to the joints that are going to be milled. This step also serves as an optional check for the less advanced user. The expert though, can still decide at this point where he or she wants to go ahead with the construction. It is for example possible to choose between different “joint-catalogs”: The first catalog will lead to a lower complexity of joints and at the same time to a higher production tolerance. This, in turn, enables a shorter milling-time and easier assembly as it is possible to join the slat-parts in both directions. The second catalog leads to more complex joint-geometries with lower manufacturing tolerance. Here, the user will get traditional jointtypes like dovetail or pin connection, simultaneously translating into a longer overall production time. In the future, this step that is currently undertaken with the mentioned software-script could be achieved by an online-tool. In this way, the user would simply be able to upload his CAD-design and edit the joints and assembly sequence in a web-browser.

1.

184

Step 3. Providing information about the raw material The third step consists of giving Joyn Machine information about the technical properties of the wooden slats to be milled: • The length of the wooden slats can be set freely, from 120 mm upwards. • The properties of the wood itself, especially if it is soft or hard wood. For example, the specific rigidity of beech, spruce, or fir can be selected. • The quality of milling can be selected. The user can either choose higher speed with regular quality (leading to fully functional joints) or slower milling for precision ornamental joints. Feed rate and speed do not have to be set by the user. However, fine adjustment is possible. Step 4. Putting Joyn Machine to work Last but not least, the milling can start. Apart from switching Joyn Machine on and adjusting the conveyer to the size of the wooden slat (a manual process), the machine is entirely controlled via a mobile device. To make this possible, Joyn Machine provides its own Wi-Fi which enables a connection with any tablet or smartphone. The user can then control, calibrate, and service Joyn Machine with a concise and user-friendly interface on the mobile browser of the device. One of the major characteristics of this interface is a clear focus on cooperative production or, in other words, on Joyn Machine and the user working hand in hand. Once a structure is selected, the user is guided through the individual production steps within a conversational user interface. This is almost like a “chat-sequence” that shows what the machine is currently doing and which subsequent steps need to be taken by the user. On the other hand, calibrating to the individual slats and the milling process itself are fully automated. The user only has to feed new slats to the machine, turn these if necessary and remove the processed material. As explained above, the actual function itself does not require any skills in specific crafts. An integrated camera inside the machine allows the user to closely survey the milling process and intervene if necessary. For optimized health and safety, the entire milling process is covered, and the user is protected against noise and dust. After the milling process is completed, the user receives a complete kit of their own design. With a digital assembly plan on a mobile device or with a printed plan, the user can now start assembling the structure. In the future, it will also be possible to print a number or QR-code on every part that is milled, using an inkjet printer that will be added to the Joyn Machine. This will further facilitate the assembly process. 185

Rethinking Wood

186

Output and use cases The remaining crucial questions must be: Where can Joyn Machine be applied, how was it already used, and what is its potential for the future? One of the remarkable characteristics of the concept is the broad spectrum of constructions that are achievable, which range from small pieces like seating furniture or frames for tables etc. to complex architectural constructions that could include bridges or even entire buildings. So far, we have mainly worked with a limited “joint catalog” allowing for joints that create angles of 45, 90, or 180 degrees. We are currently working on implementing further angles for different types of joints that can be created: corner-joints, tee-joints, extension-joints, cross-joints, end-joints, and knot-joints. In the current version, Joyn Machine can process slats with a profile between 30 × 30 mm and 60 × 60 mm. A Pavilion at Futurium The first large-scale construction that was ever created with Joyn Machine is a pavilion that will be part of an exhibition at Futurium in Berlin. Futurium is a building in the heart of the city associated with the German Federal Ministry of Education and Science, which aims to explore future technologies and concepts. In Futurium’s LAB, the exhibition “Future Architecture” will open in Spring 2019. Studio Milz has created the mentioned pavilion that was showcased together with Joyn Machine in June 2018, when Futurium opened temporarily to the public for “Werkstattwochen” (workshop weeks). For Studio Milz, the goal was to deliver a proof of concept, showing Joyn Machine’s capabilities in rapid manufacturing of a complex wooden structure. At the same time, the project was an initial opportunity to test interaction and user participation (with the guests at Futurium) in an authentic scenario and context. For the project, we teamed up with architect Patrick Bedarf who contributed the architectural concept for a large pavilion and product designer, Dustin Jessen who designed a series of small seating furniture. Once the exhibition opens, the pieces of furniture will be able to be produced by participants of the workshops offered to Futurium’s visitors. The pavilion has the following specifications: • Dimensions: 371 × 320 × 565 cm • Weight: 160 kg • Production time: 25 hours • Part count: 235 • Joint count: 505 • Different types of joints used: 9

Joyn Machine Pavilion at Futurium.

