Multi-objective optimization design of solenoid valve based on BP neural network

Abstract

The magnetic valve plays a key role in the air-flow weft insertion and injection system of the loom. In order to improve the weaving efficiency of the loom and solve the problem of solenoid valve response lag, an optimization method based on BP neural network prediction of electromagnetic force combined with multi-objective optimization genetic algorithm was proposed. Firstly, in view of the complex change of the magnetic resistance in the magnetic field caused by the change of the armature structure and position of the solenoid valve, which made it difficult to calculate the accurate electromagnetic force through the theoretical model, the BP neural network was used to predict the electromagnetic force of the solenoid valve. Secondly, NSGA-II was used to optimize the preserved BP neural network prediction model and the obtained mathematical model of the armature mass, and the Pareto frontier solution of the electromagnetic force and the armature mass of the solenoid valve was obtained; Finally, the optimal parameter combination was selected according to the standard closest to the original armature quality, and compared with the electromagnetic force of the original structure. The results show that the optimized electromagnetic force increases by 11.5%, but the armature mass decreases by 1%. The simulation experiment also verifies the effectiveness of the optimization method.

Keywords

• air weft insertion
• bp neural network
• nsga-ii
• electromagnetic force
• armature mass