In order to solve the problem of combining pump hydraulic model with motor electromagnetic calculation model, this paper adopts BP neural network combined with genetic algorithms (GA-BP) to fit the model of pump hydraulic performance. Combine pump GA-BP model with motor electromagnetic calculation model to establish an integrated model. The main work is as follows:1. Based on the basic principle of BP neural network and genetic algorithm, this paper research the method which used to optimize the structure, weight, threshold of BP neural network with genetic algorithm. And write GA-BP model of water pump with C language;2. Comparing and analysis the results of GA-BP model and basic BP neural network model, include accuracy,speed, scale of the network and the ability of generalization, to verify the effectiveness of the method;3. The motor electromagnetic model and pump hydraulic model are connected with the interface variables, to establish a pump machine integrated model. Then comparing and analysis the results of BP neural network and GA-BP.The work of this paper is supported by the national science and technology support project of china 2013BAF01B02. |