Assessment of steel inclusions cleanliness by mixed NDT methods

The present project aims to combine new technologies for the evaluation of the macroscopic cleanliness of steel products: US&Radio, micro radiography, new ultrasound applications and data fusion technique developments yield very performing tools for the detection of internal defects on as cast samples, making it possible to understand their origin mechanisms, even in under skin zone. US modelling and 3D processing bring in significant progress. The development of data fusion technique will be of interest, in particular for automatic inspection, but the application to US&Radio did not succeeded. With a GMR-sensor it is possible to detect dynamically non-metallic inclusions (alumina) in a cold strip sample with a thickness of 0.8mm. For a sensor lift-off of 0.5mm and a strip speed of 0.85m/s, the detectability limit is reached. Numerical modelling showed that, the signal decreases only of 17 % when the speed increases up to 40m/s. It is also possible to use GMR to detect surface and subsurface defects. The comparison program of the different test techniques available showed a weak relation between the results. ReinHUS (MIDAS) gives an early assessment of the cleanliness but does not take into account later possible changes in the inclusion morphology. ODL (Lamb waves) offer an assessment at least in time to sort out hot strip coils but seems to be dependent to other parameters. IDD (magnetic flux leakage) has the highest sensitivity, but the result is available just too late.