Optimal Control of a Partially Time Lagged River Aeration System Subject to Random Disturbances

Abstract

Pollution abatement by means of artificial aeration is considered for rivers which are contaminated by distributed effluent sources. The system model includes plant and observation noises to account for the poor reproducibility of water quality measurements and for the poor accuracy with which plant driving terms are known. An optimal feedback aeration policy is developed which operates upon those water quality measurements which are available at scattered measurement stations along the river. Because some types of water quality measurements are more time consuming and more difficult to make than others, a state estimator is developed for use in the feedback structure using partially delayed and intermittent data.