Three arches for a roof – case study of a multi-disciplinary design process

Abstract

In this article, the design process for the remarkable curved roof for a sports hall in Genk (BE) is traced. The project was granted to the partnership of young architects and a renowned office for structural engineering. The new building is roofed with three oversized arches, proposed as a shell structure in concrete in the competition, but executed as a steel truss system. The research is based on files from the archives of both the architect and the structural engineer. Close observations are made on six key moments in this design process. For these pivotal points sketches, documents and communication, are discussed in detail. The goal of this article is to make observations specifically on the collaborative design process between the architect and the engineer, in the specific context of a design competition. The research exposes the messy reality of a design process. It is observed how the final structure got its form, how design decisions were made and how the collaboration defined the project. It is illustrated how relative positions of team members evolve throughout the process and how progressive insight, conflict and mutual understanding are key during the design.

Introduction

In this article the design process for the remarkable curved roof for a sports hall in Genk (BE) is traced. The design results from a layered design process, in which both the architects and the structural engineers were thoroughly engaged.

In the field of design studies, a lot of research has been done on the nature of design. The broad scope, the high level of abstraction of these theories and the fact that research is mostly based on oral sources and experiments, does however not allow to get detailed insight in the course of (multidisciplinary) design processes in building design. It hinders to take the temporal aspect into account, which is necessary to understand for example the shifting positions of team members and the evolving value of statements throughout the process. These handicaps can be overcome by the bottom-up approach of case studies.

The goal is to make observations specifically on the collaborative design process between the architect and the engineer, in the specific context of a design competition. It is observed how the final structure got its form, how design decisions were made and how the collaboration defined the project.

The design process is scrutinized based on files from the archives of both the architect and the structural engineer—further referred to as the engineer. Documents such as sketches, calculation models, presentations and mail conversation are organized in temporal sequence, in order to chronologically reconstruct the design process.

Close observations are made on six key moments in this design process. They are pivotal moments to observe the roles of and the collaboration between the architect and the engineer. For these pivotal points sketches, documents and communication is discussed in detail.

The method used, is loosely based on the methodological concepts of genetic criticism. In this scholarly field in literature, manuscripts are used to study works of literature. From both archives, documents of respectively 439 pages and 766 pages from before and after the competition are composed. The selection served as the critical corpus for this research and is what distinguishes this set of documents from the archive. It is a crucial step in genetic criticism, in which a selection is made with a specific focus on the roof of the building and with the research questions on the collaborative design process in mind.

The project scrutinized in this case study originates from an Open Call competition. Open Call competitions are one of the instruments the Flemish Governmental Architect installs. It is a selection procedure in two phases: a competition for five candidates, selected out of interested participants, followed by a negotiation phase. The project is granted to prepare for public tendering, only after which a contractor gets involved.

The competition to extend the sports hall in Genk was part of Open Call 18, launched in the summer of 2009. The competition brief was to extend an existing sports complex, known for its swimming pool with a hypar roof. This building was designed by Isia Isgour and André Paduart in the seventies. It is one of the most prominent examples of a thin shell structure in Belgium (Fig. 1).

Fig. 1

Page 3 of booklet with competition proposal, showing the swimming pool with hypar roof by Isia Isgour and André Paduart to be extended in the competition. (Translation of title; ‘Pats Boem… Butterfly stroke!’. The text below is a poetic introduction to the competition proposal)

Fig. 2

The finished building. Picture by Tim Van De Velde

Fig. 3

Early collage of a sports hall with a glass brick roof.—archive Bel Architects, April–May 2010

Fig. 4

Sketches with assessment of the level of the building in relation to Paduart & Isgour’s Hypar roofed building.—Archive Bel Architects, April–May 2010

The case study focuses on the proposal of a design team consisting of the young ‘Bel Architects’ and the renowned office for structural engineering ‘Ney & Partners’. In their motivation they wrote ‘(…) Due to the combination of young design talent and experienced competence, we guarantee a fresh and at the same time feasible concept’[1]. The selection report more or less copied the statement. A general remark in italics as well states ‘Good engineers / Architectural design starting from constructive principles’ [2].

The specific ‘design identity’ of both offices are important for the specific course of this design process. In a text published in 2006 on the practice of Ney & Partners, the following can be read: ‘ (…) the task of the engineer, Ney maintains, is 'not free of obligations'… (…) the engineer, who always looks for what is called the 'structural optimum' (…) The structural optimum not only provides us with the most economical solution, but also the right solution. (…) In clear terms, the quest for the 'structural optimum' often results in the design of forms that make sure that only purely tensile or compressive forces occur in the material and that bending moments and other parasitic material stresses are avoided as much as possible.’ [3] This aspiration, to always look for the structural optimum, will prove to be crucial in the specific course of the process.

In an interview with Bel Architects the question is asked who of the two partners ‘thinks’ the most in a constructive way, since one of the associates has worked for Samyn, a known ‘architect-engineer’ in Belgium. They answer that their taste for construction is one of the important things they have in common and agree that the ‘structural aspect’ is not something that can be separated from architectural design. [4] This interest in structures and structure in architecture, probably made the architects receptive for the engineers arguments and facilitated a design process in which both the architect and the engineer can operate as equivalent design partners.

The team decided to go against the competition brief and proposed not to extend the existing building, but to build a separate sports hall. They argued that it would be impossible not to harm the existing building by attaching a large extension to it, nor to give architectural value to these new sports facilities when attached to such a prominent building.

The new building has a rectangular plan of 60 by 81 m and is covered in longitudinal direction by three oversized arches, proposed as a shell structure in concrete in the competition. Although the team had already decided on the architectural form and the structural concept during the competition phase, afterwards the roof evolved from prominent and visible concrete vaults to a concealed steel truss system. (Fig. 2)

In what follows, a short overview of the six pivotal points in this case study is made.

A ‘proper’ relation to the existing building, access of natural light in the sports hall and an integrated solution for the structure, techniques, acoustics etcetera are main concerns from the beginning of the design. The first pivotal point is when these aspirations materialize in the architectural project with three large arches. This happens during a meeting, only one week before the competition deadline. With the arches, the concept of ‘the ideal shape’ is introduced. Both the architect and the engineer agree to look for ‘the ideal shape’ of the roof. In doing so this becomes a mutual design objective. Such a concept however allows for various interpretations. ‘The ideal shape’ will evolve to a central concept in the collaborative design process.