187

Rethinking Wood

Specifications of seating furniture (medium-sized bench for two to three people): • Dimensions: 38 × 120 × 45 cm • Weight: 5 kg • Production time: approx. 45 minutes • Part count: 15 • Joint count: 24 • Different types of joints used: 10 The project at Futurium has already shown Joyn Machine’s capability of manufacturing fast, but also its reliability: Even with a complex structure like the pavilion, there is no need to worry about the question whether the design will work out and look like planned. The user process with Joyn Machine almost guarantees success.

Looking ahead Moving forward, we would like to use Joyn Machine for structures on an even larger scale—possibly even an entire building. It would also be interesting to utilize the concept’s capabilities of extreme speed (compared to traditional woodworking constructions) in the sense of public interventions. For example, complex wooden structures could be created in a very short period of time and enable new forms of architectural interaction with the public space. On the other hand, we are very interested in the concept of building a community around the idea and the hardware. We believe that tapping into the collective creative potential of people all over the world with an online-community and combining it with Joyn Machine’s potential for low-cost manufacturing will lead to fascinating projects. One opportunity could be to rent out the hardware and enable people to use Joyn Machine where needed.

188

Conclusion To conclude, we would like to highlight the substantial interdisciplinary character of the Joyn Machine project. For several specialized aspects, we teamed up with experts from various fields: interaction design, architecture, product design, mechanical engineering, software engineering, electrical engineering, process design, parametric design, and carpentry. In our opinion, this indicates a new and interesting way for woodworking. It is a good example that shows it is possible to embrace traditional concepts (like the traditional types of wooden joints that Joyn Machine is able to mill) and at the same time extend the boundaries of what is possible by employing today’s digital technology. At the same time, we see open questions concerning the concept and the machine itself: • To what extent should we “take away” tasks from the user? To what extent should the technical process occur in a “black box”? • Reducing complexity for the user will inevitably lead to decreasing possible achievable outputs with Joyn Machine. Do we therefore need different versions of hardware for the expert and the layperson? • How does our technology influence the work of craftspeople? Will it reduce demand for their skills? On the other hand, we believe that Joyn Machine will also be a tool for showcasing craftsmanship and the quality of traditional wooden joinery, which is so rare nowadays. This in turn could lead to new demand in these methods. Lastly, Joyn Machine is part of our broader vision. We believe in a future that allows people and machines (plus digital fabrication tools) to cooperate and thereby produce and “materialize” many products they need themselves, regardless of whether this is produced at home or in a mobile manner. Joyn Machine is one of several projects worldwide that embrace this approach. We are eager to explore a world where many people have access to our tool and to the necessary raw material: wood.

189

Rethinking Wood

190

New Materials and Applications In early man-made shelters and dwellings, wood was somewhat used as an objet trouvé in the form of branches ripped off by storms or washed ashore as driftwood.1 Today, the use of naturally grown and cultivated wood2 is enjoying a renaissance. Moreover, as this book’s final part shows, it has been taken to new levels. Promising developments, however, are likewise looming in the field of engineered wood, which is currently the dominant form or type of timber material used in the building sector.

In the permanent exhibition “From Primitive Hut to Skyscraper” at Deutsches Architekturmuseum Frankfurt, the primitive hut—or Urhütte—is composed of branches and leaves. 2. In this context, the notion of cultivation expands beyond that of silviculture. 1.

Wooden raw material has various possibilities for further processing. Photo: Eeva Suorlahti 191

Rethinking Wood

The engineering or reorganization of naturally grown wood into wood-based and wood-derived materials and products has a long history. One of the earliest examples of engineered wood is plywood, which was already being used as early as 2600 BC in ancient Egypt (Victoria and Albert Museum, London). With very few exceptions,3 contemporary engineered wood products generally depend on the use of adhesives. As pointed out earlier by Gerhard Fink and Robert Jockwer in part two, glued-laminated timber, for example, contains around three to five percent glue, whereas particle board contains between seven and fourteen percent glue. The use of adhesives makes the recycling of engineered wood products rather challenging and can also cause fumes that are potentially hazardous to health. The dependency on such adhesives hence weakens, or at least challenges, the promise of sustainability that is associated with using wood or wood-based products as construction materials in architecture. Numerous researchers (Venla Hemmilä et al. 2017) are currently working on bio-based, environmentally friendly glues that are both structural and suitable for mass-production. If they prove successful, these new adhesives will eventually diminish the current problems described above. Although working on different scales, which within the framework of this part and in a simplified manner could be referred to as nano, micro, and macro, the examples shown here follow the classic approach behind engineered wood products, namely that of comminuting and reassembling wood. What makes these examples stand out is that, within certain limitations, they can in principle be produced without adhesives. In the first chapter of this part, Heidi Turunen and Hannes Orelma elaborate on possible applications of cellulose and cellulose nanomaterials. Together with lignin and hemi-cellulose, cellulose is one of the main components of the natural composite material wood (Douglas J. Gardner 2002). In order to extract it, wood chips are transformed into pulp by means of mechanical, thermo-mechanical, or chemical procedures (Heli Kangas 2014). Whereas cellulose is mainly used in paper and cardboard, cellulose nanomaterials open up new fields and potential applications. In their essay, Turunen and Orelma focus in particular on new applications for cellulose nanofibrils (CNF), which represent one of the three main classes of cellulose nanomaterials (Heli Kangas 2014). Pure CNF can be used, for example, for the fabrication of films,

Very few approaches rely on mechanical connections, such as the historical arched girders developed by Armand-Rose Émy, and contemporary wood products connected by dowels, such as the so-called Diagonaldübelholz, a product of Holzbautechnik Sohm. 3.

