Integrative Structural Design of Nonstandard Building Systems: Bridging the Gap Between Research and Industry

Abstract

The utilization in architecture of computational design and digital fabrication, coupled with the exploration of new material systems, brings the potential to break with conventional ways of building. However, these emerging nonstandard structures also demand new ways of designing and proving the structure’s integrity and safety. This paper aims to develop an integrative structural design methodology and workflow to design, optimize, and validate nonstandard building systems by combining a multiscale and digital-physical approach. The methods are showcased with coreless filament winding (CFW) structures, an additive manufacturing method representative of nonstandard building systems made possible by robotic fabrication. The results demonstrate the potential of this methodology to shorten the gap between research and industry, facilitating the realization of innovative structures.

Keywords:

• Structural Design
• Digital Fabrication
• Digital-Physical Workflow
• Multiscale Simulation
• Coreless Filament Winding (CFW)

Opening image. ICD/ITKE, Maison Fibre 2021, 17th International Architecture Exhibition—La Biennale di Venezia, Italy, 2021. (Credit: C. Zechmeister, ICD/ITKE/IntCDC University of Stuttgart, 2021)

Introduction

The digitalization age has brought progress to most engineering fields, however, the ways we build have not changed since the middle of the twentieth century—with the last advances in concrete and tensile structures. In research, the introduction of computational design tools and computer-aided fabrication methods has already broken with traditional typologies, enabling the development of load-adapted building geometries and material systems with finely tuned properties (Knippers 2013). Examples are 3D printing of metals (Buchanan and Gardner 2019) and concrete (Buswell et al. 2018) or fiber placement in composite structures (Oromiehie et al. 2019). These tools open the potential for lighter, high-performative, and more sustainable structures. Yet these systems are still a challenge for simulation and analysis methods, as they are typically used to prove the integrity of building structures (Knippers 2017). This situation constitutes an opportunity to rethink conventional structural design processes that can include integrative approaches and strategies to increase confidence in the design.

A building system is considered nonstandard when either the design process or the proof of the structure’s safety needs a different approach from the methods utilized at the time of being realized. Nowadays, these would mean structures that cannot be conventionally designed and verified by an analytical method such as finite element analysis (FEA) and are not covered by design guidelines, manuals, and building codes.

However, engineers had designed and constructed buildings, bridges, and other structures with considerable skills and success long before advanced structural simulations or design codes were developed. Progress in engineering has been made by carefully analyzing and predicting the causes of failure, independently of the method, which was physical, mathematical, or simply the strength of experience (Addis 1999). Then, the material design and structural system were adjusted to prevent the predicted type of failure. Looking back into history, one can see how integrating theories with experiments was necessary to design new building systems capable of resisting and carrying loads. The advantages of digital tools combined with more experimental methods learned from the past can help develop an integrative, digital-physical structural design approach— bridging the gap between research and practice and proving safety for novel and complex non-standard building systems.

Structural Design and its Historical Evolution

Structural design is the engineering process of designing the structure to safely resist the applied forces and load effects in the most resource-effective and efficient manner. Aspects of design that are not an explicit part of the strength consideration are also implied in this process, such as sustainability, constructability, and usability (Anwar and Najam 2017). The design process connects formal architectural ideas of geometrical form to a structural system (Kloft 2005), uniquely responding to each project’s specific requirements and scenarios (Bollinger et al. 2008).

Figure 1 illustrates structural design as a process with inputs and outputs. The structural design requires load scenarios, design requirements, constraints and boundaries. These inputs produce two primary and key outputs. The first consists of what must be built: material specifications; member size and arrangement, and cross-section details. The second is a justification of the design proposal—the proof of safety (Addis 2003). Therefore, structural design is the process in which the output, material, and design specifications must include the necessary degree of confidence in the proposed structure: the confidence to begin building (Addis 1999).

Figure 1. Structural design definition with inputs and outputs. (Credit: M. Gil Pérez)

Looking into structural design history allows us to understand how engineers have faced the challenge of designing new structural systems that were beyond standards but resulted in successful outcomes at that time. History reveals how different technological advancements have influenced the construction sector and the creation of new structures, frequently linked to the development of new materials or society’s needs. Even though ancient times also taught us valuable lessons on the evolution of construction and the means to build (Ngowi et al. 2005), the differentiation of design roles only occurred during the industrialization age (Veer 2016), and it allowed us to understand the influencing factors and evolution of structural design.

The Industrialization Age—Innovative Structural Design: Lightweight and Form

The introduction of new materials and fabrication methods during this period is key for the evolution of structural design (Billington 1985), requiring new ways of design adapting the structures to the material characteristics. Figure 2 shows how early designs in iron and concrete reassembled previous construction typologies in timber and stone (a and c), while at a later stage, the design began to take full advantage of the material properties (b and d), achieving lightweight solutions and using the form as a driver for the design (Pedreschi 2008).

