Maximizing Wind Turbine Efficiency by Using Soft Switching Multiple Model Predictive Control

Abstract

Variable speed small to medium wind turbines need to cope with the intermittent nature of wind speed at lower altitudes. This imposes challenges on optimally tracking the maximum power point (MPP) during partial load and makes the wind turbine dynamics highly nonlinear. As a result, using one linear controller around a specific operating point may not guarantee acceptable performance in the other operating points. In addition, wind speed variations cause fluctuations in the output power of the turbine. The Soft Switching Multiple Model Predictive Control (SSMMPC) technique is introduced to tackle the latter problems when considering multiple linear models around various operating points (MPPs) approximating the nonlinear dynamics. The gap metric method is used to assess how close different linear models are with respect to each other. The closed loop system stability is validated using Lyapunov theory. The controller performance is investigated and compared with a bidirectional TSR-based controller through simulations using the FAST NREL 10kW wind turbine model. The results verify the improvements that can be attained by using SSMMPC in terms of higher maximum power point tracking quality, lower generator torque oscillations and smoother output power, consequently.