A novel fuzzy-multi-criteria-GIS-machine learning approach for onshore wind power plant site selection

Selecting proper sites for onshore wind power plants (OWPPs) is a challenging task due to the inherent uncertainty in the decision-making process. This paper proposes a novel hybrid methodology that combines fuzzy logic, multi-criteria decision making (MCDM), and machine learning (ML) techniques, based on geographic information system (GIS). First, we standardized seven criteria for selecting OWPP sites using fuzzy logic membership functions. Then, we employed the spherical fuzzy-analytical hierarchy process (SF-AHP) to calculate the weights of the criteria. Subsequently, the weighted aggregated sum product assessment (WASPAS) method was utilized to generate a suitability map for OWPPs in Morocco. The final suitability map's consistency was validated by comparison with existing or planned wind farms and using two ML models namely, XGBoost and multiple linear regression (MLR). SF-AHP produced a balanced weight distribution, with wind speed (23.40%) being the most significant factor, and soil texture (9.4%) being the least significant. The WASPAS method clearly identified the highly suitable areas for OWPPs in Morocco. XGBoost outperformed MLR and provided excellent testing results (R2 = 0.9982, RMSE = 0.0043, MAE = 0.0016, and MAPE = 0.5486) demonstrating a good relationship between the seven criteria and the generated wind farm suitability map. The framework validity and stability were further confirmed through comparative and sensitivity analyses.