Performance Prediction for a Centrifugal Pump with Splitter Blades Based on BP Artificial Neural Network

Abstract

AbstractBased on MATLAB, a BP artificial neural network (BPANN) model for predicting the efficiency and head of centrifugal pumps with splitters were built. 85 groups of test results were used to train and test the network, where the Levenberg—Marquardt algorithm was adopted to train the neural network model. Five parameters Q,Z,β 2,D i ,b 2 were chosen in the input layer, η and H were the output factors. Through the analysis of prediction results, the conclusion was got that, the accuracy of the BP ANN is good enough for performance prediction. And the BP ANN can be used for assisting design of centrifugal pumps with splitters. Meanwhile, the method of CFD flow field simulation was also used to predict the head and power for a centrifugal pump with splitters, and compared with that from the BPANN model. The comparison of prediction results and experimental value demonstrated that the prediction values acquired through numerical simulation and BPANN were uniform with the test data. Both methods could be used to predict the performance of low specific speed centrifugal pump with splitters.KeywordsBP artificial neural networkcentrifugal pumps with splittersPerformance prediction