Development of the Wind Map for Small-Scale Wind Energy Conversion System Using Advanced AI-Based Approach in the Indian State of Rajasthan

Abstract

Wind energy is a promising renewable resource, but accurate wind speed estimation at different heights is vital for optimizing wind energy generation and resource assessment. Conventional techniques, like remote sensing and weather stations, frequently struggle with missing data and spatial constraints. Using historical meteorological information, this paper proposes a random forest-based AI method for producing wind speed information at elevations between 2 and 50 meters for a small-scale wind energy conversion system. To improve estimates of wind resources, machine learning models are trained using wind speed measurements at 3 m, 10 m, and 50 m for projecting wind speeds at unmeasured elevations. Data preprocessing, feature engineering, model selection, and validation are all part of the methodology. The Mean Squared Error (MSE) is used to evaluate model performance, with results demonstrating outstanding accuracy (e.g., MSE = 0.125 m2/s2 at 30 m). The findings show that AI-generated wind speed data greatly improves the accuracy of energy predictions, supporting grid integration and the best possible turbine placement. The suggested method can be extended for wider geographic applications and provides an affordable substitute for conventional anemometer-based observations. After generating the predicted wind speed data, the average wind speed at each height is calculated and visualized on a graphical map based on the respective location/coordinates.