A Tailings Dam Long-Term Deformation Prediction Method Based on Empirical Mode Decomposition and LSTM Model Combined with Attention Mechanism

Tailings dams are constructed as storage for ore waste, serving industrial waste piles and drainage. The dam is negatively affected by rainfall, infiltration lines its own gravity, which can cause instability to gradually increase, leading deformation. To predict the irregular changes of tailings deformation, empirical mode decomposition (EMD) applied deformation data obtain trend periodic components. attention mechanism used assign different weights input variables overcome limitation that long short-term memory (LSTM) model only generate fixed-length vectors. lagged autocorrelation coefficient each decomposed subregion solve lagging effect external factors on Finally, in multiple directions test generalization ability. proposed method effectively mitigate problems gradient disappearance explosion. results show that, compared with control EMD-LSTM, evaluation indexes RMSE MAE improve 23.66% 27.90%, respectively. also has a high prediction accuracy remaining dam, wide practical application provides new idea research.