The second focus is on the competition proposal, published at the end of May 2010. It can be seen that the presentation is made with a specific rhetoric for the competition. The decisive tone of the statements imply that the design is already in an advanced stadium, for which the proposal is as well appreciated in the jury report. The archival documents however prove that the design was still in a premature stage, as can be expected from a competition proposal. These premature statements, and specifically the emphasis on concrete and a shell structure, do have consequences throughout the design process. Corrections on these statements have to be made later, which at a certain point introduces tension in the relation between the architect and the engineer.

The first study report by the engineers is discussed as third pivotal point. This report tries to nuance exactly the statements made during the competition. It mainly focuses on ‘the ideal shape’ of the cross section of the arches and opens the options for several construction method and materials, including concrete, steel and masonry. The focus on the cross section specifically allows to keep these options open. This angle and the conclusion that an ‘ideal shape’- a shape in which bending moments are minimized—has to be looked for, orients the further course of the process.

The fourth pivot is the definition of the cross section of the arches. After a long process in which the common ground for the ideal shape was sought for, during which both the architects and the engineers tried to fit the boundary conditions of the others into their priorities, the engineers reaffirmed that the arches should in their opinion be ‘optimal’ shapes, thus catenaries. They use classical arguments such as the economy of materials, but it is clear that this objective as well had to do with the identity and the specific design preferences of the engineers. The catenary shapes are obtained by recalibrating another boundary condition: the client agrees to reduce the free height of the sports fields in the corners of the sports hall.

The fifth pivotal point is the end of the preliminary design phase only a week later, and bundles  three occurrences. These result in a shift in focus of the design team towards an integrated design proposal with prefabricated concrete elements. First is the response of a contractor that questioned the feasibility of cast-in-place concrete shells, the execution method the engineers asked him to evaluate. Second is the response of the client on an image that showed prominent ribs in acoustic material on inside the arches. Both occurrences urge the team to recalibrate the relative importance of their design priorities. The focus is shifted towards an integrated design with prefabricated elements.

The last pivotal point is at the end of September, when the engineers (firmly) propose to shift the structural system from prefabricated concrete shells to a steel truss system.

This case, and more specifically the part of the design process after the design competition was used before in a paper [7]. In Sect. 2.3 of [7] a general, visual overview is given of how the roof structure evolved after the competition from a concrete shell structure towards a simple, steel truss system.

The case was used as an example to make some comments on the proposed chapter on ‘conceptual design’ in fib Model Code 2010 [8], a publication that serves as a basis to develop future codes for the design of concrete structures. In this code, it is proposed that ‘conceptual design’ is a phase in the design process before the phase of ‘detailed design’. In the paper it is argued that ‘conceptual’ design is not a phase before the ‘detailed’ design in the workflow of the engineer. The design process cannot be divided in sequential steps and its course is not predictable. The paper argues that ‘conceptual design’ and ‘detailed design’ are rather types of design acts and that they always co-exist to a certain degree in the work of the engineer. Acknowledging that both cooperating partners go through an equivalent kind of creative design process, is an important premise to compare and to find the common ground on these points where they develop the design together.

It goes without saying that this elaborate discussion on the process is still a large simplification of the complete process. Selections were made within the specific focus of this article: to make observations on the collaboration between the architect and the engineer.

Other specialists such as an acoustic engineer, engineers for special techniques were involved as well. They of course had a certain impact as well, but are not specifically observed in this study.

The language used in the process was Dutch. Translations have been made by the author.

Six pivotal points

A building ‘Such as the one of Isgour’

Bel architects were asked during an interview what their starting point was for the ‘three shells’ [4]. Their first idea was to make the extension underground, with a roof out of glass bricks. (Fig. 3) This would be located in front of the existing building and serve as a square. After a conversation with Laurent Ney it however became clear this would have tricky consequences. So they asked themselves, why would we organise the program underground, why wouldn’t we make a building such as the one of Isgour? The building raised from the ground, floor by floor. (Fig. 4) First it was still a rectangular box, but this missed something. The architects state that the new sports hall owes its shape to the existing building.

In the competition brief [9] as well as in the briefings [10], the importance and the value of the existing building of Isgour and Paduart is strongly emphasized. Numerous documents were collected by the architects on the buildings of René Paduart and on thin shell structures (in Belgium). Marks on the elaborate report by the Flanders heritage Agency indicate for example the roof consisting of a concatenation of five thin hypar roof elements in concrete, organized in such a way that they are not subject to bending forces. The relation to the existing building is an important focus of the design, very early in the process. The question of how to materialize this relation, however asks for extensive design labour. Sketches in Fig. 5 are illustrative to this.

Fig. 5

Ideas for a building ‘such as the one of Isgour.’—archive Bel Architects, between 21/05/2010 and 11/06/2010

Another important aspect, already present in the idea to use glass bricks for the roof, is the goal to get diffuse, natural daylight into the sports hall. At this point, Ney & Partners did some first dimensioning for a grid of steel beams to span the sports hall. It is attempted to integrate this access of natural light with acoustic solutions, thermal isolation, the structure and other aspects of the design, for which as well numerous sketches are made. (Fig. 6).

Fig. 6

Design sketch in which access of natural light is integrated with the acoustic solutions, thermal insulation, the structure and other aspects of the design.—archive Bel Architects, between 21/05/2010 and 11/06/2010

The idea to make a building with three prominent arches is validated in a meeting between the architects and the structural engineers, only eight days before the competition deadline.

This idea has grown in a very short time in the architect’s office. Two days earlier they send a three page summary of six possible roof sheds, which was then the idea for the sports hall. In these the arches cannot be seen yet [11].

