Optimal Nonlinear Predictive Control for Canal Operations

Abstract

The predictive canal operation method (PCOM) provides optimal real-time control of irrigation canals with incorporation of current and forecasted demands. PCOM utilizes an efficient, nonlinear, optimal control algorithm interacting with a fully dynamic, unsteady-flow hydraulic model. Optimal predictive controls are determined for all check structures and demand points along the canal, as well as the intake pump station, if one is present. Optimal control is performed with consideration of hydraulic conditions along the entire canal, as modeled by a Preissman-type implicit scheme for the complete St. Venant equations. The algorithm generates predictive control policies (anticipatory control) and system-wide feedback control under complex hydraulic conditions, including zero flow and closed gates. Equable distribution of shortages is achieved in cases of water supply deficiencies. PCOM is applied to Main Canal 3 of the Formoso H project, Bahia, Brazil, as a case study, and performs well under severe operational constraints. The computational efficiency of the algorithm allows implementation on microcomputer controllers for real-time use.