Bettayeb, Fairouz; Research center on industrial technologies; Algeria
Bettayeb, F.; Crti, Research Center on Industrial Technologies; Algeria
Session: NDT Reliablity 2
Time: 11:50 - 12:10
Welded components remain difficult to reliably and effectively examine cause of complex solidification process. Material grains orientations in welds generate ultrasonic beam divergence and splitting due to the re-melting process after each welding pass. The sizable size of the anisotropic grains contrasted with wavelength’s pulse, affects coarsely ultrasound propagation through severe attenuation and changes in velocity and energy scattering.
Basically ultrasonic signal flaw visibility is corrupted by many noises as electric, pulse, ringing, and structure or spurious signals. Ultrasonic noise is usually assumed to be Gauss random variable with zero averaging and a limited band power spectrum function. Various signal processing techniques were investigated to extract useful data and interpret waveform data for diagnostic and predictive purpose. However the complexity of the model order estimation carries on complicated modeling. Wavelet based auto regressive parametric model is a successful processing technique for natural signals, able to withdraw the non stable characteristics of data. In this paper, wavelet multi-scale analysis was investigated with a predictive approach, as a powerful computational tool for noise discrimination and data extraction. The analysis based correlations, residuals and interpolations calculations seems to be a well adapted signal analysis tool for viewing material micro structural dimension scales. This research shows a challenging 3D interface between material properties, calculations and ultrasonic wave propagation modeling, and indicates a linear signal energy distribution at micro structural levels. Which could point to potential outcome of ultrasonic wave signature of micro-structures at different energy scales related to matter phases. The multi-polynomial processing interpolations expose an attractor that should involve data modeling through chaos theory for a predictive material behavior purpose.