Wind characterization analysis incorporating genetic algorithm: A case study in Taiwan Strait

In this paper, the genetic algorithm (GA) is originally applied to compute the Weibull parameters for wind characterization analysis, in which an objective function required in GA for searching optimization solution has been first defined as well. Wind data analyzed are observed at a wind farm in the Taiwan Strait from 2006 to 2008. To accurately describe wind speed distribution three kinds of probability density functions are compared, i.e. the Weibull, logistic and lognormal functions. Statistical parameters including the max error in the Kolmogorov–Smirnov test, root mean square error, Chi-square error and relative error of wind power density are considered as judgment criterions. The results show that GA is a useful method, there is about 33% time saving when compared with conventional iteration method. Weibull function describes best the wind distribution, regardless of time periods. Accordingly, wind power density, availability factor and electrical energy output from an ideal turbine are assessed using the Weibull parameters; utilization rate of wind energy for the currently used turbine is discussed. Further the wind energy compensates very well with solar energy; when solar radiation is down in winter and spring, the wind power becomes greater; energy ratios for each month are calculated lastly.

► The genetic algorithm was applied for the first time to calculate the Weibull parameters for wind energy assessment.
► Weibull probability function fits the observed wind speed distribution better than both logistic and lognormal functions.
► Wind and solar energy potential in Taiwan show a great complementary relationship.