The drones diagnosing heat loss in buildings

According to the government’s Heat and Buildings Strategy, 25% of the UK’s greenhouse gas emissions come from heating buildings. One way to cut emissions and energy bills is to retrofit buildings – with the added benefit of reducing fuel poverty and health issues from cold homes.

Retrofitting aims to boost energy efficiency either by improving the ‘fabric’ of a building – think insulation or triple glazing; or by improving its energy systems, for example by installing a heat pump. To complicate things, the two are interlinked: heat pumps become more expensive to run if buildings are not well insulated. 

The usual starting point for retrofitting is to bring in a specialist energy surveyor to test where heat is leaking. The problem is, there aren’t nearly enough of these specialists for the amount of homes that need surveying. 

The scale of this challenge is “really staggering”, says Lucy Lyons, co-founder of Kestrix, a startup aiming to become the “Google Maps of heat loss”. Of the 29 million homes in UK, “almost all” would benefit from some form of retrofitting, Lyons explains. Kestrix aims to accelerate energy surveying, and thus retrofitting, by flying drones equipped with infrared cameras over buildings and analysing the images with AI. The company was recently named as a finalist for the Manchester Prize, funded by the UK Department for Science, Innovation, and Technology.

Diagnosing heat loss

Kestrix’s story started when Lucy’s co-founder Matt Goodridge was renovating his home. He wanted to make his home more energy efficient but couldn’t for the life of him figure out where to start, despite growing up with parents who renovated their homes. At the same time, he’d seen AI being applied to analyse medical images while working in the medical devices industry. He thought, if analysing images of people with AI allows us to diagnose diseases, why can’t we diagnose heat loss by analysing images of buildings?

After the two co-founders met at a climate tech venture builder Carbon13, based in Cambridge, they decided to launch Kestrix to solve this problem by applying AI to aerial images, at the moment captured by drones. Drones are a well-established technology in the building sector, for example, to inspect wind turbines and roofs. It wasn’t too much of a stretch to imagine they could be used for energy surveying too. 

Kestrix’s approach is to fly off-the-shelf drones fitted with thermal cameras about 30 to 60 metres above buildings. The thermal images are stitched together into a 3D model, which can be toggled on or off over a regular photographic view, to identify anomalies: anywhere heat is leaking will light up red.

While it might sound simple, this belies the complex building physics and thermography underneath, says Lyons. If you think of a terraced house, compared with a new build block from 2021, a bungalow or a 1970s tower block, the way heat is lost from each type of building can differ substantially. 

So, Kestrix’s algorithm estimates a U-value, a measure energy surveyors use to calculate how heat travels through materials. Normally, measuring U-values requires attaching heat flux plates to the inside surface of walls, roofs and floors, and recording how much temperatures change over time. “Simply put, leaving a bunch of hardware inside a house and taking measurements,” explains Lyons. (There are other options too, such as thermally imaging the building combined with an internal blower-door test or conducting an Energy Performance Certificate survey, albeit outputs will be mostly assumptions-based rather than based on real performance. But all require a trained, qualified person – which is where the bottleneck lies.)

However, estimating the U-value without internal readings is not a trivial task, says Lyons, and the company is working to validate its accuracy with academic partners, initially for certain housing types. Getting it right is “not a challenge we’re taking lightly,” she adds.

Verifying and accelerating retrofit projects

Once the test is complete, the client can then make an informed decision. Should they get triple glazing? Are they ready for a heat pump? Have their past retrofitting efforts been successful? 

Kestrix is currently testing out its technology in partnership with social housing providers, local authorities, and utilities. The idea is to help them plan, price and verify retrofit projects, ultimately helping people save on their energy bills and reduce carbon emissions. 

For example, the company is working with social housing provider Peabody Trust to see how well past retrofitting projects have worked, which can help shine a light onto the carbon intensity of operating homes before and after. “Today, there’s not really a scalable, reliable way to assess how a building performed before retrofit and then after,” says Lyons. “You can look at someone’s bills, but there’s no way to say [a decrease is] because the cavity insulation is performing really well.”

Kestrix is also working on a heat pump siting tool in collaboration with EDF, funded by the Department for Energy Security and Net Zero, and also rolling out a retrofitting roadmap tool for residents with Islington Council. To create the latter, the company has scanned a neighbourhood in Islington. With the data, local people will be able to access free advice on how to make their homes more energy efficient.

Next autumn and winter, Kestrix plans to scan tens of thousands of properties. Longer term, the company has set its sights on helping the government back up decisions on allocating funding in the social housing sector. With a billion pounds’ worth of funding to be distributed over the next few years, Kestrix hopes to verify the success of completed retrofits, and show where funding should be prioritised in future.

The company’s involvement will “hopefully address the not just the 25% of emissions coming from heating buildings in this country, but the plight of fuel poverty and the burgeoning health crisis that cold homes are causing,” says Lyons. “We really hope to be a part of the transition.”