Figure 2. Iron and concrete early introduction of the material in construction and design adapted to the material properties: (a) Abraham Darby, Iron Bridge over the Severn River, UK, 1781 (Credit: B. Parksy CC BY 2.0, via Wikimedia Commons, 2005); (b) Thomas Telford, Craigellachie Bridge over the River Spey, Elgin, Scotland, 1814 (Credit: Pixabay CC0, via Wikimedia Commons, 2017); (c) Robert Maillart, Stauffacher Bridge, Zurich, 1899 (Credit: Хрюша CC BY-SA 3.0, via Wikimedia Commons, 2011); and (d) Robert Maillart, Salginatobel Bridge near Schiers, Switzerland, 1930. (Credit: Rama CC BY-SA 2.0, via Wikimedia Commons, 2008)

As the knowledge of structural design increased and new materials developed, more architects and engineers were encouraged to innovate. The development of tensile structures generated the concept of structural form to drive architectural design in an integrative design approach, achieving an important design breakthrough (Sprague 2013) with projects such as the Dorton Arena by Matthew Nowicki and Fred Severud in 1952 (Figure 3a). This project greatly influenced future tensile structures, such as the work of Frei Otto (Figure 3b), designed through a model-based form-finding process that was structurally optimized following the rules of physics (Kloft 2005; Nerdinger 2005).

Figure 3. Innovative tensile structures: (a) Matthew Nowicki and Fred Severud, Dorton Arena during construction, North Carolina, United States, 1952 (Credit: Brazilian National Archives Public Domain, via Wikimedia Commons, 1952); and (b) Frei Otto, The Olympic Stadium for the 1972 Munich Olympic Games. Germany, 1972. (Credit: J. Royan CC BY-SA 3.0, via Wikimedia Commons, 2007)

From Innovation to Convention—The Standardization Process

Standardization can be defined as the process of articulating and implementing technical knowledge. If this knowledge is successfully implemented, it creates an authoritative and trusted standard. This process has taken place in history in almost every imaginable sphere: in politics, business, economics, science, technology, labor, culture, and ideas (Russell 2005). Standardization is recognized in three construction and structural design areas: materials and fabrication processes; structural systems, and design methodologies.

The construction industry could produce more, faster, and cheaper, with the standardization of materials and fabrication processes. In contrast to these economic benefits, the prefabrication of construction elements devalued skilled workers in the 1930s (Russell 2005). The standardization of structural systems constitutes the change from innovation to a convention of structural design, creating typologies (Engel 1968) that can be easily repeated and calculated. Finally, the standardization of design methods relates to the development of design guidance in the form of manuals and codes of design practice (Teichgräber et al. 2022).

The Digitalization Age—The Need for Integrative Structural Design

The introduction of digital technologies into industry characterizes the digitalization age. Its influence in architecture and structural design was not especially substantial in the first years, as the technological developments brought by digitalization focused on the digital control and automation of already established manufacturing processes. It was only in recent years, coinciding with the Fourth Industrial Revolution (Alaloul et al. 2020; Begić and Galić 2021; Klinc and Turk 2019; Scheuermann et al. 2015), that the impact on building design and construction practices became exponential (Menges and Ahlquist 2011). These developments allowed the production of complex forms that were difficult and expensive to design, produce, and assemble using traditional construction technologies (Kolarevic 2001). The Fourth Industrial Revolution aims to increase not productivity but rather flexibility, adaptability, and integration, which might significantly impact architecture (Menges 2015b).

Digitally driven design processes allow for dynamic and open-ended three-dimensional geometrical transformations, producing new architectonic possibilities (Kolarevic 2004). 3-D modeling allows for innovation not only in architecture but also in the field of computational mechanics and simulation technologies, offering opportunities to develop new structural forms (Knippers 2013). Several recent projects with complex geometries demand novel approaches to structural design and engineering, in which performance-based design and digital workflow are closely related (Kloft 2005). Data integration between different disciplines and data postprocessing methods for structural feedback that automate and accelerate iterative exchange is becoming essential (Kloft 2005). Parametric design strategies are also an opportunity for structural design. Strategies to integrate structural assessment on parametric models for early design fast feedback are already being introduced in design workflows (Preisinger and Heimrath 2014).

There is also a great opportunity to use materiality and its behavior as design factors instead of form. This direction opens multiple possibilities to break with conventional structural typologies when combined with the significant number of materials developed nowadays. The architectural and structural design processes are also challenged by embedding physical properties and material behaviors in computational design (Fleischmann et al. 2012). In addition, computation is a vital interface for material exploration in a cyber-physical process, including new modes of digital fabrication and manufacturing (Menges 2015a).

Therefore, digital design workflows need to be adjusted. Computational design, engineering, and construction should be integrated with these processes in a flexible, iterative, and agile way (Knippers et al. 2021). However, integrating structural design into the computational workflow is not the only engineering challenge that structural designers encounter. To keep up with technological developments, engineers need to rethink how to ensure emerging structures’ integrity and safety.

