DOI: 10.11799/ce202512024
A mine water inflow prediction method based on adaptive secondary decomposition and BiLSTM-Transformer
HE Weisheng
Abstract：Accurate prediction of mine water inflow is crucial for the safe production and water hazard prevention in underground coal mines.Traditional prediction methods are limited by model assumptions and parameter sensitivity,making it difficult to meet engineering requirements for high prediction accuracy.Therefore,a new method for mine water inflow prediction is proposed based on improved complete ensemble empirical mode decomposition with adaptive noise(ICEEMDAN)and variational mode decomposition(VMD)secondary decomposition technology,combined with bidirectional long short-term memory(BiLSTM)and Transformer.Specifically,the original water inflow signal is decomposed into several intrinsic mode functions(IMFs)by ICEEMDAN at first.Then,the sparrow search algorithm(SSA)is introduced to optimize the key parameters of variational mode decomposition(VMD),and secondary decomposition is performed on the signal to achieve multi-scale feature extraction.Subsequently,BiLSTM captures short-term and long-term dependency features of the time-series data,while the Transformer's self-attention mechanism strengthens global feature correlation modeling.Experimental results show that this method significantly outperforms traditional prediction methods in both prediction accuracy and robustness,providing a new technical approach for mine water inflow monitoring and early warning.
Keywords：water inflow predictionsecondary decompositionBiLSTMsparrow search algorithmmulti-scale feature extractionICEEMDAN
Publication Date：2025-12-20
Online Publishing Date：2026-01-29(First online date of this platform, not the publication date of the document)
Pages：8( 186-193 )