B 型套筒角焊缝的高效率阵列超声检测仿真方法

The ultrasonic array detection technology can effectively detect and evaluate the internal defects of type-B sleeve fillet welds in high-grade pipelines, but the selection of ultrasonic array detection scheme and process design has a significant impact on the detection effect, which makes it quite necessary to carry out ultrasonic array simulation analysis for the design and optimization of detection scheme and process. In addition, the conventional finite element simulation analysis methods have the problems such as complicated modeling and excessive computational load. This paper establishes a high-efficiency ultrasonic array detection simulation algorithm based on the time-domain finite difference method and the computational unified device architecture (CUDA) concurrent computation method, and then combined with the OpenCASCADE 3D modeling and characterization method, simulates and analyzes the acoustic beam path, defect acoustic response, and multiple detection imaging modes in the ultrasonic array detection of fillet weld structures. The following results are obtained. First, compared with the COMSOL finite element simulation, the new method improves the computation efficiency by over 95% for the same simulation model. Second, the simulation results are consistent with the actual detection signals and can provide accurate imaging and characterization of different types of defects. Third, the vector coherence imaging method can significantly improve the detection imaging quality of defects in type-B sleeve fillet welds, with a signal-to-noise ratio of defect image increased by more than 19 dB. In conclusion, the proposed ultrasonic array detection simulation method has a high computation accuracy and efficiency in acoustic simulation, which is conducive to improving the evaluation effect of ultrasonic array detection of type-B sleeve fillet weld structures and ensures the welding quality of such structures, promoting defect detection of fillet welds.