Discrete Wavelet Transform based Denoising of TOFD Signals of Austenitic Stainless Steel Weld at Elevated Temperature

摘要

Time of flight Diffraction technique is an advanced ultrasonic NDE methods, adopted in weld integrity testing. During the Time of flight Diffraction (TOFD) inspection of the stainless steel welds at high temperature, grain scattering noise is developed. In Time of flight Diffraction testing, the quality of the signal plays a dominant role in characterizing the defects. Hence, signal denoising is an essential prerequisite for the successful application of Time of flight Diffraction testing. In this work, one austenitic stainless steel weldment was artificially induced with slag defect. At high temperatures, the Time of flight Diffraction testing has been conducted on the weld piece and the resultant signals are applied over the proposed algorithms. Various combinations of wavelets, decomposition levels with different thresholding levels are applied to select an optimum denoising method. The evaluation of wavelet based denoising is achieved by calculating the Signal to Noise Ratio (SNR). Results show that the noises can be suppressed well and Signal to Noise Ratio is improved.