Neural Optimal Control Algorithm for Real Time Regulation of In-System Storage in Combined Sewer Systems

Recent advances in incorporating artificial intelligence concepts into real time control have reduced the barriers to successful implementation. A neural-optimal control algorithm is presented that fully incorporates the complexities of dynamic, unsteady hydraulic modeling of combined sewer system flows and optimal coordinated, system-wide regulation of in-system storage. The neural control module is based on the recurrent Jordan neural network architecture that is trained using optimal policies produced by a dynamic optimal control module. The neural control algorithm is demonstrated in a simulated real-time control experiment for the West Point combined sewer system, Metro Seattle. The algorithm exhibits an effective adaptive learning capability that results in improved performance of the control system.