Research on Wavelet Threshold Denoising Method of Ultrasonic Echo Signal in Ultrasonic Drilling Fluid Leak Detection System

In view of the self-developed ultrasonic testing instrument for drilling fluid leakage layer, the echo signal of the leakage position is affected by the scattering attenuation of suspended particles in the wellbore, which leads to the inaccurate judgment of the drilling fluid leakage layer. A detection method combining the ultrasonic frequency difference method and improved wavelet threshold function is proposed. Because the ultrasonic echo signal has the characteristics of non-stationarity, suddenness, focusing and compressibility, four kinds of wavelets are selected for multi-scale decomposition and reconstruction experiments. The denoising effect of the improved threshold function and four conventional threshold functions on the ultrasonic echo signal is analyzed and compared. The analysis results show that the improved threshold function can avoid the Gibbs effect and ‘excessive compression’, while fully retaining the singularity of the signal mutation and missing part, and has stronger flexibility. Using the improved threshold function and coif3 wavelet for 8-layer decomposition denoising, the signal-to-noise ratio is increased by 25.13 dB, the root mean square error is reduced by 0.63, the correlation coefficient is reduced by 6.29, and the smoothness is reduced to 2.12, achieving the best denoising effect. Therefore, the optimal denoising scheme is the improved threshold function using coif3 wavelet 8-layer decomposition.