The engineering legacy of the FIFA World Cup Qatar 2022: Al Janoub stadium

The first new-build stadium designed for the FIFA World Cup Qatar 2022 and the centrepiece of a new public sports precinct to the south of Doha, Al Janoub stadium in Al Wakrah was a true path-finder project.

AECOM was appointed in late 2012, providing full multi-disciplinary engineering and architectural design services in partnership with Zaha Hadid Architects. Working closely with the Supreme Committee's (SC) programme management team and subject matter experts, a comprehensive technical brief and performance targets were developed for a world-first building – a fully cooled 40 000 seat open-air stadium venue (in tournament mode) that could hold World Cup matches in the summer of 2022 and later reduce to a 20 000 seat capacity after the tournament.

At the time, the potential to move the tournament to the winter had not been agreed, so the team embarked on this innovative project to ensure that a summer tournament could be delivered, and in doing so helped to set the programme-wide standards and procedures for technical design, computational analysis, building information modelling (BIM), safety and sustainability that informed many of the other stadia that followed.

The following sections describe the design process, from the brief development and early concept stages and through the detailed design phase prior to handing over the final delivery to the design and build consortium. Aspects of the value engineering (VE) exercise carried out by the design and build team as well as at the construction stage are also presented in this paper.

Al Wakrah is one of Qatar's oldest inhabited areas, and the brief called for a stadium form inspired by its rich seafaring heritage. This is visible in the colours, patterns and materials found throughout the stadium, but especially in its overall abstract expression of the dhow – an Arabian pearl-fishing boat of great cultural significance in Qatar and a common sight in Al Wakrah's port.

The dhow interpretation is particularly noticeable from eye-level, and is expressed both in its elevations and overall volume, as shown in Figures 1 and 2. The north and south ends taper strongly upwards towards the roof, resembling the bow and stern of the boat.

Although the timing of the tournament was changed to the winter of 2022, the design conceived and delivered by AECOM was capable of operating in the peak of summer while meeting stringent FIFA guidelines for player comfort (FIFA, 2022) as well as the comfort of spectators and officials.

Furthermore, matches needed to be played on natural turf or a hybrid pitch and under an ‘open sky’ – that is, a fully closed roof could not be used during a match. Summer temperatures in Qatar can reach over 50°C, with surfaces exposed to direct sunlight reaching temperatures of over 85°C. Therefore the need to control the internal environment efficiently and effectively drove the team's holistic approach to the design.

In the authors' experience, many of the world's best stadia are designed from the inside out, and integration of the bowl and roof design is essential in generating a holistic design. The bowl form is in a three-dimensional (3D) parabolic geometry to ensure excellent sightlines from every seat, while the accommodation behind it is a complex arrangement of technical, welfare, hospitality, media, broadcast and commercial areas. The roof is defined by a complex interaction of factors, including architectural form, shelter, insulation, pitch conditioning, acoustic atmosphere and sound quality, ventilation, the positioning of lighting, speakers, gantries, audio-visual and communications technology, camera and broadcasting platforms, rigging systems, drainage and fire protection.

Integration of the cooling systems for Al Janoub stadium added a further dimension to this, and the early concept development focused on the following key drivers, which were often contradictory and in conflict with each other.

• The provision of shading and insulation to cool the stadium interior passively, using a combination of fixed and operable roof and envelope elements.

• The integration of extensive mechanical cooling systems and plant while maintaining clean, open concourses to enhance spectator experience and crowd management.

• Minimising ‘wind scour’ entering the bowl, displacing cool air with warm air from outside the stadium.

• Ensuring adequate sunlight and air movement at pitch level to limit the amount of active ventilation and grow-lights required for growing and maintaining healthy turf growth (directly contradicting the shading and wind-scour requirements above).

• Providing an optimal fan experience by way of a continuous stadium bowl with uninterrupted sightlines and a ‘column-free’ roof structure both in both tournament and legacy modes.

• The provision of highly resilient electrical and mechanical systems to ensure play and broadcast could continue in the event of power failure, using both multiple utility circuits and generator/uninterruptible power supply back-up.

• Ensuring that the post-tournament transformation to 20 000 seat capacity could be carried out safely with minimal disruption to the permanent accommodation or roof.

• Limiting visible or distracting shadows on the pitch, which can impact the quality of television broadcasts.