192

whereas the combination of CNF with sawdust produces a castable material mixture. The authors show that some cellulose and cellulose nanomaterials could also have a future in the area of 3D printing. Wood Foam is a new wood-based material developed at the Wilhelm-Klauditz-Institut WKI, the Fraunhofer Institute for Wood Research in Braunschweig, by researcher Frauke Bunzel. The inherent bonding capacities of wood are crucial for the production of Wood Foam. The first phase of the production process is quite similar to that of the previous case, with the difference being the length of the resulting elements. Their length is between 200 and 800 µm, and hence not in the nano but in the micro-scale. The result of the first phase of the production process is a diluted fiber suspension, which can then be increased in size with the help of tensides, such as natural proteins. The final production step is drying, for which different types of ovens can be used. Possible applications range from thermal insulation to acoustic panels. From a conceptual point of view, this material is also interesting in that some scholars compare the porous structure of wood to that of foam (Gardner 2002). Hence, one could classify this approach as a—probably unintentional—case of bio-mimicry. Steffi Silbermann, Stefan Böhm, Philipp Eversmann, and Heike Klussmann explore the potential of using wood-derived monofilaments in textile structures and components for architectural and structural applications. Bringing together wood with textile structures and principles is an approach that is seemingly gaining increasing interest.4 Using willow wood as a basic material, the research presented here aims to use custom-made tools and small-scale joints to produce quasi-continuous monofilaments, which are then further processed using weaving looms and likewise custom-designed robotic fabrication. The produced textile structures can be expected to be self-supporting, and could be used, for example, in interior architecture. However, structural building components will very likely have to use these structures as reinforcement or preforms for molded composite components, the production of which requires adhesives, similar to previous outputs of the research group, like for example the project Salix Regionalis 3D. As mentioned previously, the logic of reorganization or reassembly of their wood-derived components is key to the properties of engineered wood products. Especially in textile structures, one could argue that assembly is synonymous with organization, as their inner

Studies with a similar thematic focus were carried out, for example, as part of the research project Timberfabric at the IBOIS wood construction laboratory at EPFL in Lausanne (Hudert 2013) as well as in the work of Professor Peer Haller at TU Dresden (Heiduschke and Haller 2010). Yet another example is the research on Tailored Structures (Alvarez et al. 2018), developed by a group of researchers from Stuttgart University´s Institute of Computational Design and Construction (ICD).

4.

193

Rethinking Wood

organizational logic is directly related to the employed assembly technique and relatively independent of the chosen material. Relatively, because the material properties and sectional dimension of a yarn or filament certainly have an impact on the resulting fabric’s aperture size and mechanical properties. They may even disqualify the material for use. In any case, what is interesting here is that the relative independence of the inner structure or organization from the material composition (of the yarn or filament) is also one of the main characteristics of mechanical metamaterials (Overvelde et al. 2016). Future developments: towards a new materiality The material developments featured in this part of the book represent only a fraction of currently ongoing research on new woodbased materials and products. Yet they illustrate the diversity and potential in this field. More importantly, they show that new material developments without or with only partial use of adhesives are possible. The full scope of possible applications of cellulose and nanocellulose materials has yet to be explored. Due to the high degree of processing, these materials bear practically no resemblance to what we know as wood, neither in their mechanical properties nor in their visual appearance and haptic qualities. Nevertheless, these materials have the potential to contribute to new bio-based products and might move us closer to a circular economy. Moreover, similar to wood pulp, they may also open up new opportunities in the domain of sustainable 3D printing and replace synthetic plastics in packaging and in other applications. Materials like wood foam in particular hint at the potential of harnessing the self-bonding capacity of wood. In this context, it will also be interesting to follow the work of other researchers, like Carmen Cristescu of Luleå University of Technology, Sweden, whose research explores how to achieve auto-adhesion between veneers solely by steam and pressure (Cristescu et al. 2015). Bringing together textile tectonics and wood construction leads to light and potentially light-transmissive self-supporting architectural structures and objects. Together with load-resistant bio-adhesives, which should be available soon, this research could make an important contribution to textile-reinforced bio-composite building components. As mentioned above, what is also of interest here is that textile structures have similarities to mechanical metamaterials. In both cases, the organizational logic or pattern is more important than the consistency of the basic material (Overvelde et al. 2016). Another aspect that defines metamaterials is that they have properties that are not found in natural or conventional materials (Overvelde et al. 2016, Bertoldi et al. 2017, Kshetrimayum 2004). One example for this is auxetic metamaterials, which in contrast to conventional 194

materials have a negative Poisson’s ratio. In this context, one could claim that suppressing or overcoming certain natural properties and behaviors of wood is a similar approach, and that engineered wood products could also count, to a certain degree, as metamaterials. In conclusion, one could argue that the importance of the examples shown here and other similar ones goes beyond that of contributing to an increased variety and sustainability of bio-based materials. Together with fabrication methods like 3D printing and with regard to new material categories like mechanical metamaterials, these developments will increasingly empower designers to create tailored microstructures and material morphologies in order to achieve specific desired material properties in future materials.

