Dynamic programming applied to adaptive canal control.

摘要

This thesis presents a solution to the control of irrigation canals by applying the dynamic programming control algorithm to an adaptive canal automation system. The system consists of software improvements to the already existing system designed by Dr. Walker. The objective of the dynamic programming system is to improve control efficiency by finding optimal control strategies, to adapt to changing canal conditions, and to provide an "institutional memory" of how to control the canal. The idea behind the control framework is not to replace the water master but to develop control techniques to assist the irrigator in planning day-to-day gate movements and to satisfy irrigation demands while minimizing the fluctuation in the canal levels.;The system was tested using simulation to verify operation of the controller and the dynamic programming system. The development used a canal model derived for the use of the dynamic programming system and also for simulation tests. (Abstract shortened by UMI.).