Managing critical water assets
Cost-efficient water-cycle services
Awareness & understanding
Increase food production
Welfare & healthy growth
Ensuring safe water
Access to timely water data
Reliable observation and forecasting
Sustainable water infrastructures
Efficient and competitive
Support, prevent and improve
Innovative, fast and low risk
Cost effective and scalable
Using machine learning to reduce asset management costs without compromising service levels.
G-MW had undertaken a very large asset renewal programme automating water delivery in order to provide a high level of service equally to all customers. Some customers became concerned of the need for such a high level of service if it came with higher maintenance costs. G-MW began to examine how to manage their asset based so that could be both affordable and appropriate for customers aligned with their commercial willingness to pay, and in accordance with their stated priorities regarding the key features of the service provided.
ADASA used machine learning to study the connection between asset maintenance costs and the resulting level of customer service. Data on water deliveries, canal supply levels, assets, weather, electro-mechanical alarms and pay-rates was collected, and a model was built of the flow rate accuracy and supply level consistency. Using these models some variables such as routine maintenance frequency or the point at which batteries should be replaced were varied to determine the likely impact on service level. Expensive maintenance activities such as desilting or deweeding channels was also examined to see if service was improved as much as was generally believed.
As a consequence, ADASA was able to identify a number of factors that impacted service level, as well as changes to maintenance practices that could result in annual savings of at least $1.25M and a reduction in the asset base of about $4M. ADASA recommended changes to the Problem-Cause-Remedy hierarchy used by the maintenance staff to improve the predictive capability of the models. ADASA also wrote a data improvement plan together and defined metrics to help G-MW on its journey to improving data quality and to improve the models accuracy and robustness.
Today, G-MW have a deeper insight into the links between costs and service levels and a workbench of tools, computer programming and models from which ongoing and deeper investigation is being conducted.
Located 200km north of Melbourne, Goulburn-Murray Water (G-MW) is the largest rural water authority in Australia and one of the largest in the world. G-MW manages water storage, delivery and drainage systems across 68,000km2 (26,000mi2) of catchment and delivers water to 39,000 customers including farmers, town water corporations, hydroelectric companies and industry. They manage 6300km of irrigation channels, rivers and aquifers and 24 dams.
Operational and Planning System
Murrumbidgee Irrigation Ltd.
Enhancing asset management and customer service levels through business intelligence.
Goulburn-Murray Water Corporation
Increasing the quality and accessibility of water resource information.
Reducing energy and improving effluent in wastewater treatment.