The arches strongly remind of the design for the museum of modern art in Warsaw¹ by Christian Kerez. Copies of pages on this project in an issue of El Croquis [13], as well as a page titled ‘Kerez’ with sketches on this design, prove they were aware of this visually obvious reference (Fig. 7, left). Archival documents however show that rather than a new starting point, the reference merely helped to give the final stir to the ingredients that were on the table: a building that can relate to the one of Isgour and a roof shape that facilitates the access of natural light. (Fig. 7, right) for example shows the first interpretation of Kerez’s design as reference for curved sheds, closely related to the square sheds sketched above.²

Fig. 7

Left: Sketches that show the reference of the design for the museum of modern art in Warsaw. Right: Sketches that show the first interpretation of Kerez’s design as a reference for the sheds of the roof.—Archive Bel Architects, some days before 11/06/2010

The meeting notes from the architect (Fig. 8) are concise, but do clearly summarize the design intentions. In the cross section on top, two key words are written; ‘funicular line (parabola)’. Already at this meeting Ney & partners has mentioned their desiderata for the shape of the arches. The sketches below illustrate both the idea to integrate techniques, and the idea to make concrete longitudinal beams, on top of which the arches can for example be executed in wood. In first instance, a monolithic structure was not necessarily assumed.

Fig. 8

Meeting notes by Bel from a meeting with Ney & Partners—Archive Bel Architects, 11/06/2010

After the meeting, a collaborator at Ney & Partners prepares a drawing in which ‘the optimized shape for the arch’ is indicated in red [14]. (Fig. 9, top section.) This drawing is based on the proposal by the architects. (Fig. 9 below). In an internal message is noted that a thickness of 20 cm is assumed, which ‘looks slender but should be possible’ and that it probably is best to use the same shape for the three arches to cope with the thrust. This internal message is integrally forwarded to the architect by Ney’s project engineer, reaffirming the statements the colleague made by adding that it is theoretical, but correct. ‘In attachment as well ‘the ideal shape’ in red. With ‘ideal’ the occurrence of only pressure forces is meant.’ [15] In this way, the concept of ‘the ideal shape' is introduced.

Fig. 9

Top section: top red curved line indicates ‘the optimized shape’ for the arches, an adjustment proposed by Ney & Partners. Section below: Three options for the roof shape, as first proposed and later communicated to other collaborators by Bel. (Translation black and yellow text: volume of shell construction: … Red text: Volume sports hall with rectangular shape: … Volume with structure included: …)—Archive Ney & Partners 11/06/2010

The architects reply that the proposed shape looks good, and that they indeed should look for an ideal shape. [16] They however stress several other constraints for the shape of the section. They comment on the decreased width of the arch in the middle, which should stay related to the sports field below, and on the increased height, which makes it different from the height of Isgour’s (see Fig. 9). They continue with ‘the optimisation of the shell thickness’. The architects estimated the volume of concrete to be 970m3 (for the cost estimation), which allows for a thickness of 12 cm. They add that ‘if the shell thickness would evolve to 15 cm, this would be fine’.

Later the architects inform the collaborating acoustic and energetic engineers that they chose, together with Ney & Partners, for a shell structure [17]. They send three sections in attachment. (Fig. 9 below) The black outline has both structurally and formally their preference, but multiplies the ideal volume – the rectangular one—by a factor of 1.4 to 1.6. It is interesting to note that in this drawing, the proposed line by Ney & Partners is not taken into account.

Observations

In this first phase, the design goals for the sports hall, a building that can relate to the existing building and the admission of natural daylight, concretised in a design with three prominent arches.

The aim to make the arches catenaries is a clear, formal design input from the engineers, motivated by their concept of an ideal shape. The architect and the engineer agree on the ideal shape as design objective, but have a nuanced different understanding of what an ideal shape entails. For the engineers this refers to the idea of the optimal use of material, an important concept in their practice, as is discussed in the introduction. The architects agree on the option with the highest arches (see Fig. 9 below), but keep their own variant, which meets the boundary conditions they have set. This shape is close to the optimal shape Ney & Partners propose, but the deviation has a crucial impact on internal forces in the shells.

In semiotics, the concept of a ‘floating signifier’ exists. These are signifiers without a specific signified, which means that they can mean different things to different people. [5, 6] The collaborators have to look for a compatible understanding of the concept in the context of this specific project. To get to a final design, both have to argue and negotiate on the shared definition of what this shape will be.

Fig. 10

Section in competition proposal, presented in relation to the existing building

The most complete project. A tactical approach

The competition proposal is finalized at the end of June 2010. (Fig. 10) The booklet counts 75 pages and is divided into six chapters. The first chapter elaborates on ‘Masterplan sports village’ and mainly argues for the proposal of a separate building. The second chapter is titled ‘Phase 1: The ideal sports hall’, followed by ‘Phase two: renovation’, which focuses on some interventions in the existing building. These are followed by chapters elaborating on cost estimation, cost management and a chapter on ‘principles of sustainability’. [18]

The second chapter starts with a text on the new sports hall. This is surprisingly concrete. It for example states that the surface of the building will be 80,8 m by 59,4 m or that the colour of the sports floor will be azure blue in rubber PU-finishing. A shell structure in concrete is proposed, which is further emphasized with specific full page illustrations (Fig. 11).

Fig. 11

Full page illustrations in competition proposal. ‘White concrete’ on the left, ‘shell structure’ on the right. The subtext (left) discusses the perforations in the roof

A significant part of this chapter is devoted to technical specifications. Three pages, provided by Ney & Partners [19], are specifically on the structure. Similar notes were included on acoustics,³ HVAC-installation, lightning and fire prevention.

In the pages on the structure, the equilibrium shape, the structural principle of the arch ‘in three dimensions’, the execution methods and even foundations are discussed. In the first part: ‘For a given height and width, determined by the plan organization, architecture, access of natural light, acoustics etcetera, we looked for the most ideal arch shape. This ideal arch form is the equilibrium shape, in which the arch only works in compression.’ This is immediately nuanced: ‘Slight bending as a result of asymmetrical mobile loads, such as wind, snow and maintenance are unavoidable and relatively small. The search for this equilibrium shape is an iterative process because it depends on the thickness, which in turn depends on the shape…’ [18].⁴

Further, the issue of thrust inherent to arches is described and the thickness is discussed. ‘At this stage, the goal is a shell with a constant thickness of 20 cm. At a later stage, this thickness will have to be optimized, taking design criteria such as buckling stability, reinforcement concentrations etcetera into account. This optimization process will in any case result in a shell of variable thickness.’

The following subchapter explains that the shells can not only be observed in section, since bending action in longitudinal direction as well occurs. In the chapter on execution, two possible execution methods are evaluated – cast-in-place or with prefabricated wide slab elements.