• The incorporation of state-of-the-art lighting and audio systems to enhance the stadium atmosphere.

Optimising between these factors required significant investigation to ensure the final design was driven by science-based data as well as sound engineering judgement.

The stadium was conceived as the destination at the end of a ‘thermal journey’, which included cooled transit systems, shaded walkways and the use of sprayed water mist to cool people as they approached and entered the arena.

Human beings experience the temperature of their environment in a number of ways. The standard effective temperature (SET) is defined in the ASHRAE 55 comfort model (ANSI/ASHRAE, 2020) as a single temperature representing how comfortable a person will feel in a given environment, depending on a combination of clothing insulation, activity level, air velocity, air temperature, radiant temperature and relative humidity. All of these factors contribute to the sensation of being too warm or too cold.

FIFA guidelines for player comfort and safety are based on the wet bulb globe temperature (WBGT), as defined in BS EN 27243:1994 (BSI, 1994), which combines the variables of air temperature, humidity and radiant temperature into one measure. The WBGT is heavily dependent on the wet bulb temperature (humidity) and also the radiant temperature, which can be as high as 70°C in Al Wakrah.

In collaboration with the SC, thermal comfort compliance criteria were defined as 26°C WBGT on the pitch and 28°C SET for spectator areas.

In order to minimise the operational energy required to cool spectators and players, a cooling system was developed to create a cool air ‘bubble’ around occupied areas, rather than trying to cool the larger volume within the stadium envelope. It was essential that the air velocity and temperature were not uncomfortable or disturbing, that the temperature gradient experienced from head to foot was within a reasonable range and that thermal shock was avoided.

This was achieved using displacement cooling, whereby cold air was fed from chillers and air handling systems into a pressurised plenum below the terracing. Openings in the terrace units were coordinated with structural requirements and seat fixing details to allow the cool air to distribute from each row. Return air was fed into the system from vents located in the concourse area and through the building envelope. The sustained temperature differential across the structure was accounted for in the design.

The pitch cooling adopted some ‘borrowed’ cool air that flowed down from the grandstands, supplemented by air jets located around the perimeter of the pitch to throw cool air across the whole pitch area to give a more uniform playing environment.

In order to simulate these effects, highly innovative transient energy modelling, combined with computational fluid dynamics (CFD) modelling, was carried out to determine the impact of different roof oculus sizes, the extents of fixed or operable shading devices and different volumes and velocities of cooled air being supplied to the bowl. This time-dependent analysis included

• radiant heat from direct sunlight and reflected from the internal surfaces of the stadium

• convective heat components from the air, including the effects of air movement within the bowl

• diurnal and seasonal variations in sun path

• the potential benefits of ‘pre-cooling’ the bowl before a match, using a higher velocity of cooled air and closing the operable roof to contain the cool air

• the thermal mass benefits of precast concrete terracing as opposed to lightweight systems

• the potential impact of wind scouring the cooled air out of the bowl and ensuring the bowl temperature would stabilise within a reasonable timeframe after such an event.

The same baseline parametric geometry data as described in the roof and bowl design sections of this paper (Sections 6 and 7) was used in CFD and energy modelling packages such as Ansys (Ansys, 2023) and IES VE (IESVE, 2023) for spectator comfort, ensuring the optimisation process included all relevant parameters. Following a series of initial sensitivity and optimisation studies to determine the baseline roof shape, the extent of shading and aerodynamic effects and heating and cooling effects were simulated hour by hour for typical and worst-case radiant heat (sun path) scenarios for each month of the year. The initial simulations were based on the maximum cooled air volumes and velocities that could be delivered by the bowl and pitch cooling systems without impacting on spectator and player comfort.

The associated airflow was first simulated within the bowl to include air displacement and buoyancy effects based on a static background air movement (i.e. ignoring weather effects). The effects of external wind scouring the air inside the stadium were then combined with the baseline results using probabilistic methods based on historical and projected weather data.

Through iterations of this analysis, the final optimised roof form and air delivery design was determined, and a series of graphical visualisations was produced (Figure 3) to show the probability of overheating seat by seat and zone by zone. It was demonstrated that, for mid-afternoon matches under reasonably anticipated weather conditions, the required comfort conditions could be met for a summer tournament in all seats.

