Simulation based optimal design and sensitivity assessment of a vector-controlled induction motor drive using a multi-modal optimization algorithm

In this paper an optimized design of a vector-controlled induction motor drive system is presented using a novel parallel multi-modal optimization algorithm. The algorithm is highly parallelizable and is implemented on a 60-core parallel processor and yielded 16.4 times speedup over the sequential approach. The most robust optimal design of the vector-controlled drive system can be achieved using surrogate models that are obtained by the algorithm during the course of optimization. Using the surrogate models as a replacement of the detailed simulation model, sensitivity analysis can be conducted without any extra simulation runs.