Wind Power Prediction Based on Wind Velocity Variable and Polynomial Regression Method.

Predicting the output energy obtained from wind velocity is required to investigate the wind power characteristics at potential locations and accurate evaluation of the power fluxes generating into the wind farms. This study provides a proposed approach to predict and assess wind power, using wind turbine parameters as a learning factor and daily wind velocity dataset of 8 cities in Iraq, namely Duhok, Mosul, Kirkuk, Baghdad, Najaf, Wasit, Qadisiyyah, and Basra. The nonlinear predictive models are established via the polynomial regression technique (POR) as a less complex machine learning method to extract potential wind energy from the wind turbine model (GhrepowerFD21-50_61.2kW_21.5). According to the outcomes, the scatter points of the wind power prediction values at all locations can follow the power curve output of the suggested turbine model with high accuracy and minimal error. Also, the 3 cities of Wasit, Qadisiyyah, and Basra have best total annual wind power quantities, with more than 2000 Kilowatt (kW) for each turbine unit, outperforming Duhok and Mosul that were less than 1000 kW. In addition, the POR technique was useful in extracting annual wind energy and estimation procedures for the regression method.