A new method for short-term wind power forecasting

Utilization of wind power as renewable resources of energy has been growing rapidly around the world in the last decades. Wind power generation is fluctuating due to the variation of the wind speed. Therefore, the assessment of the output power of this type of generators is always associated with some amount of uncertainties. Accurate wind power forecasting can effectively support distribution and transmission system operators to improve power network control and management. This paper presents a new Imperialistic Competitive Algorithm-Neural Network (ICA-NN) method to enhance the short-term wind power prediction accuracy at a wind farm using information from Numerical Weather Prediction (NWP) and measured data from online SCADA. In this method, first, a prediction model of wind speed is built based on Multilayer Perceptron (MLP) artificial neural network considering environmental factors (i.e. wind speed, temperature, Humidity, geographical conditions and other factors) and then, Imperialist Competitive Algorithm is used to update weights of the neural network. The proposed method is capable to deal with jumping data; and is applicable in both wind speed and wind power forecasting.