Robust PI cascade control for a multi-mass system optimized by evolutionary algorithms

Abstract

Controlling a multi-mass system is a common problem in industrial automation, and cascaded PI controllers are a popular control structure for these systems. However, the parameters of a multi-mass system are often uncertain or can change. A robust Hinfin-control is a powerful tool in order to control systems with varying parameters. Unfortunately, the complex structure and the mathematical theory prevent known Hinfin-controllers from usage in many industrial applications. In this work two modern evolutionary algorithms, differential evolution (DE) and particle swarm optimization (PSO), are used to optimize a standard PI cascade control structure with respect to the Hinfin-norm, thus resulting in a robust control of simple structure.