Model Predictive Control Methods for Photovoltaic Electrical Energy Conversion Systems

Abstract

Solar photovoltaic energy systems (PV) have had a consistently increasing market penetration over the past seven years, with a total global installed capacity of over 500 GW. A PV installation must harvest the maximum possible electrical energy at the lowest cost to be economically justifiable. This presents many engineering challenges and opportunities within power electronics amongst which include low-cost power converter implementation, high reliability, grid-friendly integration, fast dynamic response to track the stochastic nature of the solar resource, and disturbance rejection to grid transient and partial shading.

This dissertation investigates the controls of the power electronic interface with the objective to reduce cost, increase reliability, and increase efficiency of PV energy conversion systems. The overall theme of this dissertation involves exploring the theory of model predictive control (MPC) within a range of applications for PV systems. The applications within PV energy conversion systems are explored, ranging from cell to grid integration.

MPC-based maximum power point tracking (MPPT) algorithm is investigated for the power electronics interface to maximize the energy harvest of the PV module. Within the developed MPC based MPPT framework, sensorless current mode and adaptive perturbation are proposed. The MPC framework is expanded further to include inverter control. The control of a single-phase H-bridge inverter and sub-multilevel inverter are presented in this dissertation to control grid current injection. The multi-objective optimization of MPC is investigated to control the dc-link voltage in microinverters along with grid current control. The developed MPC based MPPT controller is shown to operate with a single-stage impedance source three-phase inverter with PID based grid-side control.