Predictive sustainability analysis of installed commercial solar energy parks: a temporal and spatial machine learning assessment

Abstract

The pre-installation assessment criteria for solar energy parks have been simulated through a variety of machine learning algorithms, with predictors categorized into three different climatic time frames (present, 2050, and 2070 bio-climatic time frames) and four distinct Socio-Economic Emission Scenarios, namely, RCPs 2.6, 4.5, 6.0, and 8.5, which represent projections for future levels of radiative forcing and greenhouse gas emissions W/m². A promising new location identification was speedily achieved through the development of an ensemble distribution model using a machine learning algorithm. The total capacity (in MW) and covered area of 78 different solar parks across India from various agro-climatic zones were examined (Sq. KM). Predictions about the future viability of existing solar parks are made in this study, and the best places for new ones are suggested. It was found that 2.08% of India’s total land area, or 68,369.69 sq. km, is optimum for solar parks, given the existing climatic, solar, and land cover characteristics. Across the board, the optimal locations were increased for RCPs 2.6 (3.87% of India’s total land area), 4.5 (2.72%), and 8.5 (4.47%) by 2050. Upward trends were similarly observed in the RCP 2.6 (3.40) and RCP 6.0 (2.27%) for 2070. Solar parks are considered ideal in the western half of the country, while more moderate locations are expected to emerge in the west, south-west, and central India.