Optimal Real-Time Control of Wind Turbine During Partial Load Operation

Abstract

A wind turbine achieves its highest energy efficiency during partial load operation when it operates at the optimal tip speed ratio (TSR), thus the optimal power coefficient. In this paper, real-time controllers are developed to improve the performance of tracking the optimal TSR during partial load operation. Dynamic programming (DP) is first applied to determine the control actions that maximize wind energy capture. A DP-based real-time controller (DPRC) is then explored to overcome the high computational expense associated with DP, which limits DP to be an offline optimization algorithm. However, the DPRC is not robust against plant-model mismatch and model uncertainties. A gain-modified optimal torque controller (GMOTC) is subsequently designed as an alternative to the DPRC. The GMOTC applies internal Proportional-Integral technique to track a reference TSR, and adapts the reference TSR to the optimal TSR in real time to improve the controller robustness. The light detection and ranging technology is used to further strengthen the controller performance by providing reliable previewed wind speed measurements. Simulation results show that the DRPC generates more wind power than the standard torque controller (STC), while the GMOTC demonstrates a performance similar to that of the DRPC on wind power generation with much better robustness in the presence of modeling error. Fatigue loading on a wind turbine is another important issue that needs to be considered during control design. The analysis shows that both the DRPC and the GMOTC are comparable with the STC in generating variable torsional loads due to the torque commands.