Four illustrations (Fig. 12) are included. The one left on top shows two arched lines. The black one corresponds with the outline drawn by the architects in the other documents. The red line is the ‘equilibrium arch’ as proposed by Ney & Partners. The picture thus shows explicitly that the exact section is not found yet, which is opposed to the reassured statements made in the text. The two coloured images illustrate that the shells as well have to span in longitudinal direction. The image below is a drawing generated in final element software. A trained eye understands how this depicts bending action, although the result probably shows vertical deformations in the shell.⁵

Fig. 12

Illustration in competition proposal in the chapter ‘structure’. Left: illustrations on equilibrium shape. (An attentive observer can see how the double line actually explicitly shows the internal discussion on the shape) Right: illustrations under ‘arches in three dimensions’. Translation of text in blue added by author

The team is the only one selected for the negotiation phase. A set of questions is formulated, mainly aiming for clarifications. The architects prepare an elaborate additional booklet in which each question is answered. [21] It starts by stating that the answers will only be final during the process and that their current replies have to be understood as intentions and possibilities.

How money can be saved is included, since cost estimations exceeded the initial budget with 27% [18, 22]. A message by Ney & Partners, sent during the negotiation phase, informs that ‘Laurent launched the idea to make the arches in masonry, à la manière de Eladio Dieste.’ Although the project leader is sceptic, he mentions to the architects that according to Laurent Ney this should be cheaper [23]. Bel architects answer that they are great fans of Dieste’s work. This option to possibly use brick has an important impact on the further course of the design process.

They add that in any case they should answer the questions of the negotiation phase with an overview of possibilities for the structure, to illustrate that a lot of research will be necessary and that a part of the solution for the budget can be found here [24]. The engineers propose to ‘at least include something such as: concerning the shell construction far more research has to be done. A possible solution can be found in the execution with reinforced masonry (cf. Eladio Dieste). Possibly a lot can be saved there.’[25].

In the booklet this question is answered by suggesting possible savings can be made in the roof structure by optimizing the amount of material and/or the execution method. ‘Margin in thickness of the concrete shell: max 20 cm, min 14 cm. (…) The exact amount is difficult to determine without offers from contractors.’[21] On the optimization of execution method is noted that more research is necessary, but that options are legio: cast-in-place concrete shells, concrete shells with concrete prefabricated elements, a concrete structure with infill masonry, a steel structure with a hybrid steel plate concrete floor system, wood,… [21]

It this negotiation phase, the team both nuances their first statements made by suggesting that a lot of research still has to be done and continues to make statements that will suit the jury, by suggesting that these options probably will result in savings.

The answers are sufficient for the jury. The team is granted the project. The final report of selection concludes: ‘(The proposal) is the most complete project and makes a daring conclusion by building it in another place. (…) The collaboration of a driven architect and a good engineer makes the project exceptional, in the same way exactly this combination made the existing complex extraordinary.’ [20]

Observations

One can ask her-/himself why in a competition proposal so much energy is spent on the technical aspects of the proposal. For professionals in the specific domains, it is obvious how general these descriptions often are. The archives as well prove that it was impossible for the team to already have made such thorough studies on the design.

The approach however proved to be very effective. The proposal did not only convince with the idea to make a separate building, it is as well praised as the most complete project in the jury report. This is why the input of the engineers in a competition proposal, should as well be read in the light of the rhetoric used to try to win a competition. A proposal that looks well studied and detailed is convincing and reassuring for a client. It is interesting to note that the team stands united in this tactical approach.⁶ As shown, the engineers as well make specific statements on the used material, the execution methods etcetera.

Whereas the team wants to keep the structural principle and material open for research, the proposal clearly aims for a shell structure, made in concrete. It is agreed that is safe to recalibrate these statements during the negotiation phase by remarking that it would be good to list several possibilities for the structure. They however as well bluff on the possible savings in the budget, as can be understood from the archive and mail conversation, no thorough study has been (or could be) done (in this phase) to evaluate the cost. It is noted that already here is mentioned that a contractor should be contacted to get a specific price estimation.

The design choice to keep options open

The third pivotal point is situated in February of 2011, when the project is picked up again after the competition. The engineers publish a first version of ‘Study Report 1’⁷ [26]. The engineers note in the introduction that the goal is to set the different design parameters for the shell structure. The report is largely explanatory on structural principles and consequences. It has three chapters: shape, material and execution method. Most attention is given to the study of the shape of the section. In this article, some specific illustrations and statements are observed.

Firstly, the engineers elaborately focus on the ideal shape and stress (again) what this means for the engineers. ‘This is the equilibrium shape, in which no moments will occur. It minimizes the use of material. (…) This shape is confirmed by the results from the calculation model.’ Six coloured diagrams are included to prove this statement (Fig. 13).⁸

Fig. 13

Results from final element method calculation as images in study report to underpin the statement that no moments will occur in the equilibrium shape. – study report 1 by Ney & Partners, 29/03/2011

Apart from the first and the last diagram, the values on the scales indicate that the occurring internal forces are negligible, so indeed no moments occur. In this way, these diagrams indeed support the statements made. No result however resembles the calculation result shown in the competition proposal (Fig. 12). To get results like the ones shown here, the valleys in the middle have to be linearly supported. This cancels the beam action in longitudinal direction. The calculation model is a large simplification of the actual structural behaviour of the design. As was pointed in the competition design, the longitudinal direction is structurally very important. When this action is taken into account, the statement that no moments occur in the shell is inaccurate. The engineers isolate a certain structural principle that is on the table for discussion and advocate for their preferences, making use of simplified calculation models. Again it has to be concluded that the engineers consciously use their tools to persuade and that there is a certain rhetoric involved when using such diagrams in this context.

This however does not change the fact that indeed less moments will occur, than when the shape of the arches would be random. This is illustrated in Fig. 14. In the left image, the arches are separately supported with line supports. This result matches the result shown in the study report. In the middle and right image, only the corners of the intermediate valleys are supported. The arches in the middle have the optimal section Ney & Partners is looking for. On the right, the exact same calculation is made for the section shape as drawn by the architect in the competition booklet. (Figs. 10 and 9 below). It is clear that the moments in the shells in this direction are indeed optimized with an equilibrium shape.

