Tailings Pond Wetting Line Prediction Learning Model Based on WSTD-Informer

Accurately predicting the seepage line changes of tailings ponds is crucial for the stable and safe operation of these facilities. Still, the inherent non-stationarity and randomness of the seepage line pose challenges to the prediction task. Therefore, this paper proposes a seepage line prediction method that combines WSTD and Informer. First, the original time series is preliminarily denoised using wavelet soft thresholding; then the Informer network model is constructed to predict the seepage line changes. Multiple models are compared and analyzed to determine the advantages of the proposed hybrid model. The results demonstrate that the WSTD-based Informer model proposed in this study achieves high prediction accuracy and generalization ability.