A Novel Groundwater Burial Depth Prediction Model Based on Two-Stage Modal Decomposition and Deep Learning

Abstract

The variability of groundwater burial depths is critical to regional water management. In order to reduce the impact of high-frequency eigenmodal functions (IMF) generated by complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) on the prediction results, variational modal decomposition (VMD) is performed on the high frequency IMF components after the primary modal decomposition. A convolutional neural network-gated recurrent unit prediction model (CNN-GRU) is proposed to address the shortcomings of traditional machine learning which cannot handle correlation information and temporal correlation between time series. The CNN-GRU model can extract the implicit features of the coupling relationship between groundwater burial depth and time series and further predict the groundwater burial depth time series. By comparing the prediction results with GRU, CEEMDAN-GRU, and CEEMDAN-CNN-GRU models, we found that the CEEMDAN-VMD-CNN-GRU prediction model outperformed the other prediction models, with a prediction accuracy of 94.29%, good prediction results, and high model confidence.

Evaluate effect of 126 pre-processing methods on various artificial intelligence models accuracy versus normal mode to predict groundwater level (case study: Hamedan-Bahar Plain, Iran)

“…Decomposition methods (EMD and VMD) and wavelet transform pre-processing techniques generally showed more effective results in GWL estimation than Statistical methods ( Table 4 , Table 5 ). The findings of this study are consistent with Wu et al [ 75 ], Lin et al [ 76 ], Zhang and Zheng [ 48 ] and Panahi et al [ 77 ]. VMD algorithm improved the performance of the ELM algorithm in predicting GWL in northwestern China [ 78 ].…”

“…The analysis results revealed that the VMD technique improved the performance of the DL technique. Zhang and Zheng [ 48 ] CEEMDAN, integrated the VMD data decomposition technique into a convolutional neural network gated repetitive unit prediction model (CNN-GRU). As a result, it was revealed that CEEMDAN and VMD models improved the prediction accuracy of the CNN-GRU model.…”

“…Further, Wunsch et al [ 47 ] used ANN, LSTM, convolutional neural networks (CNN), deep learning (DL) and non-linear autoregressive networks (NARX) for prediction of groundwater level and found that LSTM and CNN perform good with a larger dataset, where DL perform better with small dataset. Zhang and Zheng [ 48 ] predicted the depth to groundwater using coupled model of empirical mode decomposition (EMD) and LSTM and results showed that the prediction rate, maximum and minimum error is 100%, 5.00% and 0.07% respectively. The wavelet transform has also been studied for groundwater levels in various research studies [ [49] , [50] , [51] , [52] ].…”

Denoising of blasting vibration signals based on CEEMDAN-ICA algorithm

“…When the prior signal is unknown, independent component analysis (ICA) 18 can self-adaptively decompose the independent components and further process the noise signal in the signal. ICA is extensively utilized in image processing, signal processing, and other related areas 19 – 21 .…”

The Application of Machine Learning: Controlling the Preparation of Environmental Materials and Carbon Neutrality

“…boosting, and bagging [33][34][35][36]. Guo et al reported the adsorption and desorption of heavy metals on the interface of sediment by using the Bayes algorithm [37].…”