Fig. 14

Comparison in final element method of the moments produced in the arches, in the direction of the arch section. Left image: arches are separately supported with line supports. This result matches the result shown in the study report. Middle and right image: only the corners of the intermediate valleys are supported, which is closer to reality. Middle: optimal section Ney & Partners is looking for. Right: same calculation for the section shape as drawn by the architect in the competition booklet. Moments do occur in the ‘optimized’ section (middle), but far less than in the ‘free form’ section (right). Image generated by the author of this article

The second observation is on the estimation for the proposed thickness of the shells. After close observation, it becomes clear that the option in masonry was still taken into account, which seems to have had a guiding effect on the arguments in this study report.

The report elaborates on the equilibrium shape being influenced by variable loads, which means that the theoretical equilibrium shape constantly changes depending on for example wind loads. The proposed idea to cope with this, is to start from the equilibrium shape for permanent loads and to look for a thickness that circumscribes the envelope of all equilibrium shapes, which results in a minimal thickness of 25 cm for the central arch. Note that this is thicker than the initially proposed thickness of 20 cm, although a reduction of the thickness was used as an argument to reduce cost during the negotiation phase. In the report is added that to equilibrate the thrust of the intermediate shell, it is estimated that the outer bows need a thickness of 32 cm.⁹

This argumentation for the thickness has a purely geometric motivation, and is not informed by material tensions or (buckling)stability. It seems to originate from the logic of masonry, since this is the only material that has very limited capacity to cope with bending moments.¹⁰ For an execution in concrete, this additional thickness is less necessary, since bending moments can be taken by the material.¹¹ This implicit focus on masonry however is not explicated in the report. The motivation to propose this option to determine the thickness seems to both be motivated by the quest to find an ‘ideal shape’ and by the attempt to keep options for the materialization of the arches open.

At last, a table (Fig. 16) with eight combinations of construction types (shell, vault, frame – see Fig. 15) and materials (concrete, masonry, steel) in the chapter on execution methods is looked at in detail. [27]¹²

Fig. 15

Three principles for the construction types in longitudinal direction. ‘option 1 (shell), option 2 (vault), option 3 (frame)– study report 1 by Ney & Partners, 29/03/2011. Translation of text in blue added by author

Fig. 16

Overview of possible construction methods – study report 1 by Ney & Partners, 29/03/2011. Translation of text in blue by author

All options are described and a comparison in cost/ square meter is given. Option A is set as the 'reference option'. Options B, C, E, G and H are estimated to result in savings of up to 20%. Options G and H are possible executions in steel. These are estimated to be 10% cheaper than the shells in concrete (option A).

An 'important remark' is added: ‘The comparison has to be understood in the preliminary phase the project is in. Besides, several other related costs like the weight of foundations are not included.' Again is noted that the best way to compare prices properly, is to ask contractors to make an offer. 'Though for this, the design studies of the team are not elaborate enough yet.’

It is concluded that from a structural point of view, option 'C' (concrete with large prefabricated elements) is preferred. The option is the most economic, leaves different execution options open, appearance is controllable and execution time can be kept short. The solutions in steel (options G, H) are mentioned as well as valuable and 'to be researched', since they have the same advantages, though for a slightly higher price. These options however are not scrutinized further in detail at this stage of the design.

Observations

Although this study report elaborately discusses several technical aspects, crucial aspects such as the construction method and the related material, are left open. This choice to focus on the optimal cross section and to keep all options open, will determine the course and the outcome of the design, because the actual impact of the shape of the cross section in different options is not evaluated. Structurally, the equilibrium shape is only crucial in the case of masonry or smaller prefabricated concrete elements, in which connections should be preserved from (too large) bending moments. As well important is that structurally, the arches are of secondary importance. The arches span the sports hall in the longitudinal direction, which primarily makes them beams.

This first study report is concluded with: ‘In addition to the structural requirements, architectural, acoustic and economic factors play a role in the choice of the desired solution. The 'optimum' shape is the one that can combine all these requirements.’ [26] At last, it is illustrated that structural principles become less straightforward when implemented in real projects. Options can be detected and weighed, but it is difficult to encompass all related boundary conditions, especially in the first phases of the design process. This is true for boundary conditions from different fields, but even so for purely structural aspects.

Based on the arguments made in the report, it could have equally been concluded that a solution in steel or concrete would allow for a wider variety of cross sections, including the one proposed by the architects in the competition. The engineers thus do steer the formal evolution of the design.

It is shown that in this study report, engineers use calculation models as rhetoric tools to prove their statements. Graphical output of calculation models seem to be very effective to do so, as they have an aura of objectivity.

Catenary trump

In the months following, the search for the ideal cross section is a main focus for both the architect and the engineer. The archive contains a large amount of sketches, copies and communication on the arch shapes, all with small, nuanced differences. To illustrate the extent of this discussion, and the subtle differences in shape they seem to entail, a set of images is included (Figs. 17, 18, 19, 20, 21 and 22). Proposals are thoroughly considered by both; The architects try to find shapes that (in their perception) are close to the catenaries proposed by Ney & Partners, and the engineers make calculations to estimate the consequences of the deviations proposed by the architects. (Fig. 18).

Fig. 17

Reaction by Ney & Partners (text in red) on proposed variants of the architect for shape of the cross section of the roof. Attachment in mail 14/03/2011 [32]. Translation of text in blue added by author

Fig. 18

Studies on ‘variant 2’ (Fig. 17) for the arches as proposed by Bel architects—Ney & Partners, 6/06/2011

Fig. 19

Different options for cross sections with arch shapes composed out of circular arches drawn by the architects (to take restrictions of formwork into account)—Bel architects, June 2011. Translation of text in blue added by author

Fig. 20

Three variants for the shape of the cross section of the roof that meet the engineers restriction of an ‘optimal shape’ (with notes on advantages and disadvantages) – Ney & Partners, attachment in mail 13/06/2011. Translation of text in blue added by author

Fig. 21

Family of catenaries – Ney & Partners, 20/06/2011

Fig. 22

Final cross section with fixed points to take in account in black and a detailed scheme of requirements of free heights for different sports activities, projected from floor plan. (yellow: volleyball, green: basketball (training and competition), black: football, pink: badminton) – Ney & Partners, 22/06/2011

This discussion on the margin to deviate from the ideal shape as proposed by Ney, culminates around the middle of June, the moment on which a second study report to accompany the end of the preliminary design phase is prepared.