Some air movement at pitch level is desirable in order to assist with the cooling process and promote the growth of strong turf. However, even modest gusts or sustained breezes could act to scour cooled air from the bowl and replace it with hot air from outside.

The CFD simulations generated by the building physics team were used to inform the design, and they were subsequently validated by a series of wind tunnel tests on 1 : 300 scale models carried out by wind specialists RWDI in the UK (Figure 4). In addition to the use of conventional pressure taps to calculate the structural and façade wind pressures, a series of smoke tests was carried out to review the aerodynamic properties of the roof and to validate the CFD studies for varying wind directions and velocities. The physical tests were found to correlate well with the CFD studies.

Final wind tunnel tests on the design and build geometry were carried out at the Politecnico di Milano.

The site was previously undeveloped, with the exception of a utilities corridor crossing the site, which was diverted as part of the enabling works.

An extensive site investigation was carried out to determine the quality of the Simsima limestone rock, which lay below a superficial deposit of wind-blown sand approximately 1 m deep. A combination of geophysical and intrusive surveys (boreholes and trial pits) was used to check for fissures and solution features that can be common in this environment; generally, the conditions were found to be favourable. Groundwater was present at approximately 10 m from ground level, which required the team to minimise the depth of excavations where possible. Contamination was also minimal, with just a small area of the site (about 20 m × 20 m) used for commercial waste disposal, which was cleared prior to construction.

Founded on limestone rock, the building is mostly supported on a combination of raft and pad foundations. The allowable bearing pressures were highly dependent on the rock quality, and a lower bound of about 500–600 kPa was adopted for most areas. The formation level was inspected on site prior to construction.

The highest loaded primary roof columns are supported on 1200 mm dia. reinforced concrete (RC) piles, bored to various lengths into the limestone to resist the combination of axial loads, uplift and overturning effects.

The basement levels comprise conventional RC retaining walls, with the limestone rock allowing steep cuts to be made safely and avoiding the need for extensive volumes of cut and fill. Waterproofing was achieved by a combination of tanking (externally applied waterproofing membranes) and water-resistant concrete admixtures.

The 3D Revit (Autodesk, 2023) model used throughout the design process was an essential tool; the extent of coordination required being comparable to that of a modern process plant or industrial building rather than a stadium, as can be seen in Figure 5. At the time, it was believed that the level of detail and information in the BIM was the most comprehensive and detailed of any major global stadium project, setting new modelling standards and component libraries for many future projects, and allowing 5D material scheduling and 4D construction phasing to be reviewed at each project milestone.

The brief called for a venue that could hold a net capacity of 40 000 during the World Cup tournament, but which could then be converted into a 20 000 capacity in legacy mode to serve Al Wakrah football club and local community requirements. This conversion process was also key to the long-term sustainability of the venue.

It was important that the overall architectural form was retained for both tournament and legacy modes, so the team embarked on numerous studies to determine the optimum arrangement.

The final stadium bowl section (Figure 6) consisted of a permanent in situ RC lower tier frame supporting precast concrete terracing and a temporary demountable upper tier constructed in steelwork. The demountable upper tier structures were built on a podium slab that would provide a flexible structure for future legacy developments. Design studies to facilitate the future development of covered markets, hospitality, leisure and entertainment zones, education and sports clinics were carried out to ensure the structural capacity and future servicing of the legacy venue would be sufficiently flexible and adaptable. Each structure was designed to be capable of being erected and dismantled multiple times, allowing the components to be transported to different locations around the world and used in numerous configurations to suit small-scale venues and temporary performances, as well as larger scale permanent structures.

The demountable terrace units were of similar design to the permanent stand, with openings provided for cool air to permeate from the plenum below. The steel raker beams were castellated or trussed to allow free air movement within the plenum as required.

A series of deconstruction storyboards was carried out to outline the transformation phase. A phased Revit model (Autodesk, 2023), combined with clash-detection tools developed in Rhinoceros 3D (RMA, 2023a) and its algorithmic plugin Grasshopper (RMA, 2023b), was used to ensure that there was sufficient headroom and access for mobile lifting equipment to demount the temporary structures without impacting on the roof structure, building envelope and other parts of the building retained for the legacy mode. An excerpt from these models is shown in Figure 7.