Proof of Safety in Structural Design

The capacity of a structure is determined according to a specific load scenario and a specific set of mechanical properties to determine whether it can or cannot carry the loads. Therefore, the proof of a structure’s safety is relative to the engineer’s decisions on material strength and loading. The engineer decides whether the methods, hypothesis, and results needed to prove safety are realistic enough (Ditlevsen and Madsen 1996). Nowadays, with standard building systems, safety factors ensure that variations in the load or material properties constitute a small risk. However, as nonstandard systems keep emerging, these safety factors are not always applicable, or higher factors are assumed to ensure the structure’s safety and reliability. Looking back into history, one can identify which tools engineers used to prove the integrity and safety of structures when the design had insufficient mathematical methods and codes to be supported.

Figure 4 gives an overview of the tools used through history by structural designers to build confidence in their designs and offers a few examples of how they were combined for different structures. These tools include physical or analytical methods and design guidance, among others, and are not developed linearly nor used as a single strategy. Consider that the engineer’s judgment is the only thing that has never changed. Engineers in the past, like nowadays, decided which available tools to use, how their models were set, and which loading conditions they assumed.

Figure 4. Structural design tools used by engineers to proof integrity and safety linked to the increment of knowledge in history (top). Examples (bottom): (a) Ad Quadratum. page 31, Viollet-le-Duc’s cross-section of Notre-Dame in Paris, France, 1919 (Credit: F. Lund Public Domain, 1919); (b) Eduardo Torroja, Zarzuela Hippodrome roof, Madrid, Spain, 1935 (Credit: Imagen No: I-ETM-115-09_04, Archivo Torroja – CEHOPU-CEDEX, 1935); (c) Frei Otto, form-finding scaled models, 1967 (Credit: Frei Otto CC BY 2.0, via Flickr, 1967); and (d) Buchan Group, Benoy’s Westfield London shopping center roof, London, UK, 2011. (Credit: Oxyman CC BY-SA 2.0, via Wikimedia Commons, 2009)

The strength of precedents is leveraged by copying already successful examples. Structural intuition is another tool related to the experience of the designers and the gap between what they knew was already built and what they believed possible (Mainstone 1999). These two basic tools, together with the introduction of Euclidean geometry (B. Addis 2003), can be linked to changes and innovations in the design of higher cathedrals with thinner walls and larger openings (Figure 4a).

In the eighteenth century, the design of structures began to be based on theories and scaled physical models, reducing the need to rely only on experience (Teichgräber et al. 2022). However, concepts of stress and elasticity were not introduced until the nineteenth century, along with the first cast iron applications (Addis 1999). Structural analysis methods were introduced in the second half of the nineteenth century, establishing the paradigm of calculability; if a structure cannot be calculated, it cannot be built (Knippers 2013). The development of graphical statics in 1866 by Karl Culmann constituted another significant step in the proof of safety methods. It is based on the graphic representation of the force in a structure, both in magnitude and direction (Pedreschi 2008). Many engineers relied upon this method as the primary form of proof. Only in the second half of the twentieth century was graphical statics superseded by computer numerical procedures (Bollinger et al. 2008).

As this knowledge increased, more mathematical models and calculations appeared. At the beginning of the twentieth century, general theories and analytical methods already monopolized design and safety processes (Bollinger et al. 2008). However, when it came to innovative nonstandard structural systems such as the Zarzuela Hippodrome roof by Eduardo Torroja, other methods of proving safety were still necessary, and full-scale structural testing was frequently performed in partial structures (Figure 4b). In addition, scaled models were also used to explore materiality and discover inspiring building forms, such as the form-finding models of Frei Otto (Figure 4c).

Advanced design simulations arrived in the 1970s with Finite Element Analysis (FEA), allowing the reliable calculation of highly statically undetermined structures and nonlinear behavior (Knippers 2017). Structural feedback can be used as a design driver by including these simulation methods in digital and integrative design workflows. In combination with the current design codes, which are based on the semi-probabilistic partial safety factor concept (Teichgräber et al. 2022), structures are designed and documented with great details of justification. Today, structures such as Benoy’s Westfield London shopping center roof (Figure 4d) can be made of thousands of plates of different geometry and thicknesses, adjusted to specific structural requirements, made possible only through parametric modeling in the structural design workflow (Knippers 2013).

Ironically, this sophisticated, well-established simulation and standard combined methodology can make it challenging to develop structural design innovations that cannot be modeled or described in codes. The task for today’s engineers is to be creative and break through conventions, developing the strategies needed to build the new generation of nonstandard building systems.

Nonstandard Building Systems and the Engineering Challenge

Figure 5 represents an interpretation of the evolution of structural design in time as a combination of developments in structural innovation and proof of safety, in relationship with both the industrialization and digitalization ages. It should be noted that the evolution of structural innovation and proof of safety is always incremental, but different factors have reinforced one or the other during history.

