An Improved Ultrasonic Imaging Method for Austenitic Welds Based on Grain Orientation Distribution Inversion Algorithm

Abstract

Traditional ultrasonic imaging inspection does not perform well for austenitic welds, in which the defect locating accuracy is insufficient due to the distortion of propagation paths caused by the weld’s anisotropy and inhomogeneity. The distribution of grain orientation (DGO) contributes to the velocity distribution in austenitic welds and influences the ultrasonic propagation paths. An improved ultrasonic imaging method for austenitic welds based on a DGO inversion algorithm is proposed in this paper. A DGO model is established using an iterative disturbance method of local parameters. The model is optimized by an evolution strategy based on the model error evaluated by propagation times of known reflectors. Inspection experiments were carried out, in which a reference specimen was used to acquire the DGO inversion result and the inspection images of a target specimen were generated based on the optimized DGO model. Experimental results have proved that the improved ultrasonic imaging method can obviously increase the defect locating accuracy and effectively improve the imaging performance for austenitic welds.