Pumped storage power plant dam deformation monitoring and early warning system based on multi-source heterogeneous data

Maintaining the safety and integrity of PSPS dams is vital under the dynamic and cyclic loading from turbine reversal and storage cycles. Reservoir fluctuations can trigger deformation, settlement, or instability, especially in seismically active or geologically complex areas. To mitigate such risks, a total early warning system was developed for the Panglong PSPS Dam in Chongqing, China, using multi-source heterogeneous data for real-time deformation monitoring. The system integrates high-resolution space-based measurements (InSAR, GB-InSAR), GNSS and ground sensors, meteorological data, reservoir records, and geotechnical surveys. Each dataset undergoes preprocessing (phase unwrapping, co-registration, smoothing, alignment, normalization) to ensure quality and compatibility. Data are fused spatially and temporally to monitor key parameters such as displacement, rainfall, reservoir fluctuation, and slope conditions. Deviant deformation is identified via threshold-based alert logic, with colour-coded warnings (Green, Yellow, Red) issued according to multi-parameter thresholds. The system highlights high-risk zones such as the upstream slope, crest, and toe, enabling alerts before safety limits are reached. Results demonstrate improved predictive accuracy and timely decision-making during construction and initial operations. This integrated framework establishes a reliable foundation for smart dam monitoring in high-risk hydropower systems.