Calculation of BSRM's inductance with PSO-BPNN

Inductance characteristic has a great effect on bearingless switched reluctance motor (BSRM) control which is difficult to solve accurately. Particle swarm optimization (PSO) is used in back propagation neural network (BPNN) inductance model. When the BPNN is trained with sufficient samples, PSO is applied to optimize weights of BPNN. Building the PSO-BPNN model of inductance and evaluating performances of the proposed model by error compute. The results demonstrate that PSO-BPNN inductance models perform satisfactory forecast accuracy and convergent speed.