Reducing bursts and background leakage with predictive pressure control

Pressure management is widely recognised as the foundation of effective leakage control in drinking water distribution systems. Conventional pressure management strategies primarily aim to reduce average network pressure, which lowers leakage flow but does not necessarily minimise pressure fluctuations. Recent research indicates that pressure variation—particularly daily—is a dominant driver for pipe fatigue and bursts. This paper presents a predictive pressure control approach that targets leakage reduction and burst prevention by reducing pressure variability while maintaining compliance with minimum service level requirements. The method combines demand forecasting with learned hydraulic relationships to determine optimal pressure setpoints in advance, enabling smoother and more stable network operation. Unlike fully measurement?based control strategies, the proposed approach does not rely on continuous real?time critical point measurements and is therefore robust to delayed communication or temporary sensor failures. The approach is operational at several drinking water utilities. Results show that predictive control achieves pressure levels comparable to advanced measurement?based control, while significantly reducing pressure fluctuations. This operational behaviour is expected to reduce background leakage, lower burst frequency, and contribute to extended asset lifetime.