Particle Swarm Optimization for Robust Control Tuning Applied to Uninterruptible Power Supplies

Abstract

This paper provides as a contribution a procedure for robust state feedback control design that optimizes the weights of a discrete linear quadratic regulator through a particle swarm algorithm. Here, the procedure is applied to the voltage control of an uninterruptible power supply, which is an important system for energy storage in microgrids, providing backup power for critical loads during utility system outages. The proposed procedure is based on the optimization of a practical objective function that encompasses the tracking error and the control signal effort considering uncertain load conditions. The control is designed offline, taking into account implementation delay of digital control signals and multiple resonant controllers. Real time simulations based on hardware-in-the-loop indicate the practical feasibility of the procedure for the application of robust control in power converters. The international standard IEC 62040-3 is used to validate the results.