Figure 5. Conceptual evolution of structural design through history in relationship to structural innovation and proof of safety developments, situating the current research challenge and aim of integrative structural design methods. (Credit: M. Gil Pérez)

Industrial progress and material developments, especially in the use of iron at the beginning of the industrialization age, produced a design revolution in which structural innovation evolved faster than the tools and methods used to design and prove safety, and this period corresponded with nonstandard building systems that required experience, models, or experiments to be built. The standardization process soon helped to produce developments in the way safety was proven which, at the point named “Engineering Progress” in Figure 5, kept improving the methods with more theories and knowledge, moving ahead of the slightly slower innovation evolution. More structures could be classified as standard in this period and were proven with a safety margin using design codes and calculation methods.

Within the digitalization age, computation and automation began to produce new forms of innovation, slowly influencing the construction sector, where conventional building methods were (and still are) very rooted. With digital fabrication and new materiality, structural innovation is rising again, creating new nonstandard building systems that cannot use the structural design methods available to assure safety. This change constitutes the current research challenge, which drives the methods presented in this paper by learning from history and utilizing the full potential of advanced structural methods.

Nonstandard Building Systems in the Digitalization Age

In the digitalization age, nonstandard building systems are emerging for lighter, high-performative, and more sustainable structures. These nonstandard systems can result from developments in the following areas: computational design processes; digital fabrication methods, or emerging material systems (Figure 6).

Figure 6. Nonstandard building systems as the intersection of the application to construction of one or more of the following areas: computational design, digital fabrication, novel materials and materiality. (Credit: M. Gil Pérez)

Computational design processes, such as parametric design, have unlocked the realization of nonstandard geometries, which are frequently complex and require a digital design workflow. The structural design of these new geometries also needs to be integrated into the digital design workflow to obtain optimal results. An integrated structural design approach reduces material consumption by adapting the design to the most efficient structural performance. Advances in digital fabrication have enabled automated construction, allowing the exploration of materiality in a cyber-physical process. The accuracy can be programmed, customizing geometry and design digitally. These new methods, especially in their development stage, usually require prototyping. Moreover, their architectural and structural design needs persistent interchange of fabrication parameters, as they can produce significant deviations in the final produced elements. Therefore, structural design should be agile and iterative, introducing physical and digital strategies. Finally, the novel emerging materials and unconventional materiality enabled by the advance of digital fabrication are also being applied to the construction sector, achieving more sustainable or lightweight solutions. Structural design codes and simulation tools frequently do not support these new material systems, increasing the difficulty of proving their viability and safety. Their design is subjected to characterization and understanding of the material behavior and fabrication implications.

Figure 7 gives examples of nonstandard systems in the digitalization age based on this classification. “Smart slab” (Figure 7a) represents an example where digital manufacturing is used to modify and improve an existing structural system—a concrete slab (Meibodi et al. 2018). One Ocean Thematic Pavilion (Figure 7b) is the first kinetic façade made with glass fiber-polymer composites that works based on elastic bending (Lienhard et al. 2013). The eco-sustainable 3D printed habitat, TECLA (Figure 7c), was built with reusable and recyclable materials, sourced from local soil, carbon-neutral, and adapted to any climate and context (Chiusoli 2021).

Figure 7. Examples of nonstandard building systems: (a) ETH Zürich, “Digital Concrete,” Switzerland, 2018 (Credit: A. Jipa CC-BY-NC-ND, via Flickr, 2018); (b) soma architecture and Knippers Helbig Advanced Engineering, One Ocean Thematic Pavilion Expo Yeosu, South Korea, 2012 (Credit: T. Fildhuth); and (c) WASP, Eco-sustainable 3D printed house “Tecla,” Massa Lombarda, Italy, 2021. (Credit: WASP CC BY 2.5, via Wikimedia Commons, 2021)

Conventional structural design methods cannot respond to the design needs of some nonstandard building systems, especially when digital fabrication or new materials are involved. Consequently, transferring the systems from research into practice is frequently challenging as structural integrity and safety cannot be proven. The current solution to these problems is the application of high safety factors to their design, resulting in design modifications and higher material consumption. This strategy contradicts the building system’s original aim, which is frequently oriented toward producing lightweight and sustainable structures. The engineering challenges vary from different projects but share some of the following aspects:

1. The workflow requires the integration of structural design from the early design stages.

2. The design and geometry are not defined until the last project stage.

3. The simulation methods frequently cannot represent the complexity of the material or fabrication system.

4. The material system or structural typology does not match conventional building codes.

5. Prototyping and testing at full scale are frequently impossible for the complete system.

6. Material characterization needs to include the parameters related to the fabrication method.

All these aspects make it evident that new approaches are needed to prove the feasibility and structural safety of the design. Motivated by bridging the gap between research and practice, and enabling their construction in real applications, the methodology presented in this paper aims to provide integrative structural design methods to design, optimize, and validate nonstandard building systems.

