Improving density logging resolution by VMD-CEEMDAN-ICWT method and its application in thin layer identification

Abstract

Data identification of density logging is a crucial foundation for logging interpretation. Because of device and other external variables, noise will inevitably be mixed in data acquisition process. Moreover, as the main technique for identifying thin reservoirs, density logging exhibits a low resolution. In this study, a new de-noising and distinguish thin-layer method is proposed, namely VMD-CEEMDAN-ICWT method. The VMD-CEEMDAN-ICWT method is designed to combine an adaptive Improved-Continuous-Wavelet-Transform (ICWT) structure, Variational-Mode-Decomposition (VMD) wavelet decomposition structure, and Complete-Ensemble-Empirical-Mode-Decomposition with Adaptive-Noise (CEEMDAN) wavelet reconstruction structure with global denoising characteristics. It integrates the superiorities of VMD, CEEMDAN and ICWT method. Ultimately, empirical-mode-decomposition (EMD), discrete-wavelet-transform (DWT), continuous-wavelet-transform (CWT) and the proposed method are utilized to analyze real data. The findings indicate that the reconstructed density log by VMD-CEEMDAN-ICWT method shows higher vertical resolution. The density log resolution about thin layer increases from 25 cm to 12 cm. It makes density logging curves more capable of identifying reservoirs, which provides high-quality data for further exploration and exploitation of oil and gas.