Experimental Validation of Extremum Seeking Control for a Midsize Hydrostatic Transmission Wind Turbine

Extremum Seeking Control (ESC) is a model-free control strategy that uses a dither signal to enable a gradient search to find an optimal operating point. The adaptability and model-free nature of the ESC can be beneficial to improve power capture of wind turbines because the Cp (power capture coefficient) vs. λ (tip speed ratio) curve is constantly changing in time and is dependent on uncertain parameters that are not easily available. For conventional gearbox wind turbines, ESC is used to continuously search for the optimal torque gain, k, of the commonly used kω2 law, improving the power capture when compared to control using a fixed k gain. Our group has expanded the studies of ESC to Hydrostatic Transmission (HST) wind turbines through simulation. The work studied the effectiveness of ESC for a midsize HST wind turbine, under steady, step, and turbulent wind, and its robustness to changes in the Cp vs. λ curve. Previous simulation results are experimentally validated here using the HST dynamometer at the University of Minnesota. Experiments show that ESC can effectively adapt to changing Cp vs. λ curves and results are compared to the simulations validating the results. Use of an anti-windup algorithm is also discussed.