Session: NDE-03-01 NDE Reliability-Modeling and Experimental Analysis

Paper Number: 122487

122487 - Enhancement and Optimization of Crack Signal Processing in Alternating Current Electromagnetic Non-Destructive Testing 

Metal structural elements frequently exhibit surface crack defects due to their operational conditions characterized by high strength, susceptibility to corrosion, and mechanical wear. To facilitate reliable forewarning and maintenance, crucial for averting safety incidents during equipment service, an accurate and stable inspection approach is indispensable. Currently, human intervention primarily dominates defect inspection, which results in restricted techniques and reduced efficiency. This paper concentrates on optimizing characteristic signals using the principles of Alternating Current Field Measurement (ACFM). Initially, a finite element analysis technique is applied to simulate and compute five parameters: excitation frequency, coil turns, excitation voltage, coil wire diameter, and lift-off height. The aim is to determine electromagnetic parameters that are well-suited for the inspection environment. Subsequently, for heightened inspection accuracy, we introduce a targeted curve-fitting method based on simulation outcomes. Specific methods are designed to address characteristic feature identification concerning crack length and depth, which includes a time-difference method and a binary-double calculation method. Furthermore, we delve into the inherent patterns between butterfly maps and crack profiles. This exploration leads to a novel crack profile prediction method grounded in butterfly map characteristics, thereby enhancing the array of crack risk assessment methodologies. Finally, we establish an experimental platform to examine various lengths and depths of groove-shaped cracks in Q235 material. This experimental setup also concurrently reconstructs profiles of cracks featuring identical characteristic dimensions. Our experimental results indicate that characteristic inspection errors for crack length and depth amount to 5.98% and 3.86%, respectively. Moreover, the inspection errors for profile attributes, based on butterfly map information, linked to crack length and depth, stand at 6.50% and 7.89%, respectively. Our experimental outcomes underscore the rationality of our characteristic signal processing and optimization methods.

Presenting Author: Yong Li China University of Petroleum, Beijing

Presenting Author Biography: Li Yong, Ph.D., Chinese, from China University of Petroleum (Beijing), mainly engaged in research on non-destructive testing technology.

Authors:

Yong Li China University of Petroleum, Beijing
Shaohua Dong China University of Petroleum, Beijing
Luming Wang China University of Petroleum, Beijing
Guanyi Liu China University of Petroleum, Beijing