Smarter tools supporting expert decisions in water management

WaterNSW is exploring how artificial intelligence (AI) can give our experienced operators faster access to important information, helping support safe and informed decision-making when conditions are changing quickly.

Managing dams during periods of high inflows and flooding requires careful monitoring, technical expertise and sound judgement. AI has the potential to complement that expertise by providing data sooner and by helping operators interpret data more quickly and consistently.

A new AI model, developed and delivered by a team of WaterNSW employees and presented at Ozwater'26 by Shahriar Sadatieh, demonstrates how emerging technology can be used alongside existing operational systems to support decisions about dam releases. Importantly, this capability is designed as decision support, not autonomous control. Human operators retain full authority over all operational decisions, with the AI providing high-precision, real-time insights to inform judgement rather than replace it.

Understanding how much water is being released from a dam is an important part of dam operations, particularly when inflows and downstream conditions can be changing rapidly. Traditionally, calculating release volumes has relied on manuals, reference tables and operator assessments.

WaterNSW’s development of a neural network model has made the process faster, more accurate and more consistent.

Using information such as reservoir levels and gate positions, an AI model was trained to estimate water releases across a wide range of operating conditions.  Testing showed the model could provide accurate estimates in real time, giving operators timely information to support their assessment of changing conditions.

A successful feature of the project was integrating the AI model directly into WaterNSW's operational control systems. This allowed calculations to occur closer to where data is collected, improving efficiency and supporting faster decision making. The project also uncovered additional benefits by helping identify unusual operating conditions, providing operators with another way of recognising when something unexpected may require closer attention.

While still only a decision support tool, the project highlights the potential for AI to assist with a range of operational activities, including maintenance planning, performance monitoring and future improvements to automated systems. The project has also highlighted opportunities where AI could be applied to similar technology across other WaterNSW assets, including weirs, regulators and storages.

As WaterNSW continues to modernise its operations, projects like this demonstrate how new technologies can complement the expertise of our people, helping us manage water infrastructure safely, efficiently and reliably for the communities that depend on it.