Forecasting of Wind Turbine Output Power Using Machine learning
Haroon Rashid, Waqar Haider, Canraş Batunlu
Abstract
Most of the countries around the world are facing huge environmental impact, and the most promising solution to mitigate these is the use of renewable energy, especially wind power. Though, the use of offshore wind energy is rapidly increasing to meet the elevating electricity demand. The researchers and policymakers have become aware of the importance of providing near accurate prediction of output power. Wind energy is tied to variabilities of weather patterns, especially wind speed, which are irregular in climates with erratic weather conditions. In this paper, we predicted the output power of the wind turbines using the random forest regressor algorithm. The SCADA data is collected for two years from a wind farm located in France. The model is trained using the data from 2017. The wind direction, wind speed and outdoor temperature are used as input parameters to predict output power. We test our model for two different capacity factors. The estimated mean absolute errors for the proposed model in this study were 3.6% and 7.3% for and 0.2 capacity factors, respectively. The proposed model in this study offers an efficient method to predict the output power of wind turbine with preferably low error.