‘As Laurent¹³ stresses, the shape of the cross section is not free to choose in the case of a shell construction (an optimal shape in which materials work optimally.) (…) We therefore recommend choosing a shape as close as possible to the optimal shape.’[28] The engineers add three options (Fig. 20), of which only the third one is valid for the architects. They have several remarks and propose to fix 5 points¹⁴ the outline should meet [29] (Fig. 22). These fixed points indicate: The width of the building (this point must not move too much outward, since it enlarges the footprint and thus the cost), the width at gallery level (at this point, the idea is still to make a passageway at the edge of the sports hall), the free height of 9 m (a free height of 9 m was prescribed for the sports fields. In previous proposals often a small kink is introduced, which enlarges internal moments in the wall significantly¹⁵). Two other fixed points are the height of the ridge (which should correspond to the height of Isgour’s building) and the height of the longitudinal beams in the armpits of the arches.

The engineers answer by proposing a ‘family of catenaries’ (Fig. 21). These take the free height of 9m into account, since this seems the most stringent condition to them [29].

To at last find a common ground, another boundary condition is re-evaluated. Two days later, the client approves a small reduction of the free height on the sports fields. (Fig. 22) [30]. The final shape is agreed upon with a very small adjustment by Ney & partners, to allow for a certain width of the longitudinal beams in the middle [31].

Observations

After a long negotiation, the shape for the cross section is settled, and with this an important step seems to have been taken in the design process. A common definition for ‘the ideal shape’ is found, by re-interpreting boundary conditions from the client. The shape meets both the requirements of the architect and of the engineers. The empty signifier has been determined.

For the architects the ideal shape rather is one that fits all the boundary conditions. These boundary conditions can be pragmatic, such as the maximal surface of the building to limit the cost. Even so they can be the result of a chosen design priority, such as the reference they want to make to the existing building in the restriction of the height. For the engineers the ideal shape is strongly related to the ‘optimal’ use of the material. They strove for catenary shapes. It has been extensively shown that this is not a purely pragmatic objective. Calculation methods do allow to accurately estimate internal forces in all shapes one can think of, so this is no practical restriction for the determination of the shape. It as well is a choice to take ‘the ideal shape’ as a priority.

The discussion on the shape of the cross section has however so far been a purely theoretical one, since no concrete material has been coupled to the execution yet.

Fig. 23

Left: structural solution to allow daylight to enter in the valleys of the arches.—Ney & Partners, 21/04/2011. Right: structural solution to cope with thrust by integrating horizontal beams or canopies at the rear of the building.—Ney & Partners, 31/03/2011

Integrated design & reality check. Shifts in design rationale

So far the evolution of the shape of the cross sections has been discussed elaborately, but this is of course only one concern in the design for the roof.¹⁶ In the meantime several other aspects are tackled, such as the study of how to allow daylight to enter into the building or the possibilities to cope with the thrust. (Fig. 23).¹⁷

In the scope of this article, three occasions are highlighted to sketch further evolutions.

First occasion: A contractor’s price estimation

The first one is a price estimation that was received from a contractor consulted by Ney & Partners. The answer comes right before the deadline of the preliminary design phase, around the same moment as the final decision for the shape of the cross section.

In a first reply to the engineers, the contractor asked ‘why concrete, and not a solution in steel? Will the design stay in concrete or can we switch to steel? An execution with prefabricated elements will practically be very difficult.’[33] The engineers answer: ‘Concrete is a logical choice if the shape is as close as possible to the optimal shape since the concrete then works optimally, and the shell can be kept fairly thin. In addition, there is a preference for concrete from an energetic point of view (thermal mass) and an acoustic point of view. Therefore, the shell will remain in concrete in the first place. However, if you expect that it can be cheaper in steel, we will have to reconsider that option.’ [34]

The contractor’s final response for a cast-in-place concrete shell results in a cost estimation 20% higher than what was estimated before. The engineers immediately communicate this to the architects, stating that if this exceeding of the budget is not feasible, the concept of the shell might have to be completely reconsidered, for example by looking for a solution in steel. [35]

The architects reply that (another) budget increase is not possible and argue for this in 6 bullet points. [36]. They state that there already has been too much discussion on the budget and that this conclusion based on one price estimation might be premature. They as well refer to the first study report, in which a prefabricated option would be 20% cheaper and add that the client will ask on what estimations these prices were based. The fifth point states that ‘the question of how these shells can be made out of concrete, within the budget, might be a question that asks for more research by the engineers’. Several references¹⁸ are added to advocate for the feasibility of shells in concrete in Belgium. At last they come back to the proposed larger thickness of the outer shells to cope with the trust, asking of this for sure is the most logical solution. (see footnote 4, 9).

A few hours later the engineers reply by stressing again that it is very hard to make a price estimation for this project due to it specificity, and that the price of the shell will stay a point of discussion until the day the prices of the public tender are received. [37] They add that when they analyse the proposed price further in detail, they assume some posts might be cheaper, which results in a price estimation similar to their current one. They propose not to communicate the contractor’s answer to the client and to organize a meeting the next week, where Laurent Ney as well would participate.

The agenda for this meeting is the assessment of a cast-in-place and a prefabricated solution in concrete, of which the last option is chosen.

Observations

At this point, an execution in concrete seems to have become a basic assumption to execute the arches with. This can be seen in the assessment that is made in the meeting between a cast-in place concrete shell or a prefabricated concrete shell, in the set of references the architects use to advocate for the feasibility of concrete shells and in the urge of the engineers to the contractor to prepare a price for a solution in concrete. The arguments of the engineers to persuade the consulted contractor for concrete—advantages for acoustics and fire protection—are valid but equally easily argued against.¹⁹ The decision for materialization of the shells seems to have been agreed upon implicitly, rather than it has been the result of a further assessment.

Another observation is that the described mail conversation is a clear point of conflict in the collaboration. The discussion illustrates the relative position of the engineer and the architect at this point. Here, the architect has taken the lead, by stating that an increase in budget is not an option, and by seriously questioning the input from the contractor. In their arguments references are used of existing roofs with arches, pointing that something similar should be possible to design. These structures however date from another era or have a completely different structural system. The architects urge the engineers to put more ‘design’ effort in the project to find a solution in concrete and repeat the cost estimation the engineers made in the first study report. The reservations made by the engineers at that point do not seem to be taken into account. The engineers react defensive but understanding, stating that the received cost estimation might be tweaked and by organizing a team meeting, with the presence of Laurent Ney.

