MAXIMISATION OF WIND FARM POWER PRODUCTION USING THE TEACHING LEARNING BASED OPTIMISATION ALGORITHM

The wake effect is a significant challenge in wind farm power production, as it can greatly reduce the output power of the wind farm. Wake redirection control (WRC) is a wake control strategy that effectively improves wind farm power production. A real-time optimisation approach is developed to implement WRC in the wind farm. This approach uses teaching learning-based optimisation (TLBO) to solve the real-time wind farm optimisation problem. This algorithm is applied to a wind farm model of a real-life 20-turbine wind farm in South Korea. The model is created using FLORISSE_M, the MATLAB version of the FLORIS (FLOw Redirection and Induction in Steady-state) model. The objective function of the optimisation problem is to maximise the total power output of the wind farm, and control variables are the yaw angles of the turbines in the wind farm. The simulations are conducted for various mean wind speeds, i.e., 6, 8, and 10 m/s, and for different wind directions, i.e., 0 to 330 degrees, incremented by 30 degrees. The results show that implementing the real-time optimisation on a 20-turbine wind farm with WRC incorporated significantly improves overall wind farm power production. The optimised yaw angles from the FLORISSE_M are validated with a high-fidelity wind farm model called SOWFA (Simulator for Wind Farm Applications). The results of the SOWFA simulation show significant agreement with the optimised yaw angles from the FLORISSE_M model. The total wind farm power output is also improved in the SOWFA simulation.