• Data-Driven Decision Making in the Water Sector
    Data-Driven Decision Making in the Water Sector

    Australia experienced perhaps its most severe period of drought in hundreds of years from 2017 to early 2020, resulting in devastating bushfires and major water supply issues. Despite coming out of this drought, many water authorities continue to face long-term challenges of water security in the context of:

    • Climate change, resulting in more frequent and severe periods of drought, extreme rainfall events impacting stormwater and wastewater treatment, and higher temperatures leading to greater loss of water through evaporation;
    • Ageing assets, with much of Australia窶冱 water infrastructure built before 1970, and many assets now reaching the end of their 50-year lifecycle;
    • Growing populations, with major urban growth areas including Western Sydney, Melbourne urban areas, Perth and Peel regions in WA, and South-East Queensland;
    • Expenditure constraints, with water authorities under significant capital constraints and budget pressure over their 5-year planning period, and consumers unlikely to bear significant price increases;
    • Changing customer expectations, including greater digital control in usage monitoring, billing, payments, services enquiries and fault management; and
    • Growing sustainability, environmental and governance expectations.

    One major factor that water authorities, engineers, contractors and product suppliers are looking to leverage is the rapid advancement in digital technology. In this article we outline three case examples of water authorities using sensor technology and data science to achieve tangible operational and commercial impact.

    Case Example 1

    A regional council was forecasting its population to more than double over the next 10 years, which in turn meant doubling water infrastructure, including a new water treatment plant. The council had started charging customers an additional fee to raise funds for its construction. The council was also aware from biannual meter readings that there was substantial opportunity to reduce water loss, but did not have good data on how to address this.

    Following a successful trial of smart sensors on a number of large water consumers, the council invested in a full rollout across its network. Simply being able to defer the capital for the new water treatment plant more than justified its investment in smart metering. As a result of the smart metering and analysis, the council achieved a significant reduction in water consumption through:

    • identifying network water leaks to develop a targeted maintenance program to reduce leaks; and
    • identifying residential property leaks and proactively contacting customers.

    The outcome was that the council determined that they would no longer need to build a new water treatment plant for the foreseeable future, and was able to reduce the average water bill by $400 per year.

    Case Example 2

    Another regional council had received numerous complaints from customers about not having sufficient water pressure during peak hours in the morning. The council was planning to install additional pressure pumps to boost town pressure, but was concerned by not only the capital cost and additional operational costs, but also the increase in water loss that would occur as a result of the increased pressure. The council had installed smart meters to help it achieve its agreed pricing target of growth by CPI minus 1% (i.e. a 1% cost reduction in real terms), but had not yet analysed the data to draw actionable insights.

    The council leveraged external data science capability to model and build an animation of water consumption by location by hour over the period of a month. This highlighted a number of dairy farms that were major water consumers between 7am-8am. The farms had large tanks of water that were used to flush the facility after the cows had been milked, which would then refill during that period.

    The council liaised with the dairy farms to put a valve on the tanks so that they would only refill if there was sufficient pressure, and at a maximum flow rate. This immediately reduced demand during peak hours, while the tanks would still be full before the next milking. As a result, the council no longer needed to make the significant capital investment in additional pumping facilities, and could avoid the increase in operating costs and non-revenue water that would have ensued.

    Case Example 3

    A major city water authority was having an ongoing issue with vacuum sewer events, causing significant unplanned and labour-intensive maintenance. A fault would occur and contractors would need to identify where the issue was by manually checking suspected pumps. The city council analysed the commercial payback of improving maintenance practices and proceeded to install two sensors on each vacuum pump 窶 one to alert if the pump had been open for more than 30 seconds, and another to alert if the water level had risen too high.

    This immediately provided a live view of the entire vacuum sewer network, enabling maintenance teams to immediately locate and resolve issues. Additionally, three months of this data was fed into machine learning algorithms, which was then able to predict with 78% accuracy when there would be a vacuum sewer event in the next two days.

    As a result, the data and insight from machine learning enabled the council to fundamentally change its maintenance approach. It moved from reactive, highly manual maintenance, often out of business hours and on weekends, to proactive, targeted maintenance during normal business hours at a substantially lower cost. This resulted in a major reduction in network failures.

    These examples illustrate simple yet powerful examples of the potential for water authorities and contractors to leverage low-cost sensor technology and data science to increase asset life, reduce non-revenue water, and reduce capital costs and operational costs.


    If you would like to discuss the challenges and opportunities you are facing in the water sector or other asset intensive industries, and how Mainsheet has assisted management teams to develop their strategies and improve performance, please contact:
    Dan Sunderland | dan.sunderland@mainsheet.com.au
    Gerard Moody | gerard.moody@mainsheet.com.au