Wind Turbine Torque Oscillation Reduction Using Soft Switching Multiple Model Predictive Control Based on the Gap Metric and Kalman Filter Estimator

Abstract

A new multiple model predictive control (MMPC) is reported to regulate the output power of the National Renewable Energy Laboratory (NREL) 1.5 MW baseline wind turbine (WT). The maximum power point tracking in the wide range of the partial load regime without losing any control performance is the goal of the proposed algorithm. To do this, the whole area of the partial load regime is divided into adequate linear models using a mathematical tool called gap metric. Next, we set different MPCs (controller bank) to each model. The controller selection is conducted based on wind speed value estimated using a neural network estimator. The sudden change between controllers applies a hazardous chattering to control signal. To tackle this problem and reduce the chattering, a new soft switching MMPC (SSMMPC) is introduced. Furthermore, the stability of the closed-loop system is guaranteed using Lyapunov theory creating novelty in this work. Moreover, to validate the proposed algorithm, SSMMPC and classical MMPC are implemented on the NREL 1.5 MW baseline WT using fatigue aerodynamic structure and turbulence simulator. Finally, the simulation results and comparisons demonstrate the effectiveness of the proposed controller in the tracking error, oscillation in control signal, and mechanical power.