AI finding a home in the water industry

With water scarcity an increasing problem, artificial intelligence could help mitigate water management problems and enable water utilities to better serve their customers.

Digital technology is having a significant impact on all sectors of society, not the least of which are the industrial and utilities sectors. Advances in digital communication technology, data science and artificial intelligence (AI) will change all aspects of these industries in the near future, and for the water and wastewater industries, these advances are enabling new approaches to the planning and management of water supply and wastewater systems.

Water utilities have been involved in digital networking and data analytics for years, and like many other industries already have instrumentation and telemetry systems that collect and transmit data and report continuously on system status.

However, the rapidly advancing world of the Internet of Things (IoT) and data analytics offers so much more, and is beginning to offer water utilities the opportunity to make real service improvements for their customers.

AI can be defined in many ways and perhaps means different things to different people. However, it is possibly best described as intelligent response computer systems. The main benefits of AI include not only the ability to identify patterns in data, but also to learn from these patterns as a result of endlessly refining relationships between variables.

Improving customer service

Speaking at Ozwater’19, THIESS Data Science and Analytics, Innovation and Technology Manager Virginia Wheway said the capability of data systems moves from reporting through to analysis, with analysis then enabling monitoring and, eventually, prediction.

“Reporting is about what has happened. If things are not happening efficiently, you can then move to analysis. But you can’t move onto analysis until you have reliable data,” she said. “Of course, the Holy Grail is prediction. And the step after that is prescription; that’s when the machines take over and start prescribing what to do in any event.”

Early experiments in expanding the use of digital technologies are already bearing fruit. For example, Sydney Water’s experiments with the IoT have already provided the benefit of the better customer relations that results from better information about outages.

Having better information about service disruptions that impact customers has resulted in financial saving from customer rebates paid due to disruptions. Notifying customers about water disruptions in advance and keeping them informed about progress on faults has resulted in a 40% reduction in inbound calls.

Also speaking at Ozwater’19, Sydney Water Customer Hub Manager Darren Cash said that the technology is allowing the organisation to improve its understanding of the types of problems customers are having.

“Generally, we have to wait for the disruption to happen before we are able to do anything about it,” he said. “But we want to get to the point where we can identify issues before they affect our customers. There are a few enablers that will allow us to do that, and we see IoT as one of those.

“We might still just be responding to something that has happened, but this experience is going to give us an indication of where we need to take predictive action to help our customers in the future.”

Everyone is talking about ‘smart water’

But better customer relations is only the beginning of the story. Xylem CEO Patrick Decker said that ‘smart water’ will be “the next disruption”.

Decker said there are three dominant pain points for the water industry:

1. Water loss across the distribution network, a major financial burden for utilities.
2. Stormwater overflow, with climate change seeing record weather events taking place that place greater stress on water and sewer networks.
3. Rising energy consumption and associated costs.

He believes the role of smart water technology is in being able to embed more intelligence in the actual equipment and hardware and to be able to overlay software and data analytics on top of that hardware.

“This could be data for an individual piece of equipment or an entire network, which is turned into actionable insights for our customers,” he said. “These insights can help them reduce their water losses on clean water distribution networks, help them more effectively manage stormwater overflow situations, or reduce their energy consumption on the wastewater side of the network.”¹

Smart meters: still an open question

One of the enabling technologies for better water system management with AI is the smart meter. However, smart meter data needs to be managed carefully to be sure the data is interpreted correctly. For example, differentiating a household that has developed a leak from one that has consistent water demand requires advanced pattern recognition. Just looking for meters that never show zero flow and assume it signifies a leak could potentially generate many false positives.

But in any case, at this time only a small percentage of water meters around the world are smart. Until water utilities can better distribute smart meters, there will be some delay in achieving the promise of AI and machine learning from smart meter data.

Smart meters not needed for everything

Smart meters aren’t the only tools for gathering and data.

An example of the successful use of digital technologies in minimising water loss is a water pipe failure prediction tool that was awarded a UTS Eureka Prize for Excellence in Data Science in 2018.

Developed by the CSIRO’s Data61 Smart Infrastructure Team, the tool utilises AI and machine learning. The research behind the project involved collaboration with a variety of utilities on failure predictions and related problems, such as water main failure prediction, sewer corrosion prediction and active leakage detection. Data was gathered from 27 national and international water utilities covering nearly 9 million pipes and 700,000 failure records.

The tool is already helping utilities in Australia and overseas realise the data they collect has practical applications, including pinpointing the most at-risk pipes in their network.

Melbourne Water has also had some success in utilising AI to optimise water pumping, in order to reduce energy consumption. A custom program mines operational data to ‘learn’ the most efficient pump configuration at any given time, and is expected to help the pumping station reduce energy costs by as much as 20%.

The program utilises historical data to determine the most energy-efficient combinations of pumps and the associated speeds necessary to achieve the right flow rate, and is able to take into account other data such as reservoir level, available pumps and past performance. It determines optimal pump calibrations and sends them directly to the pump system without any human intervention — an example of the AI determining the best settings and applying them in real time.

Water scarcity: AI to the rescue?

In 2018 UNESCO published a report stating that around five billion people are expected to be living in countries or regions tackling water shortages by 2050, and in Australia, our experiences with the drought and massive water shortages in our rural areas should be an incentive to make better use of technology to manage water usage.

AI has the potential to fundamentally transform the economics and productivity of water management in years to come. In turn, more intelligent water solutions will become available to areas suffering from water scarcity.

Conclusion

Across many industries, but not least in the water and wastewater sectors, AI will be a driving force for change across industry. In a future where water scarcity is an ever-growing problem, the application of AI solutions could help mitigate water management problems and enable water utilities to better serve not only their customers, but the environment as well.

Reference

1. Henderson C 2018, Xylem CEO Patrick Decker: Smart water is the next disruption, Xylem, <<https://www.xylem.com/en-us/making-waves/water-utilities-news/xylem-ceo-patrick-decker-smart-water-is-the-next-disruption/>>

Image: ©sock.adobe.com/au/Chinnapong