Вероятностные нейронные сети в задачах управления комбинированными системами водоочистки

Abstract

The shortcomings of the modern systems of the selecting management strategies in times of unforeseen factors and the shortcomings of the current approaches in the management of water treatment plants combined are analyzed. The theoretical foundations of probability theory of the neural networks and Bayesian networks are considered. On the basis of self-organization formed Kohonen network the quality of water, where the control parameters used BOD, pH, concentration of suspended solids and the concentration of nitrate the probabilistic neural network evaluation of dynamics of images is synthesized. It has tested its functionalefficiency and confirmed the prospect use. The block diagram of the corresponding control system with integration into it PNN-network described. The direction of the future research is formulated.