Sydney Water combines a world-first robotic solution with predictive analytics to improve the performance of wastewater infrastructure and reduce annual maintenance costs.
Maintaining Sydney’s 100 year-old underground sewer system is no easy task. The ageing infrastructure requires continuous monitoring to minimise the risk of pipe failure and protect the health of the community and environment.
Sydney Water spend $40 million annually managing 26,350km of sewer systems across Sydney, the Illawarra and Blue Mountains. In order to inspect the pipes, the team must enter live sewers and walk up to 6km to gather core concrete samples, or use CCTV cameras to access restricted spaces. This type of manual inspection is high-risk, slow to deliver lab results and dependent on human judgement.
iPipes Principal Investigator, Professor. Sarath Kodagoda (right) & Mechatronics Engineer, Vinoth Kumar Viswanathan (left), at UTS Tech Lab
iPipes at UTS Tech Lab and Sydney Water have partnered to rethink pipe maintenance and develop the world’s first robotic sewer assessment toolkit. A buoyant robotic system, weighing 40kgs and spanning 1.2m in length, enters sewers through a 60cm diameter manhole then expands up to 1.5m inside pipes, allowing for safe inspection in confined spaces.
Equipped with sensors such as ground-penetrating radar, pulsed Eddy current sensor, 3D mapping and CCTV, it provides a complete assessment of the pipe structure, including visual and sub-structure information. The robot is operated via a remote-control station, where Sydney Water can identify defects in real-time.
The toolkit also contains a handheld device, the same size as a drill, which is used to manually measure the thickness of pipe walls and detect corrosion levels and other problems. These solutions deliver crucial timely data that improves decision making capabilities, reduces OHS risk and minimises environmental, social and economic impacts. The toolkit is estimated to have decreased maintenance and repair costs by up to 10%.
From 2021, the robotic solution will be enhanced by a failure prediction model developed by the UTS Data Science Institute. Using advanced machine learning techniques and multiple sources of historic pipe assessment data, the predictive model pinpoints locations at greatest risk of failure so that robotic systems can be deployed accurately and efficiently.
This multidisciplinary robotic capability represents a major technological advancement for utility companies managing underground or hard-to-reach assets. The smart tool is non-destructive, reduces maintenance costs and out-of-service times, improves accuracy, negates the need to send humans into dangerous situations and reduces the likelihood of catastrophic pipe failure.
The fully functional prototype will be deployed and refined over the coming years while a commercialisation pathway is developed. The team is aiming for broad service availability by mid-2024.