This work made it possible to compare the experimental and simulated radiographs obtained with CIVA of step wedges in Dural and ferritic steel, with two types of detectors.
In general, these comparisons highlight CIVA's ability to simulate X-rays in the tested configurations, based on the various criteria presented in this report.
Considering the profiles lines along the holes and steps, 90% of the pixels have relative deviations between simulation (MC) and experiment of less than 9% for full-wedge for both detectors (67% without taking into account the diffusion).
Some significant discrepancies have been identified, and their causes analyzed:
- The dose deposited on the detector, as simulated by CIVA, is inferior to the one obtained experimentally. The magnitude of the emission spectra of the reflected tubes proposed by CIVA does not correspond to that of the radiation emitted by the experimental X-ray tubes. It is therefore necessary to systematically modify the exposure time or the intensity of the simulated X-ray tube to adjust the simulated dose to that measured experimentally.
- Over-evaluation of the noise level for both models: this estimate is based on a DQE value that is supplied by the user as the input parameter of the simulation. There is no provision in CIVA for providing a spatial frequency-dependent DQE curve, and the modification of the FTM by the user does not appear to have an impact on the DQE. An evolution of the noise generation model implemented in CIVA to remove these limitations should therefore make it possible to improve this point.
Back to Analysis