Design and implementation of artificial neural networks to predict wind directions on controlling yaw of wind turbine prototype

Abstract

Wind energy as one of the new renewable energies has an important role in replacing fossil energy sources in Indonesia. To make the wind turbine's performance more efficient in extracting energy from the wind, it is necessary to control the actuation movements pitch and yaw of the horizontal wind turbine. Controlling the yaw actuator can increase the absorption efficiency of the power to the rotor face towards the direction of the wind. The purpose of this work is to be able to predict the direction of the coming wind, then move the turbine rotor into the predicted direction. In this work, a wind turbine prototype is used with a precision of 5.3%, then for the data acquisition section, a wind direction sensor is built to change the amount of wind direction to a quantity that can be measured in units of degrees, and anemometer to measure wind speed. In making the wind direction prediction algorithm, an artificial neural network (ANN) method is used with input parameters such as wind speed, temperature, humidity, pressure, and altitude. Data acquisition is done at one-minute intervals with long data collection for one day, 1072 data are obtained, the data is then fed to the ANN model that has been prepared. Based on the results of tests that have been done, it is found that the lowest mean absolute error that can be achieved in the model is 0.4% and the model then be used to control turbine yaw movement using straight-forward method of combined method with movement error of seven percent, it is found that using this method can reduce the use of yaw movement while achieving same rotor rpm rate.