Correlation analysis and comprehensive evaluation of dam safety monitoring at Silin hydropower station

Dam failures pose catastrophic risks to human life and property, necessitating robust safety monitoring systems for risk mitigation. However, the specific contributions of distinct monitoring modalities to dam safety remain inadequately characterized, particularly regarding their differential impacts on structural integrity assessment. This study investigates the correlation between diverse monitoring modalities and dam structural safety through a comprehensive analysis of the Silin Hydropower Station dam. We analyzed 324 datasets collected from nine types of monitoring sensors installed across 36 dam cross-sections. Statistical analyses including one-way ANOVA, cluster analysis, and principal component analysis (PCA) were employed to quantify the influence patterns of monitoring parameters. The safety impact levels of all 36 cross-sections were systematically ranked, establishing a prioritized reference framework to inform decision-making in dam safety management. Unlike conventional dam safety assessments that predominantly rely on subjective empirical judgments, this study introduces an objective methodology integrating principal component analysis (PCA) of heterogeneous monitoring data across multiple dam cross-sections. The analytical outcomes were systematically quantified, hierarchically ranked, and visualized through multidimensional mapping techniques. The results demonstrated that variations in fissure (X2), horizontal displacement (X3), tilt (X4), stress (X6), soil-displacement (X8), and denotes water-level (X9) exerted highly significant effects on dam safety (p < 0.001). The first two principal components cumulatively accounted for 74.1876% of the total variance, with eigenvalues reaching 6.6769. In the comprehensive evaluation, cross-section T4 (T4) obtained the maximum score (0.8500), while cross-section T35 (T35) showed the minimum score (0.0175). In conclusion, the analysis revealed that X9, X8, X2, X3, and X4 exerted significant impacts on dam safety, while cross-section T4 achieved the highest comprehensive evaluation score. This approach employs Principal Component Analysis (PCA) with integrated scoring to reduce multivariate dimensionality, enabling rapid identification of key monitoring sections critical to dam safety, and demonstrates broad applicability for dam safety monitoring.