Case Studies

Modernising Sewer Pipe Maintenance with Robotic Sensors


In 2022, South East Water used CCTV to conduct a sewer pipe condition assessment which indicated that the pipe required maintenance. Due to the limitations of CCTV inspections, the extent of the works required weren’t evident and the pipe collapsed before repairs began, incurring significant costs.

Across the country, a better solution is needed to accurately assess the condition of old sewer pipes.

Challenge

Water authorities own and operate sewerage networks consisting of thousands of kilometres of pipes under our cities and are responsible for their maintenance and repair.

But current methods of maintaining sewers are not very effective, leading to pipe failures such as leaks and collapses.

LaTrobe University lecturer and project research lead Alex Stumpf explained the problem with the current maintenance method.

“A robot goes down and takes a video image of the pipes. There’s a human operator that sits there watching this video for hours on end and then they classify the pipes on a scale from one to five,” Stumpf said.

“It’s only as good as what the human eye can perceive, so that tends to lead to errors. And if they miss something, then you can get some pretty disastrous results.”

Often, pipes can’t be repaired but must be replaced entirely which can be very expensive. Add to this the cost of clean ups of damages and spills and even potential fines. Annually, this costs water authorities $982 million each year, the equivalent of $60 per adult Australian (ACA).

Pipe leaks and collapses can also pose risks to human health and damage the environment, for example, if sewage spills enter waterways.

Objective

South East Water, in partnership with La Trobe University, are developing a system which includes a robot to go into the pipes and collect data on their conditions.

“The challenge we’re trying to solve is to create more information to help inform water authorities on how their sewer infrastructure is performing,” Stumpf said.

SmartCrete CRC has invested $218,000 into this project to advance an already-in-the-making robotic-based sewer pipe condition assessment system.

The system will assess the pipe’s condition by collecting data and using AI to turn it into actionable insights into how much longer the pipe will work for and whether it needs repairs.

The project will field test the robot to make sure the sensors are collecting accurate data.

“We always need to clarify with industry, Is that a suitable method? Is what we’re proposing going to actually work?” Stumpf said.

“Operationalising the robot really needs to be informed by industry. So having this opportunity to be working with industry and funded by SmartCrete CRC is a really good way of doing research and developing things and trying to translate research and impact out of the university.”

Impact

Once operational, the condition assessment system will give South East Water and other water authorities a way of maintaining their pipes and keeping them in good condition for longer.

It will provide a more accurate way of working out when repairs need to be done, which is anticipated to reduce the number of pipe failures. There will be less risks to the environment and human health, and fewer costly clean ups.

It will also result in significant savings for water authorities each year, which can potentially be transferred to customers – the homeowners and businesses who pay reduced bills.

Next steps

The project lays the groundwork for future research to test the condition assessment system across the city, and  collect data to train the AI. Once the AI system can accurately assess pipe conditions, it will be ready for industry adoption.

The system would also help refine the existing code for assessing pipe conditions, making it more precise and better at preventing pipe failures.

Project title

    Robotic Based Sewer Pipe Condition Assessment

 

3 years
Duration

 

$1,288,138
Value (Cash + In-Kind)

 

In progress
Status

Partners

  • South East Water
  • La Trobe University