Pumped Storage Power Station Site Selection in Abandoned Mines Using XGBoost-SHAP and AHP

作者：W. Shen, Xiaogang Li, Jiaxi Zhang · 发表于：Advances in transdisciplinary engineering · 年份：2025 · DOI：10.3233/atde250818 · 被引用次数：1 · 研究领域：Advanced Algorithms and Applications、Geoscience and Mining Technology、Belt Conveyor Systems Engineering

Against the backdrop of synergistic development between mine ecological restoration and new energy, the importance of site selection for abandoned mine pit transformation into pumped storage power stations (PSPS) has become prominent. At the same time, traditional expert-driven decision-making approaches exhibit limitations. This study focuses on PSPS site selection in Fushun mining areas, conducting an in-depth analysis of 127 successful global cases to identify critical factors influencing site selection precisely. By employing the XGBoost-SHAP integrated model, this research quantifies the contributions of these factors and reveals their priority ranking. Results indicate that slope (SLO) and road distance (RD) are the two most dominant determinants. Through single-factor sensitivity analysis, SHAP dependence plots further identify optimal ranges for key parameters. Moreover, the study innovatively integrates machine learning with multi-criteria decision-making by converting XGBoost-derived feature weights into standardized judgment matrices for the Analytic Hierarchy Process (AHP), thereby constructing a hybrid intelligence evaluation framework. Validation demonstrates that the western Fushun mining area exhibits superior suitability for PSPS development. The proposed methodology successfully overcomes the subjectivity and inflexibility inherent in traditional expert-reliant models, establishes a systematic and verifiable scientific decision-making paradigm, and resolves ...