195

Rethinking Wood

196

From Pulp to Form: Future Applications of Cellulose Heidi Turunen and Hannes Orelma

Various sources and forms of cellulose In nature, cellulose is an abundant material that can be found in large quantities not just from trees, but also from all kind of plants in general. Plants produce cellulose through photosynthesis, and therefore cellulose materials are carbon neutral and renewable. The main function of cellulose is to work as the main constituent of cell wall of plant or wood cells giving the mechanical character for plants. Wood cells of hardwood and softwood consist typically of 40–45 percent of cellulose, 20–30 percent lignin, 25–35 percent hemicelluloses and the rest of the material are additives (Stenius et al. 2000). Wood pulp has been processed from wood chips, where cellulose has been separated and dissolved into fibers. The disintegration of cellulose fibers from the wood can be achieved by utilizing mechanical or chemical methods (mechanical pulping versus chemical pulping). Outcomes of these processes can be found in everyday life in various kinds of products, such as in papers, cardboards, laminates, composites, or insulation materials. Wood cellulose can also be further disintegrated for polymeric cellulose by using chemical dissolution. Textile fibers have been traditionally manufactured utilizing this process and the outcome of these regenerated fibers are materials such as viscose, modal, and Lyocell. Cellulose derivatives are chemically modified polymeric cellulose, where outcomes can be additives in food production such as carboxymethyl cellulose (CMC) or cellulose acetate, which is a common material, for example, in eyeglass frames. Wood fibers can be further fragmented into nanosize particles by using mechanical, chemo-mechanical, or chemical disintegration methods, resulting in cellulose nano- and microfibers (CNF), also called nano- or microfibrillated cellulose, or cellulose nanocrystals (CNF), also called nanocrystalline cellulose (Klemm 2011). The main difference between CNF and CNC is the aspect ratio, which is smaller in CNCs due to the removal of amorphous cellulose parts by acid hydrolysis. Therefore, CNF has far higher gel forming and film forming abilities. Processing methods when achieving nanosize requires a large amount of water. Aqueous CNF suspension, made by mechanical grinding can contain 97–99 percent water. Whilst enzymatic meth-

Material experiments with casted wood. Photo: Eeva Suorlahti 197

Rethinking Wood

ods can reduce the amount of water by around 80–85 percent when using, for example HefCel-method (High-Consistency Enzymatic Fibrillation of Cellulose) patented by VTT Technical Research Center of Finland (Hiltunen et al. 2015). Lower water content reduces shrinkage after water has been removed from the material for use. In Nordic countries, the main research focus has been around CNF materials due to better material balance in the manufacturing process. The manufacturing processes of CNF do not lose material, as the manufacture of CNC with acid hydrolysis does, where up to 50 percent wood material is lost by hydrolyzed material. In addition to processed cellulose, nature also utilizes bacteria to produce similar nanosize cellulose that wood based CNF, called bacterial or microbial cellulose. Bacterial cellulose can even be grown at home within a few weeks under certain circumstances. The reason why wood chips are processed to wood pulp and wood pulp to micro or nanocellulose are varied. One reason is that when scaling material particles, new material features emerge. These can be, for example, film formation capability or increased strength. In addition to new material features, economic value of the material will increase. When machining pulp towards smaller scale, the good properties of cellulose, such as recyclability and biodegradability, are still possible to maintain up to the product level, when avoiding additives which reduces unwanted environmental impact of the material.

Wood

Timber

Wood Pulp

Micro/Nanocellulose

Dimensional Range

Meters ⇆ Centimeters

Meters ⇆ Millimeters

Millimeters ⇆ Micrometers

Micrometers ⇆ Nanometers

Machining

Chipping, turning

Sawing, milling, sanding, joining, carving

Pulping

Mechanical grinding, enzymatic treatment

Examples of Application Areas

Electricity pylons, burning wood chips

Wood construction, engineer wood products, furniture, interiors

Papers, cardboards, composites, laminates, insulation materials

Raw material for timber

Raw material for pulp production

3D printing materials, cell culturing

Raw material for micro/nanocellulose

Added value will increase • Processes become more complex • Economic returns and expenses will increase • New material properties will emerge

Figure 1. Wood-based cellulose can be processed at nanosize scale, when special material features occur. Photos: Eeva Suorlahti

