In Victoria, Intelligent Water Networks (IWN) has a Data and Analytics Program focused on trialling new and emerging technologies to help the water industry optimise data management. Here, Utility explores its recent projects and findings.
For the Victorian water industry, intelligent data can deliver cost efficiencies and improve customer service. It will also play an increasingly important role in meeting the challenges of operational complexity due to aging infrastructure, climate change and population growth.
This goes to the core of why Intelligent Water Networks – known as IWN – exists. IWN trials new technologies and innovations for its Victorian water corporation members and provides development opportunities for their emerging leaders.
Through its dedicated Data and Analytics Program, IWN looks for tools and opportunities to improve data management and analysis in priority areas identified by its members.
As a safe sandpit to trial new ways of doing things, IWN’s work helps member corporations understand the technologies available to address industry challenges. In this article, we take a look at two projects from the Data and Analytics Program over the last 12 months.
Automating the review of sewer cctv footage
There was a time when CCTV technology marked a new era in managing sewer networks. However, with footage in Victoria’s water industry still manually reviewed to detect faults and classify sewer defects and condition, it’s time for another technological stride forward.
IWN Data and Analytics Program Manager, Kevin Hellier, said the program began looking into this after its industry working group identified the manual review of sewer CCTV footage as a priority area for improvement.
“Water corporations often have a backlog of sewer footage to look through,” Mr Hellier said. “We wanted to look into technology’s ability to automate this and how accurately and efficiently it could do it.”
In late 2020 and early 2021, IWN worked with ten of its member water corporations to trial the automated review of CCTV footage using artificial intelligence (AI) and visual recognition technology provided by Blackbook.ai/AWS, VAPAR and INLOC Robotics.
Each vendor trial progressed in parallel, with different numbers of member water corporations. Each corporation inputted at least 2km of gravity sewer CCTV footage into their vendor platform. The trial tested the technologies’ functionality, including the effort involved in using it and a baseline accuracy assessment.
IWN’s trial report is available to its members, including indicative set up costs. Project Coordinator, Sophie Bangs, said the trial found the technologies to be very promising. “They have the potential to deliver significant time and cost efficiencies,” Ms Bangs said.
“All vendors demonstrated a promising similarity in results, however more work is needed to determine the accuracy of the AI technologies against consistent structural condition benchmarks.”
There are also further opportunities to explore, including how to refine the technology to improve the consistent defect classification and whether it can be adapted to provide intelligence to the camera operator to augment footage capture. IWN will assess interest from its water corporation members before exploring these opportunities.
Cleaning big data for water corporations
Another improvement opportunity for the Victorian water industry is the accuracy and interpretation of big data for asset management.
This year IWN held two trials of an emerging data cleansing software being developed in Australia, called SensorClean, which aims to make it easy to clean and visualise operational time-series data for decision making.
Each trial involved two Victorian water corporations and focused on:
1. Cleansing data to assess the value in water supply peak flow and non-revenue water analyses
2. Cleansing data to assess the value in sewerage pump station energy management Mr Hellier said the software enabled time-series SCADA data to be uploaded, water quality anomalies to be classified, and datasets to be cleaned for visualisation and analysis.
“The trials demonstrated the value of improving the quality of time-series data from asset monitoring for various business outcomes, such as water system capacity planning, non-revenue water analysis and managing sewerage pump performance,” Mr Hellier said.
“The main benefits identified were time saved in issue investigation, elimination of nuisance data, and improved data sets for asset planning and reporting.”
Since the trials, Australian business FSA Data has released SensorClean to the market. IWN is yet to determine interest from its members for further trials to assess SensorClean’s ability to provide other benefits, such as identifying the impact of wet weather events for power usage comparisons.
