On leaning algorithm and soft sensor model of swage disposal based on process neural network

Abstract

To solve the problem that water quality of sewage disposal process (such as BOD) is difficulty to measure on-line, meanwhile considering the characteristics of sewage disposal process which is related with time. A soft sensor method for water quality of swage disposal based on process neural network (PNN) was proposed in this paper. On the basis of learning algorithm based on orthogonal function basis expansion, in order to improve the learning rate, the function momentum adjustment item was introduced, moreover, genetic algorithm was used to optimize learning rate and realized learning rate adaptive adjustment algorithm. The soft-sensing model was trained and simulated by a lot of observed data, the experimental results show that the method is effective. So it can implement the real-time and close-loop control of sewage disposal process and have a broad perspective in application.