Linear quadratic optimal control of multi-cascaded canals

Water delivery and distribution for irrigation through canals is the most important part of the internet of water (iWater). The automatic control level of this component directly affects the efficiency and risk control ability of the whole water supply system, and is of great significance for efficient management and utilization of water resources. By discretizing and transforming the integrator-delay model, discrete-time state-space model is formulated. Then optimal state-feedback controller is designed based on linear quadratic (LQ) theory and uses canal automation simulation software SOBEK along with MATLAB to test the proposed controller on the multi-cascaded canals of Changma South Irrigation District (CSID) in Gansu Province. The simulation results show that LQ optimal controller can effectively handle designed changes in water demand, stabilize the system in a reasonable time and steadily maintain the water level within the operational range, and thus the performance of the designed controller is much better than classic decentralized PI controller. In short, LQ optimal controller can effectively cope with time delays, coupling characteristics and unscheduled water intake disturbances inherent in multi-cascaded irrigation canals, and thus achieve reliable water supply service and efficient irrigation canals operation control. However, since LQ controller only uses feedback control logic, it cannot cope with excessively large unscheduled water intake disturbances effectively.