شبيه سازي داده هاي سرعت باد با استفاده از مدل هاي استوكاستيك

speed simulation using stochastic models

In the last decade wind energy has been exensinly used over the world. It is necessary to have long period wind speed data and their statistical characteristics for determined wind energy generated, using of Wind Energy Eonversion Systems and decreasing the damage of wind erosion. In this study, Weibull distribution, Markov chain and time series methods were used to generate hundred 100 years tree hourly wind speed values at Zabol and Yazd stations. In Markov chain first, transition probability matrix were developed from the real wind data, cumulative probability matrix were determined and then wind speed generated. For time series method one model was developed for each month and in most months AR2 and AR5 were used. Result showed that time series and Weibull distribution have better performance in average and standard deviation criteria and for Autocorrelation and Skewness parameter time series and Markov chain had better application. Regarding to serial correlation in wind speed series time series model had better application and suggested for wind speed generation.