Markov Models Based Short Term Forecasting of Wind Speed for Estimating Day-Ahead Wind Power

In order to meet the growing demand of energy, renewable resource utilization has increased in recent years. Wind is the source to a significant percentage of renewable resources and wind farms harvest this energy into electricity with the help of wind turbines. These turbines are very costly to set up and require high amount of maintenance. Accurate short term (from 30 minutes up to 6 hours ahead) wind energy forecasting is therefore important for optimal scheduling of the wind farms. The paper explores the usage of Markov Chains for forecasting wind speed during a short-term period (day-ahead hourly wind generation forecasts for an individual wind farm). The proposed prediction model depends on one variable factor - wind speed, for a specific wind turbine. The geographical location under study is taken at Jodhpur in Rajasthan, India. The performance evaluation of the proposed method is calculated using the different statistical error measures like RMSE, MAPE and MAE.