11Application of Machine Learning Optimization Techniques in Wind Resource Assessment

Udhayakumar K.1* and Krishnamoorthy R.2

1Department of EEE, Anna University, Chennai, India

2Information Technology, TANGEDCO, Chennai, India


Wind energy is a preferred energy in electricity production because reducing fossil fuels-based generations reduces pollution in the atmosphere. In this study, the wind characteristics and wind potential assessment in Onshore, Offshore and Nearshore location of India, viz., Kayathar, TamilNadu; Gulf of Khambhat; and Jafrabad, Gujarat, were statistically analysed with wind distributions methods. Weibull, Rayleigh, Gamma, Nakagami, GEV, Lognormal, Inverse Gaussian, Rician and Birnbaum Sandras and Bimodal Weibull-Weibull Distribution were used for resource assessment. The wind distribution parameters estimated with parameter methods is compared with Machine Learning optimization techniques Moth Flame Optimization method (MFO) for accuracy. The results show that the Moth Flame Optimization performed well on parameter estimation. The wind speed distributions mixed Weibull, Nakagami and Rician performed well in calculating potential assessments. The bimodal mixed Weibull distribution performed better than other distributions. The offshore Gulf of Khambhat area, Gujarat, has steady wind speeds ranging from 7 m/s to 10 m/s with less Turbulence Intensity and highest Wind power density, 431 watts/m2. Wind power generation peaks during South-West Monsoon ...

Get Explainable Machine Learning Models and Architectures now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.