Research and Application of Wavelet Neural Network Based on the Optimization of Genetic Algorithm in Centrifugal Compressor's Performance Prediction

In order to better predict the performance of centrifugal compressor and find out hidden problems as early as possible, we combine genetic algorithm, wavelet theory and artificial neural network together to establish the performance predictive model for centrifugal compressor based on wavelet neural network optimized by genetic algorithm. This predictive model can help us monitor the performance change of centrifugal compressor. The simulation experiment has shown that this model owns the advantages such as simple algorithm, stable structure, fast convergence speed, high identification precision and strong generalization ability, which all have great practical use.