Downhole Liquid Level Depth Detection Research Based On Acoustic Signal Blind Separation

In the process of oil exploitation, the well fluid level is a very important parameter,it has significant reference value for formulating reasonable oil well production plan. The general way is acoustic method to measure liquid level depth. Due to the oil collars in complex situations, making the sound waves affected by various factors in its propagation and reflection process, thus surface wave and the reflected wave is not easy to identify,so it is impossible to measure the oil level depthcorrectly.This article focuses on the reflected acoustic signal separation, identification process to detect well fluid levelaccurately and design oil level detection system.Analysis of the acoustic method to measure the level of the basic principles of movement, combining with the well site survey data to establish the acoustic signal propagation model of the oil collars. Digital signal processing, wavelet transform and single blind separation principle research and analysis, and simulation analysis wavelet transform and denoising effect overall on empirical mode decomposition of single channel blind separation.From raising the viewpoint of oil well liquid level measurement precision and speed, using MFC visual interface design the well liquid level measuring system, which has drawn waveform zoom, drag and drop capabilities, can be very intuitive analysis of the reflected acoustic signals, and the signal analysis and processing, thus calculate the liquid level depth of the oil well correctly.For the reflected wave signal is not easy to identify,in this paper,based on the overall experience mode decomposition method of single channel blind source separation separation reflected sound wave signal and noise signal; wavelet transform has multi-resolution characteristics, so it is used to test the level of the reflected wave position; short average magnitude difference function to identify the characteristics of having a period for section hoop reflected wave cycle recognition, resulting acoustic wave propagation velocity in the well, combined with the level of the reflected wave detected position to detect the value of the well depth level.Experimental results show that the overall empirical mode decomposition of single channel blind separation in complex noise, which can effectively separate the reflected acoustic signals to improve the well fluid level measurement accuracy.