PERFORMANCE OF SEVERAL HISTORIC CANAL CONTROL ALGORITHMS ON THE ASCE TEST CASES

Authors: Wahlin, Brian; Clemmens, Albert

Submitted to: Workshop on Modernization of Irrigation Water Delivery Systems
Publication Type: Proceedings
Publication Acceptance Date: 10/21/1999
Publication Date: N/A
Citation: N/A

Interpretive Summary: Water is becoming a scarce resource, and irrigation districts are under pressure to use water more effectively. Computerized automatic control of irrigation canals can potentially improve the operation efficiency of irrigation districts. Many canal control algorithms have been developed; however, a direct comparison between these algorithms has been difficult. Recently, a set of standardized tests for control algorithms was developed in an attempt to make comparisons between these algorithms easier. Several historic control algorithms were run using these standardized tests, and their performance relative to each other was examined. These results should be of use to irrigation districts and consultants.

Technical Abstract: A wide variety of algorithms have been used in an attempt automatically to control water levels in irrigation canals. Simple proportional-integral (PI) control logic (ELFLO) has been used for downstream water level control, but with mixed results. Heuristic downstream water level controllers (Canal Automation for Rapid Demand Deliveries, or CARDD) have also been proposed but not rigorously tested. Both algorithms represent a series of gate position controllers for the downstream pool water level with one controller per pool. Because of the hydraulic interaction between the pools, the individual gate position controllers have a tendency to interfere with each other. Decoupling is one way to remove the interactions between the individual controllers by using a partial decoupler (Decoupler I) or a full decoupler (Decouplers I and II). To make comparisons between different control algorithms easier, the ASCE Task Committee on Canal Automation Algorithms developed a series of test cases to compare the general suitability of various control algorithms. These ASCE tests were performed on the standard PI control logic, the PI control logic with hydraulic decouplers, and the heuristic CARDD control logic. Initially, the control algorithms were tuned under known hydraulic conditions. The control algorithms were then tested under slightly different hydraulic conditions in an attempt to determine how robust the algorithms are. The performance of the PI control logic was improved with the addition of Decoupler I. No significant improvement was seen with the addition of Decoupler II. CARDD was not able to handle the ASCE test cases as well as the PI control logic. All of the controllers appear to be fairly robust in that they can handle the untuned test cases with only minor degradation in performance.