LQR CONTROL DESIGN FOR A DC-DC CONVERTER USING SENSITIVITY FUNCTIONS
Open Access
- Author:
- Doddabasappa, Sushamshushekar
- Graduate Program:
- Electrical Engineering
- Degree:
- Master of Science
- Document Type:
- Master Thesis
- Date of Defense:
- July 02, 2019
- Committee Members:
- Javad Khazaei, Thesis Advisor/Co-Advisor
Peter Idowu, Committee Member
Scott Van Tonningen, Committee Member - Keywords:
- LQR Control
Sensitivity Functions
Sensitivity Analysis - Abstract:
- To fulfill the ever-increasing energy demand and to curb the innumerable threats that conventional energy generation techniques are posing, we are forced to explore the use of renewable energy sources. The contribution of renewable energy sources in total primary energy consumption has been steadily rising in the past few decades. The increased penetration of the renewable energy sources into the power system network have provided enormous opportunities and their own set of challenges to deal with. Hence, the study of various control strategies and their performances employed to integrate the renewable energy sources to the existing power system network is of prime importance. In this research, a novel technique is proposed to analyze the performance of the close loop system used to integrate a renewable energy source to the power system network. A solar photovoltaic array connected to a load interfaced via a dc-dc converter is considered as the test system in this research. An optimal control theory based linear quadratic regulator (LQR) controller is designed for the closed loop operation of the system. Derivative based sensitivity functions are formulated to validate the performance of the designed closed loop controller. The contribution of closed loop controller to the stability and overall system response is analyzed using the sensitivity functions. The system characteristics such as settling time and percent overshoot are modified to meet the desired requirements using the derived sensitivity functions. A simplified model of the system is built using MATLAB/Simulink and several case studies are performed on the test system to validate the claims of this research.