Mid-term predicting of wind turbine power generation using Artificial Neural Networks

Wind turbine power has a great effect on environment, power system operation and market economics. Predicting wind energy has been a vital part of wind farm planning and power generation with the aim of reducing fuel cost and reducing greenhouse gas emission. This paper presents Artificial Neural Network (ANN) approaches for forecasting monthly wind energy by climatological variables. The ANN was trained by two methods in multi-layer neural networks, first is Radial Basis Function Network (RBFN) and second is feed-forward neural network with Levenberg-Marquardt algorithm for variable hidden unit. The neural network gets wind speed, temperature and humidity as inputs (climatological data) and produce wind turbine energy as output. The algorithm is evaluated using simulation performed with MATLAB. The acquired results of the integrated model have shown high accuracy of about 100%. Therefore, the proposed approach can be used as an efficient tool for prediction of wind energy in the all of the locations in Iran or any country that don't have wind farms. This approach can be useful for predicting of electricity power generation on built wind farm and also can be useful for locating new wind farms in order to get higher efficiency in electricity power generation.

کلیدواژه ها:

Wind turbine power ، Artificial Neural networks ، Mid-term Predicting ، Climatological Variable ، Iran