A Topology Optimization of Hydroelectric Generator Using Covariance Matrix Adaptation Evolution Strategy

This paper presents a topology optimization of hydroelectric generator. The present optimization is based on evolutional method to easily consider multiple features, such as iron loss and waveform distortion. The machine shape is determined through the weighted sum of Gaussian functions, and their connection weights are optimized using covariance matrix adaptation evolution strategy (CMA-ES). The rotor shape of a hydroelectric generator is optimized using CMA-ES to maximize the output voltage and to minimize the iron loss. Numerical results show that the iron loss can be reduced with keeping the output voltage. In addition, its reasons are investigated.