Weibull model for wind speed data analysis of different locations in India

Wind speed data should be fitted by a suitable statistical model like Weibull to determine expected number of hours per year in the critical wind speed range for a slender structure, which is required to determine the expected number of stress cycles in the projected working life of the structure. Apart from this, for the assessment of wind energy potential wind speed data should be fitted by an appropriate probability distribution. In the present scope of study, wind data of various locations of India have been fitted by Weibull model. Wind speed data are initially sampled in knot by Indian Meteorological Department and later converted into integer km/h before supplying them to the end user. Due to this conversion, wind speed data cannot be properly fitted by Weibull distribution and in this regard, the choice of appropriate class width becomes very much important. Without the choice of appropriate class width, estimated Weibull parameters become biased which would yield incorrect estimation of expected number of hours in critical wind speed ranges as well as wind energy potential. After taking appropriate class width of 4 km/h, it has been found that Weibull model is an adequate model to describe wind speed distributions of India. Weibull model has also been compared with other models such as Gamma and inverse Weibull distributions to establish its suitability than the others. In this study, the values of Weibull shape parameters vary from 1.3 to 2.3, whereas the values of scale parameters vary from 1.4 m/s to 6.5 m/s. The validity of Weibull model is also verified with a target confidence interval of 90%. The uncertainties involved in the estimation of available wind energy potential as well as the expected number of hours per year in critical wind speed ranges have also been considered due to random variation of wind climate in each year.