Study on Wind Characteristics Using Bimodal Mixture Weibull Distribution for Three Wind Sites in Taiwan

Some wind speed distributions in Taiwan have been found deviating from conventional Weibull distribution. In this paper the mixture Weibull distribution was adopted to analyze the wind data observed at three wind sites having different climatic environments. The Kolmogorov-Smirnov test and wind potential energy were considered as indicators to show how the mixture Weibull function characterizes wind speed distribution. Relevant mathematical expressions are derived originally for wind energy assessment. The results show that the mixture Weibull function performs quite better than a conventional Weibull function particularly for a region where the wind speed distribution reveals two humps on it. The similar result is obtained also when wind power density is considered. The maximum errors of cumulative distribution function between observation data and mixture Weibull function are always below the critical value of 95% confidence level in Kolmogorov-Smirnov test. The relative percentage errors of wind potential energy between time-series data and theoretical values from mixture Weibull function never exceed 0.1%. It is found that the distribution pattern of wind speed would affect a lot to the electrical energy generated by an ideal turbine.