尺寸
焊接
无损检测
射线探伤
超声波检测
覆盖
点(几何)
超声波传感器
计算机科学
数字射线照相术
工作流程
欠采样
仪表(计算机编程)
测量不确定度
射线照相术
结构工程
材料科学
近似误差
声学
各向同性
体素
标杆管理
统计能力
作者
Mariana Burrowes,Davi F. Oliveira,Cintia G. Ferreira,Olga Maria Oliveira de Araújo,Cesar Giron Camerini,Gabriela Ribeiro Pereira
摘要
ABSTRACT This study presents a microCT‐anchored evaluation of flaw‐sizing performance in CRA‐clad pipe welds inspected using digital radiography and three phased array ultrasonic testing (PAUT) procedures. Two full‐scale carbon steel pipes clad with Inconel 625 were machined with 23 volumetric EDM defects located in both weld overlay and Triple Point (TP) regions. High‐energy microCT scanning, performed at 130 μm isotropic voxel resolution, provided metrologically robust three‐dimensional reference measurements against which all non‐destructive testing (NDT) results were benchmarked. Quantitative comparison demonstrated that radiography achieved full flaw detectability and the lowest absolute and relative sizing errors relative to the CT ground truth. PAUT performance was strongly dependent on procedural configuration: Procedure 2 exhibited the smallest sizing deviations but detected fewer than half of the flaws, Procedure 3 provided improved detection but systematic oversizing—particularly for small flaws—and Procedure 1 achieved full detection but displayed the largest variability owing to the absence of proper calibration. Region‐specific analysis enabled by CT spatial localisation showed that all PAUT configurations exhibited increased error in the Triple Point (TP) region, whereas radiography remained region‐insensitive. Bland–Altman analyses further revealed modality‐dependent systematic biases and limits of agreement. The results demonstrate the utility of microCT as an independent dimensional reference for benchmarking NDT performance in complex welded and clad structures. The CT‐anchored workflow introduced here provides a generalisable methodology for quantifying detection–accuracy trade‐offs and assessing the influence of procedural variables on ultrasonic inspection reliability.
科研通智能强力驱动
Strongly Powered by AbleSci AI