After recent reports regarding the background to the departure of the CEO of Thames Water have emerged, it appears from the press coverage that one of the contributing factors impacting Thames Water’s performance is its struggle to address the significant leaks in its network. The BBC reported that Ofwat had fined Thames Water £120 million, to be paid back to their customers as compensation. According to Thames Water’s website their 2017-18 leakage from their network totalled 695 million litres a day. That is the equivalent of 278 Olympic swimming pools every day of the year. The Financial Times says that the leaks are equivalent to 180 litres of water every day for each property served by Thames Water.
Earth-i have been working on technology to help locate potential leaks using satellite imagery. This process uses near infrared satellite imagery to create normalized difference vegetation index (NDVI) to highlight areas of high moisture vegetation near water pipelines. These areas of vegetation show up particularly well after a period of dry weather. Training machine learning (ML) models on the NDVI signature of water leaks allows the use of Artificial Intelligence (AI) to automatically scan the images, speeding up the detection dramatically when compared to traditional leak detection methods. Using this approach, water companies could target their leak detection teams to areas of high likelihood of finding leaks.
As summer approaches, Earth-i is again capturing satellite imagery to analyse for leak detection, not just in the UK but globally. We are keen to work with both water utility companies as well as regulators to bring our technology and expertise to help reduce wastage of such a precious natural resource.
As part of Severn Trent‘s ongoing strategy to use new technologies to help identify and fix water leaks, Earth-i completed a successful trial in the summer of 2018. Click here to read the press release.
If you would like more information about how Earth-i is helping utility organisations, please email firstname.lastname@example.org.