Sensing trouble in the sewer system

Some 20 % of the UK’s ageing sewage structure is damaged, leading to 5,000 collapses and 200,000 blockages each year. Many sewers are too small for inspectors to crawl through, while video cameras mounted on wheeled trolleys do not always pick up on cracks, nor can they see through water. Checking video footage is also time-consuming – a one-kilometre stretch of sewer generates about three hours of footage.

A team of scientists at King’s College London have now developed an ultrasound sensor that can scan areas of pipe below water for cracks or growing roots. A separate laser system projects a circular pattern on the pipe walls, identifying defects where changes occur in the pattern via a video camera.

“Once we have collected the data and stored it digitally, we then subject it to analysis with intelligent computer software based on neural networks,” says Dr Kaspar Althoefer of King’s College. The group’s computer is ‘trained’ to recognise the characteristics of a defect from the sonar or camera data, allowing it to detect any defects automatically and pinpoint their position in the pipe.

As well as providing autonomous damage detection, the new system is more easily able to distinguish between genuine defects and apparent defects, which are actually reflections of light from the wall of the pipe.