Current Distribution Optimization by Using Genetic-Algorithm Based On-Off Method: Application to Pellet Injection System

Abstract

The current distribution in the electromagnet is optimized by using the geneticalgorithm based on-off method so as to maximize the acceleration performance of the Superconducting Linear Acceleration (SLA) system. In the SLA system, a pellet container is accelerated by the interaction between a shielding current density and an applied magnetic field. By using the equivalent-circuit model, the distribution of the shielding current density is approximated as a set of the multiple current loops. In contrast, the current distribution in the electromagnet is represented by means of the on-off method. As the method for optimizing the current distribution in the electromagnet, two types of genetic algorithms are adopted. The results of computations show that the pellet velocity for the optimized current distribution is 1.3 times as fast as that for the homogeneous current distribution.