Comparison of wind turbine LQG controllers designed to alleviate fatigue loads

Abstract

This paper focuses on the design of Linear Quadratic Gaussian (LQG) controllers for variable-speed horizontal axis Wind Turbines (WTs). Those turbines use blades pitch angle and electromagnetic torque control variables to meet specified objectives for Full Load (FL) zone. The main control objectives are to reduce the structural dynamic loads and to regulate the power of the WT. The controllers are designed in order to optimize a trade-off between several control objectives. Four different LQG controllers are designed. Their control model is progressively more complex. The first one takes into account a rigid simple behavior, the second control model considers the first mode of flexibility of the drive train, the third model takes into account the tower flexibilities and the fourth that of the blades. In the same manner, their optimization criteria considers for each controller a new control objective to alleviate fatigue loads in the drive train, then, also in the tower and finally also in the blades. The evaluation of the fatigue loads affecting the WT components are based on a Rainflow Counting Algorithm and the Miner's rule (RFC). The results indicate a significant reduction of fatigue loads in the drive-train when its flexibility is taken into account in the control model.