Efficiency Enhancements of Wind Energy Conversion Systems Using Soft Switching Multiple Model Predictive Control

Abstract

The intermittent nature of wind speed imposesa nonlinear behavior on the dynamics of the wind turbine. Consequently, relying on one linear controller tuned to a single specific operating point cannot guarantee a feasible performance in the whole operating region despite providing a fast solution. Besides, wind speed variations cause oscillations in the output power of the wind turbine. To tackle these issues, the Soft Switching Multiple Model Predictive Control (SSMMPC) technique is introduced in which the nonlinear dynamical model of the wind turbine is approximated by defining multiple linear models around various operating points. The gap metric is used as a mathematical tool to construct a bank of multiple linear models that results in a bank of linear controllers. By assuming a relatively small gap threshold value between the neighboring linear models, the soft switching is guaranteed. In this way, there will be no unexpected impulsive behavior in the control input during switching from one MPC to another. The closed loop system stability is studied using Lyapunov theory and a novel switching stability proof based on probabilities is provided. The SSMMPC performance is investigated and compared with a TSR-based controller and a bidirectional control method through simulations using the NREL 10 kW wind turbine model and experimental results on a 7.5 kW wind turbine drivetrain. The results corroborate the improvements that can be attained by using SSMMPC in terms of more captured energy, better maximum power point tracking quality, lower generator torque oscillations, and smoother output power curve.