Advanced Noise Reduction Technology Helps Improve The Accuracy of Mining Subsidence Monitoring

In modern mining, accurate monitoring and analysis of mining subsidence are important links to ensure safe production. The noise problem in the time series of Global Navigation Satellite System (GNSS) technology, as the core means of mining subsidence monitoring, has always been a challenge for researchers. In order to improve monitoring accuracy, this paper proposes an innovative denoising method that combines the Hybrid Grey Wolf Particle Swarm Optimization (IPSOGWO) algorithm with Improved Adaptive Noise Complete Set Empirical Mode Decomposition (ICEEMDAN) combined with Wavelet Thresholding (WT), providing a new approach for solving noise problems in GNSS time series.

IPSOGWO&ICEEMDAN: Powerful combination

Firstly, the IPSOGWO algorithm optimizes the hyperparameters of the ICEEMDAN algorithm, making it more suitable for processing GNSS time series data. The ICEEMDAN algorithm decomposes time series data to extract a series of Intrinsic Mode Functions (IMFs), effectively separating noise and useful signals from the data. The introduction of IPSOGWO not only improves the decomposition accuracy of ICEEMDAN, but also greatly reduces the computational cost.

Multi scale permutation entropy and wavelet thresholding: fine processing

After extracting IMF components, this paper uses the multi-scale permutation entropy method to screen out IMF components containing noise, and applies wavelet thresholding to perform secondary processing on these noise components. The wavelet thresholding method can effectively suppress noise while preserving the main features of the signal, making the processed IMF components purer. Then, the denoised IMF component is recombined with the noise free IMF component to obtain the final denoised result.

Experimental verification: Excellent performance

To verify the effectiveness of the method, researchers conducted experiments using simulated signals and measured data from an automated monitoring station in a mining area. The results show that compared with traditional wavelet thresholding, complete ensemble empirical mode decomposition (CEEMD), and GWO-ICEEMDAN methods, the denoising method proposed in this paper exhibits superior performance. Specifically, this method not only significantly reduces noise in the time series, but also improves the fidelity of the signal, providing more reliable data support for subsequent settlement analysis of the working face.

Practical application: Improve monitoring accuracy

In the practical application of automated monitoring stations in mining areas, this new noise reduction method has demonstrated its unique advantages. By efficiently processing GNSS time series data, researchers can more accurately monitor the changing trends of mining subsidence in mining areas and promptly identify potential safety hazards. Especially in complex mining environments, the robustness and accuracy of this method are particularly important, as it can provide more scientific decision-making basis for mining managers.

The proposed joint wavelet thresholding method of IPSOGWO and ICEEMDAN provides an innovative solution to the noise problem in GNSS time series. With the continuous advancement and improvement of technology, this method is expected to be widely applied in more fields, such as earthquake monitoring, building deformation monitoring, etc., providing more accurate monitoring data for various industries. In the future, as more research deepens, we look forward to this technology further improving monitoring accuracy, ensuring production safety, and promoting the sustainable development of mining and other fields.