Comparison of Wind Speed Probability Distribution Models for Accurate Evaluation of Wind Energy Potential: A Case Study from Kerala, India

Assessing the spatial and temporal distribution of wind speed is essential for the proper design and performance monitoring of wind farms. This study evaluates the efficacy of 8 novel probability distribution models (Ram Awadh, Prakaamy, Pranav, Ishita, Sujatha, Akash, Shukla and Weighted Pranav) for assessing the wind speed distribution using 39 years' of data from 6 stations in Kerala, India having varying topographic features. The two regression-based test statistics (AIC—Akaike’s Information Criteria and BIC—Bayesian Information Criteria), A-D test statistics and histogram analysis are performed for evaluating the goodness of fit. Based on the regression and A-D test analysis, the Weighted Pranav distribution has shown the best fit for stations having low and middle altitudes while the Shukla distribution has proven to be the best fit based on A-D test statistics. The histogram-based comparison of the trend curves showed that the geometric features of the distribution are also significant to confirm the overall fitness of a model for the selected domain of the parameter. We suggest that Weighted Pranav, Shukla, Ram Awadh and Prakaamy distribution are preferred for representing the wind speed variations based on low-altitude measurements in locations having significant topographic variations.