A Bayesian Optimized Variational Mode Decomposition-Based Denoising Method for Measurement While Drilling Signal of Down-the-Hole Drilling

Abstract

Measurement while drilling (MWD) emerges as a reliable technique for assessing rock mass properties. However, the measured MWD signals are often contaminated with noise, leading to distorted signals. To address this issue, this paper proposes a denoising method that utilizes variational mode decomposition (VMD) and wavelet soft thresholding (WST). The proposed method employs Bayesian optimization to adaptively determine the optimal VMD parameters. The decomposed modes are then further denoised using WST with appropriate thresholding and optimal wavelet decomposition levels. To validate the effectiveness of the proposed method, four synthetic signals with varying input signal-to-noise ratios (SNR) and two sets of measured MWD signals were evaluated. The results demonstrate the surpassing denoising efficiency of the proposed technique in terms of higher output SNR, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">R</i> <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> and lower RMSE values. Moreover, the denoised MWD signals show promising results, with reduced amplitude and volatility while preserving the characteristics of the original signals.