Development of MPC for an Uncontrolled Prime Mover Driven SCIG based Distributed Generation System

Abstract

A finite control based model predictive control (MPC) is developed for the squirrel cage induction generator (SCIG) fed to stand-alone load via a back to back converter. The proposed MPC can handle both source and load side variations through the combined action of both the generator and load side controllers. The generator side converter's control performance is three-stage, whose objective is to keep the dc-link voltage constant for the operating scenarios. A load voltage controller with the inclusion of an LC filter parameter is proposed to maintain the load voltage amplitude and frequency. Further, the developed scheme is flexible for various types of cost functions. The generator side converter's cost function is modeled in the rotor flux reference frame, whereas the load end converter is implemented in a synchronously rotating reference frame. The system is tested for the uncontrolled prime mover driven distributed generation scheme. The proposed scheme's efficacy is validated with a 33% variation of source and 50% load perturbation. The machine and load variables' dynamic performances are found to be satisfactory under the considered operating conditions.