LQR-Based Control with Gravity Compensation for a Wind Turbine Pendulum System.

Abstract

In this project, we present the design, construction, and control of a pendulum-like system propelled by a propeller mounted on a brushless motor. Programming and data acquisition were performed using an Arduino UNO board connected to a computer through the Arduino Integrated Development Environment (IDE). The measurement of the pendulum's angular position was conducted using a variable resistor. However, a drawback of this approach is its susceptibility to introducing noise into the signals. To address this issue, a Kalman filter was implemented in the analysis. The system was mathematically modeled as a second-order transfer function with underdamped poles and identified using the adaptive Gauss-Newton method. Experimental tests were conducted with step response trials, employing pulse width modulation (PWM) as the input and the resistor voltage as the output variable. The pendulum position was controlled using a Linear Quadratic Regulator with Gravity Compensation (LQR+G) and a Proportional-Integral-Derivative (PID) controller. Finally, a comparative analysis of performance was carried out between both approaches.