198

Cellulose used in architecture and the construction industry Traditionally in the architecture and the construction industry, cellulose-containing products have been two-dimensional, such as building papers or cardboards. The reason for this two-dimensional thinking has been based on production facilities at the factories. However, shapes have been folded or rolled to manufacture three-dimensional products, such as covers of the packages, cardboard boxes, or paper rolls for multiple uses to serve other industries. In addition to building papers and cardboards, cellulose has been used as insulation materials on boards or blown inside wall structures. Cellulose has become a common insulation material due to easy processability and installation. In addition, cellulose-based insulation materials are made from renewable resources, compared with, for example, mineral wool or plastic-based insulation materials. However, insulation properties are currently moderate compared to insulation materials produced from non-renewal resources. Moreover, cellulose, an environmentally-friendly construction material, does not withstand moisture and becomes moldy in humid conditions where water vapor cannot escape from the insulation. Due to this, cellulose-based materials require knowledge of design of structures and attention at the installation phase. In addition to the previous ones, cellulose fibers have been used as an additive in concrete and paints. The purpose has been its use as a rheology modifier and to reduce water bleeding from the material structure. CNF materials have also been utilized successfully in concrete applications as a reinforcing material to strengthen the concrete (Peters et al. 2010). Properties of cellulose materials Cellulose as well as micro- and nanocellulose have many peculiar properties, which could add value for the end-products. Those properties might have not yet been applied in the construction industry, due to the early stage of the research and production, which might be just at the stage of viability. Especially properties at the nanoscale of cellulose materials have been marginally used in the construction industry. Interesting properties at the nanoscale include, for example, the specific undulating surface areas of nanomaterials and strength caused by hydrogen bonds within the material (Klemm et al. 2011). These undulating areas benefit the manufacturing of stronger materials via increased bonding ability. Nanoscale hydrogen bonds within the material enable strong, but lightweight structures. In addition to the above-mentioned properties, the renewal raw material is relatively inexpensive to produce, as current disposable applications, such as the paper industry will prove. However, when applying this material to the construction industry, several properties might need further research and improvements. 199

Rethinking Wood

These include, for example, the fire sensitivity of the material and behavior in wet or humid conditions. Currently, in cellulose fiber and fibril materials, water can lead to a decrease in strength and changes in form, such as swelling. This can be seen as weak property, unless the application areas could be found in biodegradable temporal applications. In any event, focusing to this aspect in material development, areas of application could be expanded. Nevertheless, the material is currently breathable which can release the moisture, not only to preserve, but also to prevent unwanted molding effects. The right kind of design with suitable end-use application, which takes into account the inherent properties or developed features of cellulose, can lead to new sustainable products, shapes, and design. New sustainable applications, which utilize, for example properties of nanosize cellulose, are possible to generate by using multidisciplinary material research, with the help of chemistry and new technological innovations.

Shaping cellulosic materials for the future Combining new technologies or technologies common among other materials can enable new openings for cellulose materials in future. The consistency of cellulose and nano- or microcellulose materials can be diverse; materials can be soft foam, hard bricks, fine filaments, or thin film. Shaping the material depends on material consistence, texture, and size of material particles. Due to this difference, methods to modify or shape the material are varied. Cellulose materials can be foamed to generate lightweight, but soft and airy materials. Filaments can be spun, enabling knitted or non-woven textiles. Thin films can be created by casting, but also 3D printing or extrusion are possible if three-dimensional objects are needed. In addition, shapes can be created by using molds, which saves material, because waste material is minimized by cutting excessive material out. However, the above-mentioned methods are not common in timber architecture and construction. Because of this, many shaping possibilities, which cellulose enables, are impossible to implement with timber or engineered wood products. Compared to traditional woodworking methods, where the main shaping method is the removal of excessive material, shaping possibilities with cellulose are more diverse, even though cellulose is also the main material in timber. When comparing cellulose to timber, or even engineered wood products, it can be seen that cellulosic materials such as wood pulp or nano/micro cellulose are very homogenous masses. The material is uniform and there are no variations in the structure, unlike timber which has knots and wood grains. The uniform material consistency enables versatile manufacturing and shaping methods both in dry or wet form. Because of this, design-oriented research related to cellulose materials, takes inspiration from manufacturing methods outside of the traditional paper or wood industry, such as 3D printing or casting. Applying methods beyond conventional 200

