The RBF Neural Network Optimization Based On Hybrid Hierarchy Genetic Algorithm And It’s Application In BOD Soft-sensing

With the industrialized development of society, the sewage treatment of cities and industries has become an imminent important topic in the field of environmental protection. Unforeseeable interference is very large in the biological process of sewage treatment which has some typical nonlinear, unstable and multivariable characteristics. So it is difficult to establish accurate mathematical model, and the key parameters of water quality (especially the biochemical oxygen demand BOD on behalf of sewage treatment performance) can not be measured online.According to these problems which the key parameters of water quality are hard to measure online monitoring in the process of sewage treatment. This thesis proposes a BOD soft-sensing model based on RBF Neural Network Optimization Based on Hybrid Hierarchy Genetic Algorithm.At first, this thesis uses the method of principal component analysis to perform both dimension-reduction and de-correlations in input space in order to simplify the inputs of RBF Network, which can reduce the input variables from10-dimension to5-dimension by not less than85%cumulative contribution rate.Then the structure and parameters of the hidden layer can be regarded as a whole, and the coding is chromosome. This thesis uses hierarchical genetic algorithm with the structure of double gene on chromosome to optimize these parameters of RBF hidden layer, while the output layer weights are determined by the least square method. The soft sensor model has strong adaptability and good fault tolerance performance by using the hybrid hierarchy genetic algorithm to optimize RBF neural network model. Finally, the soft sensor model presented in this thesis is compared with other models, the simulation results show that the new model presented has good control accuracy and stability in this thesis.This thesis uses the hybrid hierarchical genetic algorithm and RBF neural network in BOD soft-sensing, the new method not only can promote the real-time closed loop control of the sewage treatment process, but it also has a positive significance for developing soft measuring instrument with intelligent control algorithm in the other industrial process.