Predictive control as an intelligent tool to manage water distribution networks

Fresh water is becoming a major concern in actual societies as it represents only 2.5% of the total Earth water reserves. Some recent studies point the year 2025 when 2 of every 3 persons will be affected by the lack of fresh water. This paper presents a predictive controller strategy that is implemented on a modern automated water canal where sensors and actuators are controlled via a PLC (programmable local controller) network supervised by a SCADA system (supervisory control and data acquisition). This canal prototype is composed by a set of distributed sub-systems that manage local control to assure the water level in each canal pool, defined by discharge gates (control variable) and water off-takes (disturbances). All local controllers are connected through an industrial network to be assessed by a SCADA system where the global overview and the centralized predictive control algorithm is running. Extreme severe situations have been simulated and the obtained results proved the very good robustness of the developed controller