technologies, various shapes and forms are possible to achieve that take advantage of the varied consistency and properties of cellulose materials. Visually, cellulose materials can be manipulate easily. Pure wood pulp mass can be dyed to achieve a uniformly colored material. In addition, depending on the hardness of the surface, patterning or printing on the material are also possible to implement. Cellulose is also easy to chemically modify through available hydroxyls to tailor material surfaces or to produce chemically programmed materials. This approach can be utilized to produce a wide range of functional properties ranging from magnetic materials to super hydrophobic and antibacterial materials. By combining functional materials to shaping methods, new product applications can be generated. Case one: cellulose nanofibril films Nano and micro scale cellulose has an interesting property, namely a film forming ability due to high surface area and large hydrogen bonding. Films, or nanopapers as these are also called, are almost transparent or translucent 20–50-micron-thick sheets. These films can also be polished to increase translucency (Nogi et al. 2009). In addition to this film forming ability, films made out of cellulose nanofibrils have an excellent elastic modulus. The elastic modulus of a cellulose nanofibril film has been estimated to be around 14 GPa (Henriksson and Berglund 2007) and 15.7–17.5 GPa (Syverud and Stenius 2008) whilst elastic modulus of single cellulose nanofibril is 145–151 GPa (Iwamoto 2009). This means that inside the material several hydrogen bonds are formed which give it strength. In addition to the above-mentioned properties, the films can also act as an oxygen barrier. This property has already been used in the food packaging industry to protect food from the unwanted effects of air. In addition to oxygen barrier properties, a film also behaves like wood in humid conditions. The film can absorb moisture from indoor air and release it in dry conditions. Visually films are versatile. A suspension of cellulose nanofibrils can be dyed before casting them into film form. Films maintain translucency if the colorant is liquid, and transform towards opacity if pigments or other coloring particles are added. However, thin films are very fragile to handle in shaping. Therefore, it is necessary to develop suitable production method for laminating films together for strengthening. Future product applications could be found in the areas where material can be two-dimensional boards or shaped three dimensionally, or where it is possibly to see through the film, partially or completely. Simultaneously, the material can act as a space divider or protective material for various purpose, such as displays. Case two: cellulose nanofoams Although cellulose pulp has already been utilized in insulation materials for decades, these do not benefit properties of the nanosize 201

Rethinking Wood

Figure 2. Cellulose nanofibril films dyed by using natural colors. Research team: Heidi Turunen, Department of Architecture, Aalto University, Timo Kaljunen, Vesa Kunnari, Jaakko Pere, VTT Technical Research Center of Finland Ltd. Design Driven Value Chains in the World of Cellulose (DWoC), research project. Photo: Eeva Suorlahti

202

materials. Cellulose nanomaterials allow the production of low-density foams or aerogels, which unique characteristic includes the possibility of an extremely high surface area (Pääkkö et al. 2008). Therefore, these materials could be used for air filtering, insulation, and acoustic applications. Case three: cellulose nanocomposites The term “composite” means that at least two materials have been combined together to achieved new and better properties which neither material can achieve alone. Currently, the tendency seems that materials from renewable resources are used in composites to an even greater extent, thereby replacing oil-based materials. Traditional composite materials are difficult to recycle due to the strong bonding between reinforcing elements and matrix materials. Therefore, current activities have been targeted to monomaterial composites, as recycling is far easier without compromising mechanical properties. Cellulosic materials have been added to composite materials, enabling an increase of the environmental efficiency of the composite alongside plastics. Cellulose can be utilized in composites as a reinforcing material in polymeric matrix or in both reinforcing and matrix materials. The latter composite type is called all-cellulose composite (Huber et al. 2011). Cellulose is an excellent material for composites, due to biodegradability and high modification potential. In addition, this natural biopolymer imparts the material combination warmth and tactility. Warmth is not an inherent property of oil-based plastics. Materials combined with cellulose can also have a good strength to weight ratio. Overall, less material can be used to achieve strong and durable materials. Water resistance is not the best quality of nanocellulose or cellulose nanofibers, but can be improved by material research. Though, timber which consists of the most cellulose and regenerated cellulose used in textile industry are water resistant. Mixing cellulose with other ingredients will enable hard three-dimensional shapes by pressing or using molding techniques. Future application areas include the use of features of cellulose or micro/nanocellulose that bring added value to the composites, not currently exploited. Case four: 3D printed cellulose In architecture, 3D printing is an emerging manufacturing method and several new implementations have been generated using different materials. The role of the material research in large-scale, 3D printing is crucial, especially when there is a need to understand the potential of this production method more widely. Because cellulose comes from renewal resources, environmentally-friendly products are possible to generate due to the inherent properties of the material. Currently, cellulose materials have been 3D printed successfully in several research projects. However, the scale of the printed object is small, and there 203

Rethinking Wood

Figure 3. Casted Wood is produced by combining fine sawdust and cellulose nanofibrils (CNF). Colorants can be added to the material before casting. In this material study, cellulose nanofibrils act as a binder, while fine sawdust is the filling material. Three-dimensional shapes can be made by using molding techniques. In addition to creating a base for the composites, the material research can also be seen as a preliminary material development for large-scale, 3D printable wood-based fluid materials. Heidi Turunen, Department of Architecture, Aalto University, Design Driven Value Chains in the World of Cellulose research project. Photo: Eeva Suorlahti

204

are still several open questions, such as how to solve shrinkage of the moist printing paste and how to get layers to adhere to each other as the printed layers increase. The easy modification of cellulose allows the utilization of cellulose materials by using paste and thermo-printing methods. These two methods enable the production of different kinds of structures, where paste printing allows to maintain of the native structure of cellulose material. The thermoplastic cellulose can be utilized as an alternative material for traditional synthetic 3D printing materials and processed with home 3D printers. Nonetheless, material research encourages progress towards mid-scale printing and applications with small steps. Material development with a design driven approach has great potential for sustainable applications, also in architecture and construction industry, if over time the method becomes profitable in architectural applications. Future application areas could be where 3D printing creates added value compared to current manufacturing methods. In addition, those advantages can also relate to economic benefits, areas which support combining design ideas to manufacturing with ease, or to find solutions for demands which have previously been difficult to fulfill.

