Optimizing the Charging Parameters of Linear Motor Flux Pump with BP Neural Network and Genetic Algorithm

Charging superconducting coils with linear motor flux pump is affected by many parameters. In order to pump the maximum charging current, we accumulated a large amount of data in the experiment. Due to the principle of linear flux pump is not entirely clear, in this article, we mine the functional relationship between three important parameters (AC current, DC current, frequency) and charging current based on historical data, and adopt the method of BP neural network to predict an optimal charging scheme of flux pump by using the genetic algorithm to optimize the BP neural network parameters. Experimental results show that this method is more efficient than the manual random method, which provides a direction for flux pump design and charging parameter adjustment.