BPS Case Study: Sustainable Bank HQ

Building Performance Simulation in Early-Stage Design: Case Study 01 – Sustainable Bank HQ, Shanghai

The previous article in this series introduced some of the guiding principles our interdisciplinary engineering teams work to in order to deliver integrated building performance, from the very first sketches. Openness to innovation, programming sustainability into the DNA of a design (while sometimes even influencing building form and massing), and identifying a design’s further developable performance potential were identified as key factors that contribute to a project’s quality.

We also asked the question how, in a fast-moving early competition design context, building environmental performance can be demonstrated while a design is still highly flexible and needs to take into account the narrative demands of competition proceedings. This first case study, a continuation of the ‘Building Performance Simulation (BPS) in early stage design’ article, intends to shed some light on the interface of high design, environmental performance and interdisciplinary engineering consultancy.

Fig. 01 – Simple Irradiation Studies (image and simulations: Buro Happold; Architecture: Mecanoo)

Buro Happold worked with architects Mecanoo in an interdisciplinary competition team (MEP, structures, environmental design / building performance simulation, fire engineering) to support this single-phase invited competition. The client solicited proposals for its new HQ in Shanghai, which is intended to embody the newly founded institution’s identity of sustainability, social responsibility and resilient, mindful economic development. Despite our team not winning the overall competition, we reinforced our strong working relationship with Mecanoo which has lead to many successful collaborations.

Fig. 02 – Shanghai Climate Studies – Sky Radiation, Wind Direction/Speed, Temperature / Comfort Ranges (BuroHappold)

As Shanghai’s climate has both very warm, humid summers and comparatively cold winters (Köppen climate class Cfa – humid subtropical), the design had to work hard to mitigate undesirable solar gains through the facade, create comfortable work place and exterior roof spaces, and still be highly spatially efficient.

The client brief requested a total glazed ratio of no more than 50% , which added additional challenges to satisfy daylight utilisation in the offices, while allowing a sufficient glazed percentage to visually open up the public areas, e.g. around the street front and roofscape cafes. In general, the client’s aspirations were ambitious, aiming to use the exterior natural environment as much as possible to achieve excellent internal building comfort conditions while relying on passive design principles, and achieving the highest possible GB/T 50378 (“Three Star”) and LEED ratings, with according impact regarding the energy use targets.

Fig. 03 – Site Plan and Axonometric (Mecanoo)

The very early stages of the design process especially considered the urban situation of the former EXPO 2010 site, one of whose plots will be developed for this project. Opening views to Shanghai (approximately to north) and keeping the long office tower facade axis oriented south as much as possible were explored; simple irradiation studies (fig. 01) identified the most solar-exposed facades and contributed to avoiding long east or west-facing elevations. Ideally, this very early-stage involvement of specialised engineering disciplines (especially environmental design) has a significantly positive impact, and is pursued by Buro Happold whenever possible.

Another special feature of the design was the introduction of a slanted, diagonal cut to the south-west corner, reducing the floor plate depth towards the lower floors and leading to significant self-shading of the envelope, especially in summer due to high prevalent sun angles. In general, the floor plate depth was kept to within one 8.4m grid in the tower, as deeper floor plates (in combination with reduced facade glazing areas) would lead to sub-optimal daylight utilisation (and an 8.4m grid is already pushing daylight autonomy utilisation to the limit). Similar depth-limiting daylight zoning strategies were pursued across the entire design, with sensitive spaces positioned and shaped for maximum daylight access, and by placing deeper convention and cafe spaces in the larger plinth (some equipped with skylights for deep plan natural lighting).

Fig. 04 – Facade Performance Design Sketches (Buro Happold mark-up on Mecanoo views) & South-East Perspective (Mecanoo)

Following the massing definition, many facade design typology variants were explored by the integrated team, focusing on means of self-shading through local depth increases (e.g. overhangs), geometric folding and differentiation of the facade layout per orientation (fig. 04). The limit of the total glazed area posed a significant challenge, necessitating that some areas of the plinth (as seen in fig. 05) without overly onerous daylight targets experience a reduction in transparency. To ease the introduction of fine gradients, a main visual facade grid of 1.2m was chosen, though the actual curtain wall panel dimensions were envisioned to be much larger, in order to minimise thermal bridging at the interfaces.

Fig. 05 – Facade Folding Strategy and Perspective View of the Plinth (Mecanoo)

Mecanoo and Buro Happold finally adopted a strategy of using adaptive, non-equilateral triangular facade folds sitting under horizontal static shading bands to structure the facade environmental response. In the final design, the glass gradient gradually opens up from the more enclosed plinth towards the office zones, with subtle shifts in glazing orientation. The shifts are made possible by either flipping the triangle orientation in plan or assigning glass to either the longer or shorter triangle side.

Once the building was locked in plan and the facade principles set, which shift or flip is appropriate for each elevation subsequently became the main question to be answered through building simulation. Both aesthetic and multi-domain performance considerations (daylight utilisation and visual comfort, primary energy use reduction) thus met at the facade interface.