Digital workflows were adopted for the stadium seating bowl design. Bowl setting out is governed by a huge array of factors and functional requirements, most of which are interrelated by complex geometrical parameters. Most stadium architects now use scripting of some form to generate the baseline geometry, and add layers of bespoke parametric tools with their own experience to rapidly test options and hone in on the optimum fan experience.

For the structural engineer, the design calculations are often relatively straightforward. Terrace units are simply supported beams, and raker beams and columns are basic framed structures, albeit with large forces and sometimes complex dynamics and human–structure interaction to consider. However, traditionally, the most time-consuming aspect is the generation of the structural geometry, especially for bowls with radial grids, parabolic section geometry and cookie-cutter profiles. This is particularly true when the architecture will continue to evolve and change throughout the project duration as stakeholders and brief requirements also evolve. Rhino and Grasshopper were adopted as a data hub to take the basic bowl surface and define the geometry of all of the supporting structure. This included the precast terracing, raker beams, raker top profiles and primary columns. Components were sized from a pre-optimised dataset for varying spans and loads.

The bowl was designed in accordance with Dynamic Performance Requirements for Permanent Grandstands Subject to Crowd Action (IStructE, 2008) for scenario 3: ‘Commonly occurring events including, inter alia, high profile sporting events and concerts with medium tempo music and revival pop-concerts with cross generational appeal’ (IStructE, 2008: p. 3). With the possibility of the stadium hosting music concerts and other events in legacy mode, the permanent lower tier was also verified against the more onerous scenario 4: More extreme events including high energy music concerts with periods of high intensity music’ (IStructE, 2008: p. 3).

Due to the desire to achieve a lightweight, economical and demountable structure for the temporary upper tier, and the architectural requirement to limit the amount of visible bracing and disruption to the concourses, the design adopted a route 2 analysis as defined in IStructE (2008). Rather than being based on limiting the natural frequencies of the bowl structure to 6 Hz, which can result in an inefficient structure, the route 2 assessment considers human–structure interaction for all modes of vibration and superposes their results to determine the acceleration of each point of the structure. This allowed the designers to develop an economic demountable grandstand solution that not only achieved scenario 3, but also allowed the design team to identify what additional event management or temporary propping requirements might be required to upgrade to a scenario 4 requirement if needed.

The roof comprises a fixed outer roof and operable inner roof. The operable inner roof is described in more detail in a companion paper.

The outer roof consists of a quarterly-symmetrical arrangement of long-span steel arches, shown in Figures 8 and 9. The primary steel roof structure was designed to be independent from the RC bowl structure down to foundation level, ensuring that the long-span steelwork retained a symmetrical and predictable stiffness distribution under the various applied loads.

The primary arches are formed in efficient 3D trusses. They span the length of the building, approximately 230 m, and their primary supports are large sculptural fabricated steel buttresses on the north and south stands. In the four side quadrants, secondary arches are supported by cantilevered cores of fabricated steelwork and RC. The cantilevered cores were subsequently changed to raking concrete towers in the design and build stage.

The arches curve on plan and in section to follow the profile of the central roof aperture. Additional trusses cantilever from the buttress structures to create the feature overhangs at the north and south ends of the stadium.

The primary arches work together with the secondary truss arches, which create the roof forms over the east and west stands. The secondary truss arches support the secondary roof structure and also act as props to the mid-span of the primary arches.

A continuous perimeter ring truss constrains the horizontal spread of the roof structure and is supported on inclined perimeter columns that are integrated with the façade.

The structure between the trussed steel arches was originally proposed to be a series of parallel glulam timber arches. This acted to complement the architectural representation of a boat construction and was also proposed as a lower embodied carbon dioxide (ECD) alternative to traditional steelwork. Extensive research was carried out to determine the viability of engineered timber in the aggressive environmental conditions in Qatar. In addition, a back-up structural steel option was also developed as a cost-saving alternative, and that was taken forward in the final construction.

The roof supports a 360° access gantry and an extensive array of sports lighting, architectural lighting, sound systems and localised cooling equipment, suspended from the primary and secondary structures above.

The structural geometry is deeply intertwined with the expression of the architectural shape. Parametric optimisation techniques were vital to allow various design options to be tested and prototyped from concept design through to the detailed design stages.