Second occasion: calm simplicity of the first idea as a priority

The second occasion is one of the responses of the client to the presentation that was given to conclude the preliminary design phase. The front image shows a model on which several ribs are added on the shells, indicating lamellas in acoustic material.²⁰ (Fig. 24).

Fig. 24

Front image of booklet to present the preliminary design phase.—Bel Architects, 5/07/2011

The formal reaction of the client indicates that the proposed solution has a radical impact on the spatial perception. The client regrets that the calm simplicity of the shells as proposed in the competition phase, in which various issues were solved simultaneously (stability, light, acoustics etcetera), is lost and asks to look for another solution to solve the acoustics of the sports hall [38]. Note that the client does not stress the materiality as a quality of the competition proposal.

Third occasion: acoustics as a priority

These two occasions reorient the focus of the team. They start looking for prefabricated elements in which indeed lighting, thermal insulation and acoustics are integrated. (Fig. 25) Acoustics had not been a priority in the design process before, probably because the preliminary studies in the competition phase stated that the shape of the hall did not have to be disadvantageous for the space acoustics²¹ [39].

Fig. 25

Proposal for prefabricated elements in which techniques, acoustics and ridges for stiffness are integrated. (Translation of subscripts, from top to bottom: Connection of cupola, Lighting fixture, Sprinkler) – Bel Architects, 8/8/2011

This leads to the third occasion. On the 22^nd of August Ney & Partners start scrutinizing perforations to enable acoustic material to be placed behind the structure. The ratio of the openings should be between 50 and 70%. They realize that this implies very large openings.

These openings are this large that they imply a truss system rather than a shell structure [40]. (Fig. 26).

Fig. 26

Test for perforation ratio’s in the surface. Ratio’s from top to bottom: 44%, 50%, 48%. The ratio was prescribed to be between 50 and 70%. Ney & Partners – 22/08/2011

The ideal shape: a three hinged arch and a steel truss system

The engineers start looking for perforated prefabricated elements, in which at first the execution in concrete is questioned again. Such tests are shown in Fig. 27. Note that the shape of the catenary cross section in nowhere questioned again.

Fig. 27

Stages in the development of the (concrete) truss system. Left: triangular net with horizontal lines (24/08/2011, Ney & Partners). Right: triangular net with verticals.(09/09/2011, Ney & Partners)

On the 19th of September, Laurent Ney has a phone call with Bel Architects, as can be seen on the notes in the archive of the latter. The note ‘steel = realistic’ is framed. This leads to the final mail discussion that will be highlighted in this article.

The architects send a long e-mail after their ‘short conversation’ with Laurent Ney. [41] They do understand that steel can be a beautiful solution, if the resulting impression is related to the one proposed in the competition. A long list of elaborate questions follows, ranging from the exact amount of savings this steel structure will entail (‘up until now it has always been said that a steel structure would be the most economical. We are thus interested in the exact margin on the budget’), how flashover for fire safety will be solved, what build-up will be proposed for the roof structure etcetera. The engineers react rather reserved, answering to several questions by questioning what is expected of the structural engineers or by stating that these questions are not for them to be solved.²² [42]

Following these answers, the architects conclude that it will be better ‘to continue the path they have been taking since the competition. A building with a shell structure in concrete: cast-in-place or prefabricated. Based on the vagueness of your answers, it seems impossible to justify this proposal to the client. To us, it seems the task of the engineer to make a feasible proposition that matches the image proposed in the competition, within the cost estimation proposed by you’. [43]

Laurent Ney replies internally to his colleagues, with the question to verify: ‘I thought I was clear before. A solution in concrete with the acoustic requirements as they are now is, in my opinion, just nonsense. (…) The architectural quality of your design is in the SPACE. I think it is a pity to endanger the design with such an inefficient choice of material. I thus formally advise against continuing the design with a concrete solution. The risk in terms of budget is too large. Thank you to take this remark in account.’[44]

It is not clear whether this message was actually sent. The next day during a meeting the structure in steel is validated.

The decision to change the structural system initiates a last recalibration in the definition of what the ideal shape means. The clarity and simplicity of the final structure in longitudinal direction only gradually grew into the design of the steel structure.

The idea to make consecutive vertical three hinged arches emerged around the time of the shift in structural system. (Fig. 28) A three hinged arch has two advantages: ‘All of its forces and stresses can be completely and precisely determined by hand calculation methods.(…) Another advantage is that they are more forgiving in the sense that, for example, the vertical settlement of one support will not occasion supplemental stresses in the arch segments.’[45] This again fits the goal of the engineers to make optimal use of material. It is not necessary to make such three hinge arches today, as mainstream advanced calculation methods can easily cope with highly indeterminate systems and the ratio of the height over the span of the truss results in a very rigid system.²³

Fig. 28

First calculations for three hinged arches. (Translation of notes: Left: HEB300 S235, laterally supported every 4 m (secondary structure) Right: HEB 340 S235, laterally supported every 4 m (secondary structure)—Ney & partners – 19/09/2011

In the final structure 4 longitudinal steel trusses are integrated in the valleys of the arches. The length of the building is divided in three parts: two trusses span 32,5m, with a smaller one that spans of 17,5m in between (Fig. 29).

The diagonal ends of the trusses are less optimal from a structural perspective. Here the structure made place for technical shafts integrated in the columns. Due to the very low self-weight of such a system, thrust is of a less decisive order than in case of concrete. The trusses are as well tilted in the extension of the arches, which is why they can help deviate the thrust to the supports.