Coreless Filament-wound Structures as a Nonstandard Building System

Fiber-polymer composites (FPC) in architecture have been used since the 1950s (Bank 2006). However, in contrast to other construction materials, such as steel or concrete, its use and development in construction did not occur immediately. This delay can be explained as a result of the lack of appropriate fabrication methods. Fabrication techniques (Peters et al. 1991) such as hand layup, filament-winding, or pultrusion could only produce continuous plates, and due to cutoffs or mold materials, these techniques often produced a significant amount of waste. Consequently, FPC applications were mainly modular structures in which it was possible to reuse the mold multiple times (Knippers 2017).

Only after 2012 could different materiality in composite structures be appreciated, with the development of the additive manufacturing (AM) method called coreless filament winding (CFW) at the University of Stuttgart. This AM technique allows the fiber filaments to align with the stress flow through a project-specific fabrication-oriented design that can produce larger, more geometrically flexible components. This alternative approach aimed to reduce the mold of state-of-the-art filament winding to minimize the waste materials produced during fabrication (La Magna et al. 2016). Since then, computational design inspired by biomimetic principles and integrated with engineering methods and fabrication feedback has unveiled unexpected and inspiring materiality (Menges and Knippers 2015). Just as had happened historically with iron and concrete, digital technologies have now helped overcome limitations and allowed the full potential of fiber composites to be utilized. The resulting CFW structures constitute a novel nonstandard system that can be situated at the intersection of nonstandard fabrication, material, and geometry (Figure 8). It should be noted that the same classification and characteristics can be similarly applied to other AM techniques.

Figure 8. Coreless filament wound structures as a representative nonstandard building system. (Diagram credit: M. Gil Pérez; left and right images credit: ICD/ITKE University of Stuttgart)

The robotic fabrication technique of CFW utilizes light and discrete steel frames with anchor pins to provide the boundaries for filament winding, resulting in lightweight lattice structures. A typical fabrication setup for CFW consists of an industrial, six-axis robot arm, a resin bath, a spool holder, a custom-built end effector, and the frames or supports for winding (Dambrosio et al. 2019). These can be stationary or mounted on external axes or manipulators. The configuration of the setup and the included elements directly impact the fabrication solution space that can be tailored to generate different component types (Gil Pérez et al. 2022c).

The robot pulls the fibers through the resin bath during fabrication for impregnation. The robot then winds the wet fibers from anchor to anchor, following a planned robotic path called the syntax (Zechmeister et al. 2019). The syntax is repeated, creating fiber interaction until the designed material amount is reached. After the component completion, it should be cured in an oven at a high temperature for a specific time, depending on the resin specifications. The curing process allows the composite to reach its final mechanical properties, at which point the frame is removed.

The fiber-polymer composite material system used in CFW can vary depending on the fibers and resin system chosen. The fibers are used as continuous filaments or yarns supplied in spools. The unidirectional composite bundles present highly anisotropic properties that are a function of the composite’s fiber volume ratio (FVR). The longitudinal direction has much higher stiffness and strength values than the transverse direction. The anisotropy ratio (as the ratio of the axial to transverse properties) varies with the type of fiber (Harris 1999). For example, high-performance fibers, such as carbon fibers, can have an anisotropy ratio of about 10, while this ratio is much lower in natural fibers. Other factors, such as possible defects in the fibers, might result in variability of the final mechanical properties. For CFW, it is essential to understand the properties and possible deviations of the system chosen, whether technical or natural fibers, to adjust the design to the material behavior. Therefore, an integrative approach is necessary for designing these structures, where computational design, simulation methods, and fabrication feedback are incorporated into a digital-physical workflow (Zechmeister et al. 2023).

From 2012 until 2017, several pavilions and building demonstrators have investigated the architectural and structural possibilities of CFW building systems (Figure 9a–e). The robotic fabrication design space was first explored with monolithic and modular structures by developing different setups or winding configurations.

Figure 9. Overview of CFW projects built between 2012 and 2021 by ICD/ITKE, University of Stuttgart: (a) ICD/ITKE Research Pavilion 2012, Stuttgart, Germany; (b) ICD/ITKE Research Pavilion 2013–14, Stuttgart, Germany; (c) ICD/ITKE Research Pavilion 2014–15, Stuttgart, Germany; (d) Elytra Pavilion 2016, London, UK; (e) ICD/ITKE Research Pavilion 2016–17, Stuttgart, Germany; (f) BUGA Fibre Pavilion, Heilbronn, Germany; (g) Maison Fibre, Venice, Italy; and (h) LivMatS Pavilion, Freiburg, Germany. (Credit for all images: ICD/ITKE/IntCDC University of Stuttgart)

However, to understand the system’s complexity and scale up the structures, modular configurations (such as Figures 9b and d) are more viable as they allow the evaluation of single components. From 2019 until 2021, three modular CFW projects were developed and built (Figure 9f–h) with very diverse research objectives, ranging from performative long-span to hybrid building systems and sustainable structures (Gil Pérez et al. 2022d). The BUGA Fibre Pavilion aimed to cover a larger span than previous projects with a dome-like structure. Previous modular projects inspire the component’s design, but in this case, the winding frames are separated in space, creating a bone-like shape (Dambrosio et al. 2019). Maison Fibre corresponds with developing the first fiber-timber hybrid components used as slabs in a multistory building system application. The material combination into a hybrid system in which both materials are structurally performative was the most novel aspect of this project, together with the actual application; a CFW structure became walkable for the first time (Dambrosio et al. 2021). The LivMatS Pavilion explored the design possibilities of a new material system: flax fiber composites, previously not used in any CFW application. The design is adapted to the material needs, aiming to explore more sustainable architecture (Gil Pérez et al. 2022b).

