Real-time particle swarm optimization based parameter identification applied to permanent magnet synchronous machine

Particle swarm optimization (PSO) has been widely used in optimization problems. If an identification problem can be transformed into an optimization problem, PSO can be used to identify the unknown parameters in a nonlinear model that is used to describe a system. Currently, most PSO based identification or optimization solutions can only be implemented offline. The difficulties of online implementation mainly come from the unavoidable lengthy simulation time to evaluate a candidate solution. In this paper, a technique for faster than real-time simulation is introduced and implementation details of PSO based identification algorithm is presented. Performance of the proposed technique is demonstrated through application to parameters identification of permanent magnet synchronous machine control system. The algorithm is implemented in Matlab/Simulink with the most fundamental blocks and Embedded Matlab Functions. Thus the program can be compiled to C/C++ code through Real-time Workshop and be able to run on hardware controllers like dSPACE. The proposed techniques can also be applied to many other online identification and optimization problems.