Artificial neural network based radial bending characteristics of mixed-flow pump impeller

摘要

Impeller radial bending characteristic has applied to many type turbine machines except of pump. Specially, there is no news about application on the mixed-flow pump. In this study, an artificial neural network (ANN) was used for modeling the performance of mixed-flow pump impeller. Thirty seven results were used to train and test. Many patterns have been randomly selected and used as the test date. The main parameters for the experiments are the Gamma, Betal and Beta2. Gamma, Betal and Beta2 have been used as the input layer, and η has been used as the output layer. The best training algorithm and number of neurons were obtained. At last, a new type, high efficiency mixed-flow pump impeller can be designed.