During the early concept and scheme design stages, simplified parametric models of the roof using 1D elements with stiffness characteristics representative of a prismatic truss were used to investigate the impact of changing various parameters, including the arch spread and lift, the roof oculus size, and the support and connection stiffness regimes. These sensitivity studies helped to optimise the overall structural form (and with it the steel tonnage, envelope surface area, cost and ECD) in advance of embarking on more detailed design studies, ensuring that the engineers fully understood the global behaviour of the structure. The studies were also fed into multi-objective optimisation investigations to include the pitch shading, wind scour behaviour and minimisation of the envelope surface area. These studies were further developed in the design and build stage in the definition of the final shape.

In order to respond to these processes efficiently, the design team developed scripts, principally based around Rhino and Grasshopper, as for the stadium bowl. These scripts acted as a data hub with a multi-functional, multi-objective approach. The programming environment was designed to transform initial inputs into a structural geometry, which was then redirected into several output streams and interoperable software for detailed analysis, multi-disciplinary coordination, technical documentation and scheduling.

As the tender design developed, the geometry was progressively defined in more detail, ultimately including the setting out rules of every roof member but still driven by the same controlling geometry. This included in-built offset rules for the truss lacing to meet CIDECT (Committee for International Development and Education on Construction of Tubular Structures) connection requirements (Wardenier et al., 2008), thus allowing the geometry to be exported directly to connection design software. This allowed the member size and joint stiffness assumptions to be validated in an iterative process. The final connection design was to be carried out by the contractor's engineers; the completed analysis models are shown in Figure 10. However the importance of the connection design in a high-performance structure like the Al Janoub stadium roof meant that the design intent and enveloping of forces needed to be very clear from the outset.

Thermal loads were investigated in detail from the outset due to the extremes of temperature that can be experienced in the region. The roof design was conceived with no movement joints, and relies on its inherent flexibility and articulation at the primary supports to expand and contract in response to the varying temperatures. The arches naturally ‘breathe’ up and down, with the steel supports rotating on large pinned connections and bearings. The roof was checked for a global temperature range, but also for pattern loading to reflect the diurnal variation in temperature created by direct sunlight and shading, and of the partially clad structure during construction. The pattern loading was found to have only a nominal impact on some of the local design, but global design was dominated by the overall temperature range.

Robustness against accidental damage and fire events and security and counter-terrorism requirements were also important considerations as the design developed. In liaison with the SC and Qatar Civil Defence, a comprehensive approach to security was implemented on the project with a combination of passive and active measures.

For the primary structures, avoiding a single-point failure and subsequent progressive or disproportionate collapse drove a number of the early design decisions. Analysis of the roof and its supporting structure included notional member removal, relying on alternative load paths and the mobilisation of catenary action and large deformations to avoid progressive and disproportionate collapse. The concept of interlocking arches, the symmetrical support and stability arrangement, and the inner roof truss and garages acting to tie the arches together all contributed to the overall robustness of the design. For the main building elements and smaller scale structures, the best-practice provision of continuous vertical and horizontal ties and key element design was adopted throughout.

Such considerations were essential in validating the early concepts as well as during the future detailed design stages.

The final tender design validation of the roof included a full linear and non-linear analysis to account for global and local imperfections and dynamic performance, as well as running an outline sequential construction-stage analysis to determine the impact of locked-in stresses. The latter was based on an assumed erection methodology that was provided as part of the tender package. The effects of sequentially laying out the rafters (their spread resulting in a progressive increase in lateral thrust on the steel truss arches) were considered, as was the impact of the central inner roof truss that acts to prop the primary arch trusses and tie them together. The final methodology, adopting intermediate trestle supports for the primary arches and sequential de-propping (removal of the temporary supports), was very similar to that assumed at the design stage (see Section 11).

AECOM collaborated with Schlaich Bergermann Partner (world leaders in operable fabric roof designs) to develop the concept and detailed design for the inner roof. A detailed description of the operable roof is given in a companion paper, but early renders of the combined inner and outer roof configuration are shown here in Figures 11 and 12. The operable inner roof was required primarily as a shading device, allowing the stadium to be ‘pre-cooled’ in a controlled environment prior to an afternoon match. Therefore it could be a lightweight structure with only nominal acoustic performance and did not need to be fully weather-tight.

