Results of the 2014 UT Modeling Benchmark Obtained with Models Implemented in CIVA

Sylvain Chatillon; Sebastien Robert; Philippe Bredif; Pierre Calmon; Guillaume Daniel; Francois Cartier

doi:10.1063/1.4914810

Abstract : The last decade has seen the emergence of new ultrasonic array techniques going beyond the simple application of suitable delays (phased array techniques) for focusing purposes. Amongst these techniques, the particular method combining the so-called FMC (Full Matrix Capture) acquisition scheme with the synthetic focusing algorithm denoted by TFM (Total Focusing Method) has become popular in the NDE community. The 2014 WFNDEC ultrasonic benchmark aims at providing FMC experimental data for evaluating the ability of models to predict images obtained by TFM algorithms (or equivalent ones). In this paper we describe the benchmark and report comparisons obtained with the CIVA simulation software. The simulations and measurements are carried out on two steel blocks, one in carbon steel and another in stainless steel. The reference probe is a 64 elements linear array, with .5mm element width and a gap of .1mm, working at 5 MHz. The benchmark problem consists in predicting images of vertical and tilted notches located on plane or inclined backwalls. The notches have different heights and different ligaments. The images can be obtained considering different paths (direct echoes or corner echoes). For each notch, the full matrix capture (FMC) have been recorded in one unique position with the probe positioned such that than angle between the probe axis and the notch direction corresponds to 45 degrees. The results are calibrated on the response of a 2mm side drilled hole. For each case, TFM images have been reconstructed for both experimental and simulated signals. The models used are those implemented in CIVA based on Kirchhoff approximation. Comparisons are reported and discussed.

Results of the 2014 UT Modeling Benchmark Obtained with Models Implemented in CIVA : Solution of the FMC-TFM Ultrasonic Benchmark Problem Using CIVA

Identifiers

Collections

Citation

Export

Metrics

107