Reinforcement Learning Based Wind Farm Layout Optimization

Abstract

Wind farm layout optimization is a the major decision factor for maximum utilisation of wind energy for large scale wind farms. As more methods are being researched to reduce losses in the wind power plants, none more effective in reducing the over all loss than the loss due to wake effect. The arrangement of location of the turbines influence the power generation as well as levelized cost of energy. In order to minimise over all loss of the power plant, effective positioning of the turbines is needed. In this study, a novel turbine layout optimization method utilizing reinforcement learning is implemented for a wind farm. Turbulence intensity and the deficit velocity due to wake loss from Gaussian wake effect is used as the input for the model. The simulated results from the wind resource assessment software, WAsP suggests that the proposed method is effective for the number of turbines used in the study.