A number of arrangements were considered at concept stage, including central garages and radial cables, inflatable panels and more complex mechanical systems with solid panels and angled fins. The chosen concept was unique at the time, adopting a folding fabric shade that retracts from a central truss gantry and into a pair of garages located at the north and south ends of the roof.

Prestressed cables span between the garages and the central truss. The cables support carriages that are fixed to the fabric and are operated by electrical winches. The fabric is progressively prestressed as it traverses the oculus, by virtue of the fanned arrangement of the cables. The final prestressing upon closure is carried out by hydraulic actuators located on the central truss gantry.

The effects of pattern loading, partial closure and ‘accidental’ scenarios resulting in uneven loads were all considered in the design, impacting on both the inner and outer roof designs.

Different fabrics were considered for structural performance, foldability, maintenance requirements and other performance parameters. Wind tunnel testing included aeroelastic tests to ensure it was not susceptible to dynamic excitation. As predicted, the structure was sufficiently stiff to avoid these effects.

The challenge of creating a comfortable environment for players and spectators, packaged within a dramatic and iconic architectural form, required a completely integrated approach from every design discipline.

Although sustainable design has been on the agenda for many years, it is only recently that the construction industry as a whole has made a concerted effort to truly understand and measure the lifecycle carbon dioxide emissions of a building. In 2012, the goal set out by the SC was for a LEED (Leadership in Energy and Environmental Design) (USGBC, 2023) and GSAS (Global Sustainability Assessment System) certified building (GSAS, 2023) building, but it also recognised the need to create more bespoke and challenging performance targets for the Al Janoub precinct during both the tournament and its legacy, which included setting targets of kilogrammes of carbon dioxide ‘per seat’ for the stadium itself.

ECD metrics were established for the local and international supply chain that would be needed to deliver the stadium programme. The early concept developments were measured against approximate area- and volume-based ECD calculations. The ECD of the structure and envelope, considered to be the greatest contributors to the total ECD of the building, was assessed for different options using the parametric model, and subsequently using carbon dioxide parameters built into the primary components of the highly ambitious BIM. Such benchmarking and the use of BIM to monitor carbon dioxide metrics throughout the design process should now be commonplace under the Engineers Declare climate and biodiversity emergency mandate (Engineers Declare, 2023), but at the time this was a very forward-thinking requirement from the SC and it played an important role in focusing the engineering team to adopt lean design principles and intelligent material selection. This included, for example

• working with the SC to establish a green procurement framework, including the stipulation of recycled content for steelwork and reinforcement, challenging limits on wastage and setting minimum cement replacement requirements

• adopting sustainably sourced timber

• specifying ISO 14001 supply chain compliance (ISO, 2015)

• optimising the cut-and-fill balance across the site and reusing arisings where possible for sub-base and fill material

• minimising areas of hardstanding and maximising soft landscaping

• maximising the potential for off-site prefabrication

• operational energy assessments and the use of low-energy and low-water solutions where practical in the building services, building management systems and utilities designs

• designing for a sustainable post-tournament legacy with demountable and reusable components.

The decision to move to a winter tournament was confirmed late in the detailed design phase of the project. This possible change had been anticipated by the team and so a series of studies was collated to determine the changes that could be made to the design with the reduced cooling requirements that a winter tournament would allow.

It was noted that it was still desirable for the stadium to perform as a summer venue in legacy mode, but the performance criteria could be relaxed in this scenario. Therefore, it was determined that the following changes could be made to the design.

• Reduced capacity and area of cooling plant, allowing a reduction in the overall building volume and surface area.

• Increased fixed roof oculus size due to reduced requirements for shading and insulation.

• Reduction of the building volume gave opportunities to re-plan some internal spaces and further optimise the amount of temporary tournament accommodation that could be located externally.

Some constraints existed, as an enabling works package comprising the foundation, basement and parts of the lower tier were already under construction, but the design was subsequently adapted to suit the revised performance criteria as part of the design and build main contract, which allowed a number of significant cost savings to be made.

In light of the change to a winter tournament, the reduced content of the wider sports precinct and a programme-wide cost model review, the SC requested a VE exercise from the contractors that participated in the design and build tender. The goal of the VE was to reduce the built volume and redefine the stadium shape, while still respecting the original concept of Zaha Hadid Architects. This challenging VE activity had to consider the constraint of the foundations, which were largely built on site as part of an earlier enabling works contract.