Although these projects seek different global objectives, the three share common engineering challenges. Changes in the syntax in which the component is wound can produce changes in the geometry. Besides, the fiber bundle’s cross-section can vary with parameters such as the component curvature or the winding resolution (number of rovings used versus passes in the layup). These variations and uncertainties are not fully defined until the final design is determined. In addition, the composite material also deviates. For example, differences in tension during winding can result in an uneven distribution of FVR along the bundles. Therefore, the structure’s mechanical properties are unknown and must be investigated during the design phase.

Given all these uncertainties, and without conventional methods that can represent the system, the proof of safety in CFW structures is an engineering challenge handled on a case-by-case basis.

Integrative Structural Design Methodology

Figure 10 illustrates the methods overview and integrative workflow (Gil Pérez 2023). The design space is divided into digital-physical realms (left and right of the graphic) and structural design versus architectural design (top and bottom). This division creates four quadrants corresponding to structural simulation, computational design, structural testing, and fabrication. At the same time, a multiscale approach is represented. Items closer to the central vertical axis are related to local or small scales, while items further away relate to the global or large scale. This workflow’s multiscale design and assessment only apply to modular structures or systems that allow for local-global discretization.

Figure 10. Integrative structural design methodology and workflow overview. (Credit: M. Gil Pérez)

Given this configuration, the necessary elements for the workflow are located. On the digital side, the global design and finite element model (FEM), component design and FEM, are located and interconnected, producing a multidisciplinary and multilevel modeling approach. This approach corresponds to the first methodology, M1: Multilevel Modeling and Evaluation. On the physical side, prototyping and structural testing at different scales are used in an iterative way to support and give feedback to the simulation scheme. In the first place, material testing is used to inform the local component FEM. Different testing strategies are possible within the second methodology, M2: Structural Characterization.

Then the first design loop begins, involving the component design, component FEM, and scaled physical models. This design loop is repeated until the architectural and structural design aims and requirements are achieved. Component design and component FEM are reciprocally informed to choose design options that fulfill structural performance and fabrication strategies and further investigate those with scaled physical models. This methodology is M3: Integrative Design.

Finally, when the design is ready for the final development stage, where full-scale prototyping and testing can be involved, the last optimization and verification loop is initiated. This loop needs to be repeated the same number of times as the planned full-scale testing. More iterations will be necessary if proof of safety is not achieved, as the engineer is responsible for judging the results and deciding whether the structure is safe enough. From experience, a reasonable number of full-scale tests that allow optimization and verify structural safety should be no fewer than three iterations. A higher number of tests could be costly for the project, while a lower number would not give any certainty on the test deviations. The process and method followed in this loop is M4: Optimization and Safety Verification. At the end of this stage, the complete structure can be optimized using the simulation workflow, and the digital information is finally transferred for fabrication.

The full integrative structural design workflow and methodologies were showcased and demonstrated in their application to the three modular CFW structures presented in Figure 9f–h. The complete structural design details are described in several publications for the BUGA Fiber Pavilion (Gil Pérez et al. 2021, 2020; Rongen et al. 2019), Maison Fibre (Gil Pérez et al. 2022a), and LivMatS Pavilion (Gil Pérez et al. 2022b).

M1: Multilevel Modeling and Evaluation

Complex material systems frequently require alternative modeling approaches for two reasons. Firstly, the final design and mechanical properties are highly dependent on the geometry resulting from fabrication parameters. Secondly, the FE model should be refined according to the design process, evolving from a conceptual level to an advanced detail design level. A multi-level FE approach is flexible enough to allow the structure’s assessment with different detail and refinement levels and from a dual global/component perspective.

The final geometry of CFW structures is unknown during the design process, making multiple levels of model refinement necessary (Figure 11). The load-bearing behavior of CFW structures varies from thin-walled surface components to truss components made of linear elements. Therefore, the overall behavior is studied using a parallel approach with surface/shell models and beam models that represent the fiber placement and are used to dimension the fiber bundles. This approach helps designers use structural feedback during the design process, even in the early design stages. It identifies misleading results by comparing various models, and great modeling and computation time can be saved using these simplified representations. Other methods adapted from Classical Lamination Theory (CLT) (Bert 1989) can also be used to inform the design process of the fiber layup (Guo et al. 2022a, 2022b).