In order to guide Mecanoo’s design intent, Buro Happold built a simplified parametric model of a representative office floor plate (fig. 06) and tested out a range of different glazing ratio and facade fold setups. The facade was parametricised in Grasshopper and via Honeybee linked to an auto-zoned (core and perimeter) EnergyPlus / Daysim model. Additional glare studies were performed with EvalGlare via DIVA4Rhino, and daylight utilisation outputs visualised with the Mr.Comfy analysis plugin.

It is important to note that the facade test variants were not driven by a genetic algorithm or other automated design space explorations, but manually pre-defined through conversation and workshop tests with Mecanoo, as it was clear from the beginning that as long as targets can be met, certain design variants – from an urban and architectural design point of view – are very much preferred. Visual openness and thus underlining the lightness of the vertical massing took centre stage and had to be achieved as part of the competition design narrative.

Fig. 06 – Iterative Multi-Domain Building Performance Simulation Outputs (BuroHappold)

Eight design variants were tested with the parametric model, starting with a 50% uniformly glazed, non-folded facade baseline variant (A), via 50% (and folded) designs, and finally up to flipped glazing orientation designs with customised glazed ratios along different orientations (Cx series). Both the relative primary energy demand as compared to the baseline (A), annual glare and daylight utilisation (at 500 lux) were used as benchmarks.

Metrics revealed that folding the facade and customising glazing orientations per elevation has tangible benefits, both in primary energy use and in glare reduction for occupants close to the windows. The best performing variant (C1) featured a reduced glazed ratio of 35% on the south-west facade (where the glare readings were also taken); discomfort glare hours (without external or internal shading) were reduced to approximately 70 (down from 240 for the uniform variant A) and primary energy use reduced by approximately 8% (compared to baseline).

As the energy simulations were fully daylight-linked (occupancy schedule-driven and light level sensing), the increase in building energy performance was heavily related to variant C1’s increase in north-east facing glazing and flipping the glazed facade pleats to north, limiting direct solar gains while allowing more diffuse light to enter and deepening the daylit extent. Interestingly, the overall daylight autonomy result (41% daylit area) was slightly less than in related iterations, reaffirming that strategically shaping daylight penetration along key orientations is just as important as ensuring generally optimised levels across a floor plate.

Finally, cooling energy use was also reduced by the optimised orientation and limited area of the south-west glass, though high performance low-e glazing (g-value of 0.27) already limited cooling use and internal loads stayed invariant across the comparison matrix. To consider additional external, adaptive shading would have reduced cooling further, but was not explored during these first optimisation steps and was intended to be investigated in the next design stages. As described in the following, pursuing these strategies would have been even more important later on, as for architectural reasons, not the best-performing variant (from an energy point of view), but a compromise design was chosen.

Fig. 07 – Detail View of South-East Facade (Mecanoo)

The simulations revealed that the features of Mecanoo’s design positively impact on overall building environmental performance, with further optimisation potential embedded in the system design principles.

Returning to the parent article’s question of ‘how do we “prove” a design’s performance while (almost) everything is still in flux?’, we see in that in this instance, environmental concepts are indeed clearly legible in the architecture, and we have developed a first understanding of the interplay of design-engineering sensitivities for this specific spatial challenge. It is also the features of the architecture itself that initially define performance variables, retaining an ability and robustness to adapt during the following stages should they have been perused.

However, despite the design’s performance potential, in architectural competitions it is atypical for the ‘best-performing’ variant (as resultant from the somewhat monocentric point of view of building performance simulation) to be implemented at face value, as only the interplay of urban and architectural quality – of course in their synergy with engineering – makes for a truly exciting project.

No less was true in this instance; observing carefully, one notices that e.g. in the final renderings, the south-west glazing orientation is not flipped to north, as in the ‘ideal’ simulation iteration, but faces roughly south instead; this was an architectural decision finally made by the team, as to demonstrate openness to the city from this urban vantage point (and opening views into this direction from within the building) was deemed more important than merely claiming performance benefits in isolation.

It thus would have been the plan to solve the resultant visual comfort and energy drawbacks through alternative means during the next stage, e.g. by offsetting some of the increased cooling demand through improved external shading and by harnessing renewable building energy sources, which were already proposed in the concept (especially geothermal cooling). While otherwise outside of the scope of this article, outline MEP concepts and an energy stream map are shown in fig. 08; the morphological features of the design were intended to work in unison with the renewables strategy, while minimising demand sufficiently to allow for low-energy systems to be used throughout.

Fig. 08 – Energy Stream and Outline HVAC Schematic (BuroHappold); North-West and Atrium Views (Mecanoo)

In summary, the design worked across domains to build adaptable performance robustness into the key early-stage design features. The integrated team especially considered facade, massing and underlying MEP concepts to give Mecanoo architectural flexibility, while guiding design intent with heuristic advice, as well as cross-domain simulations that would have paved the way for future optimisations. Apart from the main decision on glazing orientation, many more granular ‘micro-decisions’ were taken that throughout the project life cycle would have demanded similar compromises to be made; however, the most important thing regarding early-stage building performance decisions is to keep the bigger picture in mind, which is one of the great strengths of working in a multi-disciplinary environment.

Finally, many of the principles we worked with in this competition later made their way into other projects, which is another positive side-effect of ‘testing it out’: finding genuinely adaptable systems is no small task, and performance intent is often refined across multiple submissions by the same integrated team. One of the next articles in this series will explore related strategies envisioned by a completely different team of architects.