Fig. 29

Left: sketches of the final trusses in longitudinal direction. Right: picture of the executed structure

Observations

The final argument for a steel structure is made by (re)defining the essence of the design. The essence is not in the nature of the structure or the material, but in the generated space. Here again however the ‘efficient’ use of materials is mentioned by the engineers.²⁴

In this second discussion, the roles are the other way around. The engineers take the upper hand. Ney informs the architects by telephone on their insight to make the structure in steel and they answer the long list of questions in a concise manner. It is interesting to note that in these questions some rather inaccurate statements are made. The architects for example say that ‘throughout the process, steel has always been said to be the cheapest solution’ or that ‘they are working with concrete since the competition’. As well interesting to note is that again Laurent Ney takes the lead on a pivotal point in the design process, although he never makes a decision but only ‘advises’ the architects to make certain choices.²⁵

The answer of the architects to the first message in which steel is proposed, is exemplary for the general image the architects have on the work of the engineer. ‘To us, it seems the task of the engineer to make a feasible proposition that matches the image proposed in the competition, within the cost estimation proposed by you.’ [43] What is ‘feasible’ has no clear definition and is something that has to be scrutinized as well throughout the process for each specific project. It is interesting to explicitly read again that this would be the realm of the engineers. Of course the engineers managed the cost estimation for the roof structure and so it is fair that the architects point their responsibility. On the other hand it can be seen in the early stages of the process, that the engineers as well do support the architects in their bold proposition, at a point where the budget is as well deliberately stretched for the sake of the idea and that they do make reservations on these estimations.

It is specific to this research to highlight such discussions between collaborators. It however has to be stressed that these discussions are common (or even vital) to collaborations as much as mutual understanding. That the shift to a steel structure in the end was supported by the architects, can be understood from their narration in an interview: ‘At first, barrel vaults were planned, but in coordination with the engineer, it was decided to use vaults that follow the more efficient catenary line. Even after the design phase, stability plays an important role. "We received a phone call from Laurent Ney to say that the roof could not be realised in concrete, but in steel," says Bel architects, "thanks to this change, we can meet the strict fire and acoustic requirements. A concrete roof would become unnecessarily complex, due to the difficult integration of more and more techniques and the in-situ implementation of concrete.’ [49]

Final observations

In this article, the design process for the roof of a sports hall is illustrated. Several observations on six pivotal points in the process are made and elaborated upon at the end of each subchapter. A short overview can as well be found in the introduction of this article. To conclude, some additional overall observations on the case study are formulated.

The team won the competition due to their bold move to propose a second building rather than an extension to the prominent, hypar roofed, seventies building. The relation to the existing building has been one of the most important concerns for the architectural identity of the new sports hall. In the early design phases and the competition, the height of the arches, the concrete as material and the shell as a structural system seemed to be the necessary aspects to let the new building conversate with the existing one. In the end, the arches were not executed in concrete and structurally they are no shells. Looking back, it could be concluded that technically, the emphasis on the catenary shape of the cross section was not necessary. Steel can easily cope with bending moments that would be introduced due to a less ‘ideal shape’.

It however is exactly through the formal evocation of a shell roof that the aimed for relation with the existing building is obtained. This evocation of a shell roof is made through the catenary shapes of the arches. Examples of the effectivity of this evocation can be found in the title of an article on the project ‘steel roof in conversation with existing roof shells [47], or in the question of an interviewer for the origin of the ‘three shells’[4]. This is why another conclusion is made for this case. The collaboration of the engineer and the architect has been crucial for the final design of the project, in which decisions on both architectural and technical aspects of the building are closely intertwined. It would be a too large simplification to evaluate the structural design without taking in account the complexity of the process.

In a collaborative design process, and even more in designs where the engineer is explicitly involved in the conceptual design phase, a mutual design rationale has to be found. The first phases of the design are about the quest for these storylines. The exact definition of this mutual rationale is subject to the further process. In this case, the concept of the ‘ideal shape’ emerges in the first phase of the design and both the architect and the engineer agree that this should be one of the design rationales for the project. Due to the different epistemological conditions the architect and the engineer operate from, but as well because they have their own identities in the design process, the team members have different priorities for the concept of an ideal shape. Looking at the concept of ‘the ideal shape’ as a floating signifier in this design process, facilitated to evaluate how the collaborative design process between the architect and the engineer evolved, how the mutual definition was negotiated and how conflict and mutual understanding are key in a such a collaboration.

The attitude to keep looking for the optimal shape illustrates the engagement the engineers as well as the architects have towards the design. ‘Often, the engineer is the watchdog of architecture, watching over buildability and affordability. Sometimes, however, a more intensive interaction develops, a reflection on stability in the genesis of the project so that structure and architecture become one.‘ [49] In this case the involvement of the office for structural engineering proved to be guiding and crucial for the final outcome of the design.

An engagement towards design implies an understanding and an ability to cope with the uncertainties of a design process, especially in the first phases of a design. The projective rather than decisive character of the first phases of the design imply a premature nature of statements. This is own to the early phases of a design process, in which a competition is inevitably situated.

It seems to be generally misunderstood that the engineer’s work, often considered as more scientific and thus ‘objective’, would be able to escape this reality of premature statements in the early design phase. The implementation of structure in reality is far more complex than the basic structural laws the engineer has in his toolbox. When engaged towards design, both cooperating partners go through an equivalent kind of creative design process. Finality of the statements made are largely related to the stage in which the design is, and this applies for the engineers as much as it does for architects.

In this case study, it is shown that the reality of this misconception is used in the design competition. The team’s selection is motivated with the participation of an experienced engineer that guarantees a feasible concept and the competition proposal is laced with several technical notes. This proves to be effective to win a competition, but as well introduces tensions and misunderstanding in the further collaboration. Acknowledging this reality would allow to bypass this misconception and might fertilize even more fruitful collaborations.

In the realms of engineering, this understanding of fluid and uncontrollable nature of the course of a process and how engineers can engage in a creative design process, is largely underexposed. At the same time, although it is generally accepted that it is beneficial for projects when engineers are involved from the start, the full consequences of engineers engaging in (early) design in collaborations neither seems to be widely understood by partners in the design teams. Although all particular on their own, such case studies are exemplary for the actual courses of complex design processes. Such references can be very helpful for (young) designers, both architects and engineers, to better understand the complex reality in which they daily operate.

At last, the research method proves to be very effective to scrutinize the design process in a profound manner and allows to deepen the course of events sketched in discourse. Several cited interviews prove that they are illustrative for the general course of the design process. This case however shows that they are large simplifications of the actual course of events. The research exposes the messy reality of a design process and allows to make a wide variety of observations. It is illustrated how relative positions of team members evolve throughout the process, how progressive insight, conflict and mutual understanding are key during the design and how the design evolves gradually towards the final result.