The main VE items were as follows.

• Redefinition of the shape of the stadium in order to reduce the dimensions and optimise the costs (see Figure 13).

• Revisions to the internal layout and finishes.

• Optimisation of the roof steel structure and the retractable roof structure to suit the updated geometry.

• Optimisation of the façade structure.

The VE proposal allowed for a saving of around 30% of the initial budget.

At the start of the VE design activities, the foundations were almost completed on site. The VE design tried to keep as much as possible from the existing foundations, however localised changes included demolitions, cutting of existing starter bars and casting of new foundations and piles to adapt the foundation to the new structural layout. For example, the secondary arch supports, stability cores and walls in the four corners (shown in Figure 14) needed to be repositioned and new foundations constructed.

The roof design concept was kept similar to the original design intent, but some changes were adopted.

• The secondary and tertiary beams (shell beams) were changed from glulam to steel in order to reduce cost and weight. In addition, the axial releases on the secondary and the tertiary beams required to control the relative movements of the steel and timber elements could be removed; thus the secondary system could provide a tying and stiffening effect to the main roof trusses, allowing for their optimisation.

• Originally, the four main columns supporting the roof consisted of 3D truss columns fixed at the top and hinged at the base. These were changed to concrete raking columns instead, with a fabricated steel connection to the roof, as shown in Figure 15. This modification enhanced the stiffness of the structure, with particular regard to the horizontal actions.

• Optimisation of the roof trusses. An optimisation exercise was carried out considering the connection design. As the roof trusses were designed using circular hollow sections (which could be affected by punching/local failure issues), the exercise considered the nodes in detail and ensured that most of these nodes did not require any internal stiffeners, thus simplifying the fabrication process. This was carried out using detailed finite-element analysis (FEA), as shown in Figure 16.

A new wind tunnel test was carried out on the updated roof geometry at Politecnico di Milano to determine the most accurate pressure distributions and allow for the maximum optimisation of the steel structure. A desk study of the dynamic amplification of the roof pressure was also carried out using FEA (Figures 17 and 18), considering the long span of the trusses and the relative slenderness of the roof structure. The calculated natural frequencies were used to determine the amplification of the baseline pressures measured in the wind tunnel test, resulting in the final pressure maps used for the structural design (Figure 19).

The roof cladding (Figure 20) consisted of

• an aluminium composite sheet for the external face

• a perforated steel sheet for the internal face

• steel framing and insulation between the two sheets.

The cladding panels are supported by a steel structure supported on the secondary and tertiary roof beams.

The roof shape was studied with a Grasshopper script in order to ensure that all the cladding panels were buildable. Particularly, the goal was to have planar panels in all the possible locations. The maximum allowable out-of-plane deviation of an unloaded panel was set to 10 mm and the overall panel performance was verified against the calculated relative deflections of the supporting structure, as shown in Figure 21.

The roof truss segments were assembled on site in a specific assembly area, moved with self-propelled modular transporters and lifted with crawler cranes. Temporary towers were used to support the roof segments until completion of the roof structure (Figure 22).

Tandem lifting was performed for the heaviest segments, as shown in Figure 23. The ring truss sections shown in the photograph weighed approximately 125 t, but tandem lifts for the primary trusses varied from 225 t to 370 t.

Once erection of the main elements had been completed, the roof structure was then activated by the controlled and simultaneous de-propping of the temporary supports (Figure 24).

The design of sports stadia is one of the most demanding and enjoyable experiences for a structural engineer. The holistic innovation that the engineering team was able to bring to this project, at the time pushing the bounds of research-led design, computational analysis and BIM, as well as collaborating with world-class experts and a progressive client, made for an exceptional experience. From the early design inception to the final product, Al Janoub stadium is a unique and stunning example of how a building form can be both functional and beautifully elegant.

The authors wish to acknowledge all members of the design team and the project and programme management teams of Al Janoub stadium. The authors are very grateful for the cooperation and support received from the Supreme Committee for Legacy and Delivery throughout the process of writing this paper; the effort and support of Eng. Hilal Al-Kuwari, Eng. Othman Zarzour, Eng. Tamim El-Abed and Dr Alexandra Kardara are particularly acknowledged.