Conclusion Cellulose is the building block of plants and wood, and gives strength to the wooden material. Because of this, the material is common in nature and integral to plants and trees. Cellulose separation from wood has a long history and outcomes of pulping processes can currently be utilized in several different ways. Currently, papers, cardboard, and insulation materials are the most common related to architecture and the construction industry. The wood pulp can be processed chemically achieving chemically modified polymeric cellulose materials or textiles. If the raw material, the wood pulp is processed towards nanoscale, new material properties occur. In addition, when dealing with nanomaterials, further processing and production methods can be different from those currently used in paper, textile, or insulation industry. Consequently, processes which are not traditionally linked to cellulose products can be taken into account. The methods that can shape the cellulose materials can be methods like film casting or 3D printing. In addition, properties of the nanosize material can be utilized in several ways. However, material research is needed to understand properties of cellulose and especially nanosize materials more thoroughly viewed from an architectural point of view. Concurrently, with the help of design, future product potentials can be visualized and mapped providing valuable information for the material researchers to set targets for their work. Thanks to this cooperation slowly but surely, meaningful and potential applications can be found.

205

Rethinking Wood

206

Wood Foam: A New Wood-Based Material Frauke Bunzel

Wood foam is a new pressure-resistant, open pore material made of one hundred percent wood. Its production requires no synthetic adhesives, due to the bonding forces inherent in wood. Raw materials used for its production are sourced from deciduous and coniferous woods in the form of wood waste, wood from thinnings, or waste materials from the paper industry. Wood foam is fully recyclable.

Production Wood foam is produced in a few steps using widely used technologies in the wood, paper, and concrete industries. As the wood is finely ground, raw materials can be sourced from wood from forest thinnings and sawmill by-products. The use of non-wood lignocellulose materials, such as hemp or straw is also possible. The first step in the wood foam production process starts by turning wood chips or coarsely shredded material into thermo-mechanical pulp (TMP) fibers using a thermo-mechanical refining process. This thermal treatment of the fibers is key to activating the wood’s inherent bonding forces at a later stage, by mechanically and thermally releasing the hemicellulose and other wood components. After this step, the fiber lengths, depending on the geometry of the refiner’s grinding discs, range between 1,000 to 2,500 µm. This stage preserves the wood’s fiber structure (Figure 1). However, using these fibers in their current form to create the wood foam only produces a fibrous wood material. In order to produce a pressure-resistant, open pore wood foam, these TMP fibers must undergo an additional modified refining process during which they are heavily ground and fibrillated (Figure 2). This fibrillation later causes the fibers to hook into each other mechanically in the wood foam, and influences both the wood’s inherent bonding forces and the mechanical properties of the wood foam. Depending on the geometry of the grinding disc, the fiber lengths are between 200 and 800 μm after this step. The degree of grinding varies the proportion of water-soluble components released. These include cellulose degradation products, hemicellulose and lignin, as well as other extractives, needed for the wood’s bonding forces. This second grinding step produces a watery fiber suspension, which is then used as the starting material for making the wood foam. Various processes can be used to foam this mass. One method involves using blowing agents that at higher temperatures turn into gas, which in turn expands the mass into foam when it dries. Tensides, such as natural proteins, can also be used to physically foam the mass 207

Rethinking Wood Figure 1. Microscopic images of TMP fibers.

Figure 2. Microscopic images of fibers after a second modified refining process.

Figure 3. Microscopic images of foams made of long fibers (left) and short fibers (right).

Figure 4. CT images of a foam made from beech fibers (left) and pine fibers (right).

Figure 5. Calculation of pore volume using CT images.

208

before drying. This step determines the density and porosity of the foam. Wood ingredients, such as lignin, cellulose, and hemicellulose are chemically activated, forming covalent bonds that hold the fibers together and give the foam its strength. This eliminates the need for synthetic adhesives that could lead to a potential health impact from emissions. The foamed fiber suspension can then be poured into molds, making it possible to form 3D shapes made of wood. Drying removes the water and activates the bonding forces inherent in the wood. The drying process can take place in convection ovens, microwave ovens, or vacuum ovens as well as a combination of these, and produces an open-pored foam. The length of the fibers determines the structure of the foam, as shown in figure 3. Foam made from long fibers shows a fibrous structure, whereas very short fibers produce a foamy structure.