Figure 11. Multilevel modeling of CFW structures including global and component levels for the BUGA Fibre Pavilion, Maison Fibre, and LivMatS Pavilion. (Credit: M. Gil Pérez)

In addition, component-based systems allow for a dual assessment, complementing the modeling strategy with a multiscale approach, dividing global and component models. The modeling can be simplified, and all loading conditions can be included globally. In contrast, the modeling can be more detailed at the component level, but the loading can be simplified. All these methods require a strong strategy for data integration. Evaluating the results is also essential, as unconventional modeling needs calibration and verification. These models are calibrated by comparing the different modeling approaches and the additional support of destructive testing. Then, a final digital-physical evaluation is required to validate the FE models and prove safety and reliability.

The evaluation of the models is made based on the governing structural behavior. In the case of CFW, this frequently manifests in the form of buckling under compression forces, which tends to produce an early failure of the composite structure (Knops 2008). For the structural design, useful strategies include limiting the maximum internal compression forces in the fiber bundles or designing the structure to achieve a higher buckling factor.

M2: Structural Characterization

When using an unconventional fabrication method or material system, relying on material data from coupon tests manufactured differently is frequently unreliable or nonexistent. Besides, existing testing methods (Whitney 1986) might not represent the building system’s complexity; therefore, customized designs to convey specific loading conditions are necessary. For example, in CFW, the resulting composite’s mechanical properties and failure modes are strongly linked to the fabrication technique, the resulting geometrical shape, and the achieved fiber volume ratio (FVR) and compaction.

Suitable small-scale prototypes and destructive testing methods to characterize CFW structures and their joint systems are shown in Figure 12. Each specimen type can be used at a different design stage to compare and find different results informing the process and calibrating the FE models. It is essential to differentiate which type of test and results are transferable to large-scale design, as small-scale specimens do not always retain the same behavior as the complete structural system. Partial testing or problem-specific experiments can be more effective than designing a coupon test to characterize all material properties (Gil Pérez et al. 2019).

Figure 12. Characterization of CFW structures through different small-scale customized testing methods. (Credit: M. Gil Pérez)

The main advantage of including small-scale material testing during the design process is not only setting material properties but the possibility of reducing the need for full-scale testing. Another strategy with the same objective is the implementation of quality evaluation techniques such as fiber optical sensors (FOS) or laser scanning of the structure (Mindermann et al. 2022a). These techniques can improve system reliability, making the design workflow more multidisciplinary, but require a more elaborated data integration approach (Gil Pérez et al. 2022c).

M3: Integrative Design

Nonstandard building systems that do not follow conventional typologies, materials, or fabrication methods can benefit from a digital design workflow. Optimization methods to reduce internal forces by adapting the form or geometry should be implemented from the early design stages to minimize the material usage of these structures. The structural design must be incorporated with an iterative and agile data integration to utilize structural feedback during the design process. Establishing a solid and multidisciplinary design framework combining the disciplines involved is crucial to produce a valid design outcome efficiently. Besides, in systems where simulation cannot fully represent reality, a digital-physical design loop allows for the checking of fabrication and simulation parameters with scaled physical models.

In the case of CFW structures, design, engineering, and fabrication involve constant assessment of requirements and parameters from all three aspects (Figure 13). In these three projects, the structural feedback is used iteratively with design and fabrication requirements showing successful results in material usage optimization.

Figure 13. Integrative design of CFW structures with the examples of the BUGA Fibre Pavilion, Maison Fibre, and LivMatS Pavilion. (Credit: M. Gil Pérez)

M4: Optimization and Safety Verification

The lack of structural codes and standardized simulation methods for new material and fabrication systems, such as CFW fiber-polymer composites, makes it essential to validate structural safety and reliability through full-scale destructive testing. This experimental hands-on approach to design and construction is not as common in the building industry as in other engineering fields, where the primary design tool is based on prototypes and tests. The design of the testing scheme needs to be planned carefully with an intelligent strategy to represent the worst-case scenario and build confidence in the design and structural system. Besides, the testing results can also be used to calibrate the FE models, unlocking further structural optimization using the already established digital-physical workflow. Like the design loop, this optimization and verification loop is iterative and multidisciplinary. Still, the number of iterations needs to be based on the planned number of full-scale tests that can demonstrate safety.

The modular design of CFW structures enables the testing of single prototypes (Figure 14), which are included in the overall design process. The tested components mirrored all fabrication and geometry conditions and were designed to represent the same load induction as the final structure, including connection systems and loading types. The FE analysis supports the decision about which component is to be tested, and it is used during the whole process to evaluate the results. After calibration of these models, comparing internal forces in the testing FEM under the force the test withstood and the internal forces in the global model under design loads allows for calculating the additional safety margin achieved. The customized testing scheme can result in an additional cost for the project, which is sometimes impossible to afford. Due to this, other methods for quality evaluation or structural characterization integrated into the workflow can reduce the amount of full-scale testing needed.

Figure 14. Full-scale testing of the BUGA Fibre Pavilion, Maison Fibre, and LivMatS Pavilion. (Credit: M. Gil Pérez)

The optimization and verification methodology described utilizing prototyping and full-scale structural testing enables the validation of innovative nonstandard systems, building confidence in their design and shortening the gap between research and industry.

