Over a billion people worldwide currently live in slums or informal settlements, where indoor temperatures can climb as high as forty-nine degrees Celsius. For them, good energy planning is no longer a matter of comfort, but survival.

But these same communities also often lack the means to plan effectively for climate adaptation and decarbonisation. To help under-resourced cities ground their energy policies in better data and speed their transition towards net-zero, Assistant Professor Ang Yu Qian from NUS Built Environment and his colleague published a study in Applied Energy detailing a new Hypercube framework – a faster, less computationally demanding modelling tool designed to put city-scale energy planning within reach of those with the most to lose from a warming climate.

Why urban energy mapping is important

For planners aiming to understand which policies will move the needle most, it is critical to know where a city’s energy goes. Through physics-based simulations, urban building energy modelling tools build a ‘digital twin’ of the city, providing key insights into energy networks.

This virtual stand-in allows planners to test different ‘what if’ scenarios: What if every rooftop was coated in heat-reflective paint? What if shading devices were installed throughout a district? “We can digitally test tens, hundreds of scenarios to see which one will give us better bang for the buck, which one will bring us closer to our targets – before we actually implement them,” said Asst Prof Ang.

When good modelling is a luxury

But conventional modelling methods that faithfully capture the full complexity of individual buildings are computationally intensive and prohibitively slow at the city scale, even when supported by high-performance supercomputers, according to the study’s first author, Dr Wang Tao. While faster, simpler approaches do exist, these methods sacrifice accuracy.

Most of these tools, moreover, require high technical expertise and have steep learning curves. Larger, more privileged cities can hire or outsource energy modelling to researchers and technical consultants, explained Asst Prof Ang, but smaller, less-resourced cities and municipalities often lack both the budgets and the technical knowhow. Many smaller cities, who have “maybe one or two planners”, resort to simple, abstracted calculations, he said. Thus, to democratise the use of energy modelling tools, the researchers set out to improve their accessibility through a new framework.

Finding the common denominator

Many conventional models pursue high modelling accuracy by analysing each individual building in detail; the more buildings a city has, the more calculations pile up, and the more computing power and time the job devours.

To bring that burden down, the research team used a simple workaround – one which Asst Prof Ang finds can be best explained with the popular children’s toy, Lego bricks.

The framework begins by breaking buildings into smaller units and grouping them based on their similarities across three parameters: their physical dimensions, how exposed they are to the sun and wind, and how occupants use the building.

“We run the simulations and analyses on these Lego blocks and then we scale them up to the whole city,” he said. This approach drives down the number of calculations needed, collapsing the computational burden. “So we call these Lego blocks, hypercubes.”

Working with hypercubes sacrifices some granularity, but at the city scale, the differences between individual buildings even out, said Asst Prof Ang. Benchmarked against ClimateStudio, an established per-building simulation method, the researchers found that Hypercube completed its analysis nearly 28 times faster, with comparable accuracy.

Further, the framework maintained that accuracy in tests across two starkly different urban fabrics – Paris, with its dense historic streets and cool, damp oceanic climate, and San Francisco, with its modern grid and mild Mediterranean weather – spanning some 3,000 and 2,000 buildings respectively. This proves the framework is “scalable across different cities,” said Asst Prof Ang.

To put the Hypercube framework in the hands of those who need it most, the researchers have placed the code in open access and are now working towards developing a user-friendly interface – possibly a phone app – so that planners do not need to work with raw code.

Paving the way to net-zero

Cities the world over have committed to net-zero targets over the coming decades. The problem, however, pointed out Asst Prof Ang, is that “the targets are usually quite clear, but the pathways to reach the target [are] usually not so clear.” With Hypercube, he hopes to close that gap, by making it easier for more policymakers, especially those governing developing and resource-poor cities or municipalities, to employ scientifically backed energy modelling and design targeted energy policies.