Properties The properties of wood foam depend on many factors, but above all on the fiber length and the ingredients of the natural substances used. The fiber length is determined by the type of starting material and the degree of grinding in the treatment, which mainly affects the structure of the foam. CT images of foams made of pine and beech fibers show the different foam structures (Figure 4) that are formed by different types of wood. Shorter beech fibers form smaller pores, and the structure of the foam is finer due to smaller web thickness. The CT image of a pine fiber foam shows a fibrous structure with larger web thickness. The CT images can be used to determine the pore volume and the web thickness, as exemplified by the foam made of pine fibers (Figure 5). The pore volume is 2–4 mm³ and the web thickness averages 0.13 mm. The foam structure mainly affects thermal conductivity, water absorption, and sound absorption. But it also determines the foam’s mechanical properties. The wood components of deciduous and coniferous woods are markedly different. This varies the strength of the bonding forces inherent in the wood, which are activated by the wood’s components, and influences the mechanical properties of the foam. In addition to the foam structure and the substances in the wood, the properties also depend on the density of the foams, which can be produced between 40 and 250 kg/m³. Due to the various influence factors, no defined property values can be assigned to wood foam. Figures 6 and 7 thus show how transverse tensile strengths and compressive strengths change with density. Both compressive strength and transverse tensile strength increase exponentially with density. The relatively large differences between the values at the same density is due to the starting materials, whereby both the type of wood and the degree of grinding have a major effect. The exponential 209

Rethinking Wood

200,0 180,0

TRANSVERSE TENSILE STRENGTH [KPA]

160,0 140,0 120,0 100,0 80,0 60,0 40,0 20,0 0,0 40,0

60,0

80,0

100,0

120,0

140,0

160,0

180,0

200,0

220,0

DENSITY [KG/M³]

Figure 6. Transverse tensile strengths of wood foams depending on density.

0,080

Figure 7. Compressive strengths of wood foams depending on density.

WOOD FOAM POLSTYRENE

THERMAL CONDUCTIVITY [W/MK]

WOOD FIBER INSULATION BOARD 0,060

0,040

0,020

0,000 0

20

40

60

80

100

120

140

160

180

200

DENSITY [KG/M³]

Figure 8. Measured thermal conductivity of wood foams of different densities and as a comparison of polystyrene and wood fiber insulation boards.

Figure 9. Water absorption of various wood foams made of beech and pine fibers.

Figure 10. Fire behavior according to EN ISO 9239-1 of wood foam and in combination with metal as a hybrid material.

210

increase in strength results from increasing the contact areas, and thus the number of covalent bonds, between the fibers, which in turn affect the strength of the foam. Thermal conductivity measurements show that low-density wood foam has low thermal conductivity. Thermal conductivity depends solely on the density of the foam, and not on the type of wood or the degree of grinding (Figure 8). The thermal conductivity of low-density wood foams is comparable to polystyrene and wood fiber insulation boards. Due to its open porosity, wood foam’s water absorption is very high. The water stored in the pores can be released without changing the structure and strength of the wood foam. The water absorption of pine wood foams is lower due to the wood’s components (Figure 9). The water vapor diffusion resistance factor is 1.7 and thus lies in the same range as mineral wool. Since the fire behavior of new materials is very important for application, wood foam was fire tested according to EN ISO 9239-1. Wood is classified as B2, i.e. normal flammability, and thus wood foam received the same criterion. Figure 10 shows the burning distance and the flame duration based on the wood foam density. In addition to a wood foam made from previously used fibers, a wood foam of very fine fibers was also produced and tested in order to evaluate the influence of the wood foam structure on fire behavior. As it is possible to combine wood foam with other materials, a metal-wood foam hybrid material was also tested. The burning distance and flame duration are not affected by the wood foam structure. Both foams can be classified as B2, i.e. normal flammability. The combination with metal greatly reduces the burning distance and the duration of the flame. Another important aspect of porous materials is sound absorption. This was measured in comparison to expanded polystyrene (EPS) foam (Figure 11). When measuring the sound absorption coefficient, the thickness of the sample is decisive. Thus, a thinner 30 mm EPS sample shows a lower sound absorption than a thicker EPS sample. The sound absorption coefficient of a thin, low-density wood foam sample is comparable to a thick EPS foam sample, thereby demonstrating the high potential of wood foam as a sound absorber. The sound absorption coefficient of a higher density wood foam is comparable to EPS foam in the same sample thickness, and so the open porosity of the wood foam contributes to a very good sound absorption. Because wood foam is made of one hundred percent wood, the foam can be completely recycled. It is also possible to produce new wood foam from used wood foam.

211

Rethinking Wood

1,0

BEECH, 70 KG/M³, 30MM

0,9

PINE, 150 KG/M³, 30MM EPS, 30MM

SOUND ABSORPTION COEFFICIENT α

0,8

EPS, 80MM 0,7 0,6 0,5 0,4 0,3 0,2 0,1 0,0 0

200

400

600

800

1000

1200

1400

1600

FREQUENCY [HZ]

Figure 11. Sound absorption levels of wood foams of varying densities compared to EPS foams.

Potential Applications of Wood Foam Consisting of 100% Wood

Potential Applications of Wood Foam Consisting of