PUB introduced Aqua Suite to provide advanced process control and analytics

The challenge

Digitising Singapore’s water system to achieve improved operational performance

PUB, Singapore’s National Water Agency is a statutory board under the Ministry of Sustainability and the Environment (MSE) of the Government of Singapore, and is responsible for ensuring a sustainable and efficient water supply in Singapore. PUB regulates and oversees the country's entire water supply system, which comprises the water catchment systems, drainage systems, water works, pipeline network, water reclamation plants and sewage systems.

The solution

Real time advanced analytics provide vital insight into water performance

In a pilot project started in 2019, Aqua Suite Pure was deployed at PUB’s Ulu Pandan Integrated Validation Plant (IVP) for a two-year trial to validate the reduction of energy consumption in water reclamation processes, chemical use and reliance on manpower, while increasing water quality.

Aqua Suite was introduced to provide the operator and management with predictive insights, while improving plants’ performance through machine learning and control. Connecting to the plant’s Supervisory Control and Data Acquisition (SCADA) system to gather data and control key processes, it sends the data to the Aqua Suite cloud solution, through a secure data diode. With the data collected in near real-time in the cloud, the software is able to track actual performance of the on-premise solution, via a digital twin.

More advanced analytics are available to the operator through the cloud, including data-driven predictions, anomaly and drift detection, performance indicators and soft sensors. It also raises early warnings for specific events that require attention from the operator. Meanwhile, the on-premise part of the solution keeps optimising real-time performance. The approach of predicting patterns, learning process relationships and integrally optimising assets on a holistic level is key to this solution and can be applied across the whole water cycle. The limited amount of historical data, rather a few days to weeks instead of months or even years needed for providing optimised setpoints, makes this system easily and rapidly applicable, and therefore robust.

The impact

Self-learning predictive control provides significant benefits

Results at the IVP have also shown that Aqua Suite’s advanced analytics and predictive control have successfully predicted the influent ammonium load several days ahead, learning the operational performance of treatment and controlling treatment processes. Its smart autopilot has proven that it can perform unattended operation with self-learning predictive control. During the trial, it has also achieved significant aeration flow reduction of up to 15%, with corresponding energy savings. In addition, it has autodetected several deviations and has provided early warnings as part of the operational support. This has led to stable operation and effluent quality.