Shi Zhanqun, Xu Xiaoyu, Ma Jiaojiao, Zhen Dong, Zhang Hao
School of Mechanical Engineering, Hebei University of Technology, Tianjin 300130, China.
Sensors (Basel). 2018 Apr 2;18(4):1070. doi: 10.3390/s18041070.
Small cracks are common defects in steel and often lead to catastrophic accidents in industrial applications. Various nondestructive testing methods have been investigated for crack detection; however, most current methods focus on qualitative crack identification and image processing. In this study, eddy current pulsed thermography (ECPT) was applied for quantitative crack detection based on derivative analysis of temperature variation. The effects of the incentive parameters on the temperature variation were analyzed in the simulation study. The crack profile and position are identified in the thermal image based on the Canny edge detection algorithm. Then, one or more trajectories are determined through the crack profile in order to determine the crack boundary through its temperature distribution. The slope curve along the trajectory is obtained. Finally, quantitative analysis of the crack sizes was performed by analyzing the features of the slope curves. The experimental verification showed that the crack sizes could be quantitatively detected with errors of less than 1%. Therefore, the proposed ECPT method was demonstrated to be a feasible and effective nondestructive approach for quantitative crack detection.
小裂纹是钢材中常见的缺陷，在工业应用中常常会导致灾难性事故。人们已经研究了各种用于裂纹检测的无损检测方法；然而，目前大多数方法都集中在裂纹的定性识别和图像处理上。在本研究中，基于温度变化的导数分析，将脉冲涡流热成像（ECPT）应用于裂纹的定量检测。在模拟研究中分析了激励参数对温度变化的影响。基于Canny边缘检测算法在热图像中识别裂纹轮廓和位置。然后，通过裂纹轮廓确定一条或多条轨迹，以便通过其温度分布确定裂纹边界。得到沿轨迹的斜率曲线。最后，通过分析斜率曲线的特征对裂纹尺寸进行定量分析。实验验证表明，裂纹尺寸能够以小于1%的误差进行定量检测。因此，所提出的ECPT方法被证明是一种用于裂纹定量检测的可行且有效的无损检测方法。