Discussion and Conclusion

This paper presents an integrative structural design methodology and workflow for designing and validating nonstandard building systems. Three modular structures fabricated through the Additive Manufacturing (AM) Coreless Filament-Winding (CFW) technique are used as a case study. They can be considered a nonstandard building system that combines unconventional geometry, material, and fabrication: the produced systems are component-based; robotically fabricated fiber-polymer composite lattice structures. Other AM structures will share similar characteristics and challenges, making them the perfect target to apply the integrative design methods shown in this paper.

The methodology should be used with flexibility and adapted to the needs of each nonstandard building system. Figure 15 illustrates how different nonstandard parameters require different partial methods: digital workflow; structural characterization, prototyping, and testing. The project requirements will influence the number of design loops needed or the level of integration with other disciplines. As the structural design goal in any nonstandard building system will still be the design, optimization, and ultimately the proof of the system’s safety, the workflow presented in Figure 10 can be considered a complete set of methods and strategies for an integrative structural design scheme. Nonetheless, the steps and iterations for a specific structure still need to be reconsidered or adapted to its design and structural requirements.

Figure 15. Integrative structural design methods for the different types of nonstandard building systems. (Credit: M. Gil Pérez)

The strategies proposed to prove the safety of novel nonstandard building systems do not differ from those used historically (Figure 16). On the contrary, the proposed methodologies overlap with those already used by great engineers who helped develop innovative structures during the industrialization age. A great lesson from history is that experience and physical methods can support the design when the level of unknowns is significant. Utilizing the full potential of digital engineering techniques with experience and physical methods can be key to overcoming current engineering challenges.

Figure 16. Proof of safety strategies proposed by the integrative structural design methods for nonstandard building systems. (Credit: M. Gil Pérez)

In conclusion, it was demonstrated that the workflow needs to include multidisciplinary collaboration, emphasizing the importance of data integration techniques. The design should be performed iteratively with fabrication and structural feedback, resulting in a digital-physical approach. Material optimization is completed when the aesthetic and performance objectives are satisfied using the feedback loop between design and FEA assisted by scaled physical models.

The FE modeling should be flexible and adapted to a multiscale and multilevel scheme, including models that run in parallel with different definition details to identify misleading results. These structural simulations are informed by small-scale characterization testing to set initial assumptions and are later calibrated through full-scale destructive testing during the optimization and verification stage. The full-scale testing represents the worst-case loading scenario that produces maximum internal forces. The test setup should be modeled with an FEM, capturing load induction and boundaries from the actual test setup. This model calibrates the material parameters, comparing displacement and force with the mechanical test.

Finally, safety and reliability are proven by calculating the ratio between internal forces obtained in the full-scale test FEM under the maximum load withstood and the internal forces in the calibrated component FEM. Including full-scale prototyping and testing in the design makes further optimization to adjust fabrication and structural parameters possible. Additional quality evaluation methods can be implemented to reduce the number of full-scale tests, building confidence in the design and structural performance.

Advancing towards an integrative design approach that can be generalized for any nonstandard building system, a more extensive evaluation of parameters should be integrated into the multidisciplinary computational design framework. In addition, if these structures aim to produce more sustainable architecture, evaluating the structural system against other relevant factors, such as life cycle assessment (LCA) indexes, could be implemented as a variable in the early design stages (Mindermann et al. 2022b).

In summary, the integrative structural design as shown through its application to CFW structures revealed the methodology’s potential to prove the safety and reliability of nonstandard building systems, shortening the gap between research and industry and facilitating the increment of innovative structures built. Future research should extend the methods to identify the adaptations needed to validate other nonstandard building systems.

Author Statement

Marta Gil Pérez: conceptualization, methodology, software, investigation, formal analysis, data curation, validation, writing—original draft, writing—review and editing, visualization, project administration.

Jan Knippers: supervision, writing—review and editing.

Acknowledgments

The work presented in this paper was partially supported by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany’s Excellence Strategy – EXC 2120/1 – 390831618.

Data Availability Statement

Data sharing does not apply to this article as no new data were created or analyzed in this study.

Additional information

Notes on contributors

Marta Gil Pérez

Marta Gil Pérez is a Postdoctoral Researcher and Research Group Leader at ITKE, University of Stuttgart. Her research area lies in the structural design of nonstandard building systems, focusing on sustainability and codesign. She was part of the engineering team of successfully built projects such as the BUGA Fibre Pavilion, Maison Fibre, and LivMatS Pavilion.

Jan Knippers

Jan Knippers is a Structural Engineer and, since 2000, Head of the Institute for Building Structures and Structural Design (ITKE) at the University of Stuttgart. His interest is innovative and resource-efficient structures at the intersection of research, development, and practice. Since 2019, he was Deputy Director of the Cluster of Excellence “Integrative Computational Design and Construction” and Dean of the Faculty of Architecture and Urban Planning.