A Comparison of PI vs LPV Controllers for a Doubly-Fed Induction Generator in a Microgrid

Abstract

State-space controllers have been central to modern control theory and are often preferred over classical controllers such as PI and lag-lead controllers. This is because they offer flexibility in design, such as being able to simultaneously incorporate performance and robustness requirements, and have a solid mathematical foundation. Recently, doubly-fed induction generators (DFIGs) for wind turbines have been modeled as linear parameter varying (LPV) systems, a special class of time-varying linear state-space systems, for which LPV state-space controllers can be designed. The LPV controllers have been introduced for DFIG mainly because of their novelty and that they can offer the typical advantages that come with state-space controllers. However, no comparison has yet been made regarding the control performance of LPV controllers versus more traditional PI controllers for DFIGs. In this paper, LPV controllers have been designed using linear matrix inequality (LMI) techniques and their performance is compared to PI controllers via simulation studies on a DFIG in a microgrid. Our study shows that if a mathematical model represents the DFIG accurately, then both controllers can be tuned to give comparable performance.