Development of Wind Speed Forecasting Model Based on the Weibull Probability Distribution

Wind is a variable energy source. The power output of a wind turbine generator (WTG) unit, therefore, fluctuates with wind speed variations. Accurate models reflecting the variability of wind speed is hence required in both reliability evaluation of power systems containing wind energy and analysis of the impact of integration of wind power to electric power networks. This paper presents an improved maximum likelihood method for wind speed forecasting to calculate the two parameters of Weibull distribution. The model overcomes some of the disadvantages such as complicated calculation procedure and slow convergence associated with the maximum likelihood method. The method proposed in this paper extends the applications of Weibull wind speed probability distribution based wind speed forecasting. As compared to the method of mean wind speed and standard deviation for calculating Weibull parameters, the proposed technique can maintain higher accuracy even in complex terrain areas. The range of application and the efficiency of the model are also examined by comparing the forecast error and autocorrelation coefficient of the proposed model with those of time series models.