PI Controller Tuning Optimization for Grid-Connected VSC using Space Mapping

Abstract

Controller tuning of voltage source converter (VSC), using voltage oriented control (VOC), has a significant impact on the system stability against disturbances. To this end, optimization algorithms are sought in order to achieve optimum dynamic performance. Such algorithms are normally applied to numerical simulators that are time-intensive, which hinders the design process. This paper proposes, for the first time, space mapping (SM) optimization algorithm for controller tuning in a grid-connected VSC system. SM is a surrogate-based optimizer that utilizes a ‘coarse’ model that is less accurate, yet extremely fast, to guide the optimization of the numerical, time-intensive ‘fine’ model. The development and employment of both models are presented in this study. Subsequently, the accuracy and efficiency of the SM algorithm are assessed by simulations. It is found that SM optimization algorithm can return a quasi-optimum solution in less than half the time taken by optimizing the fine model with adequate accuracy.