Model Predictive Control of an Experimental Water Canal

Abstract

This paper presents the first experimental results of a model predictive controller designed to control the water levels of a modern automated water canal, located at the University of Évora, Portugal. The controller is able to correctly handle known-in-advance water offtakes, while considering the most relevant physical constrains of the experimental setup. The dynamic model used is based on discretized Saint-Venant equations linearized for the steady-state regime taking into account the hydraulic structures present, such as gates and a water offtakes valves. The controller is implemented on a PLC (Programmable Logic Controller) network supervised by a SCADA (Supervisory Control And Data Acquisition) system. The experimental results demonstrate the reliability and effectiveness of the proposed control scheme in real-life typical situations, including gate malfunctioning and extreme water offtake conditions.