A Fuzzy Error Characterization Approach for Crack Depth Profile Estimation in Metallic Structures Through ACFM Data

The estimation of fatigue crack depth profile is one of the most important issues in nondestructive evaluation techniques. In this regard, providing a method which is able to determine the error rate of the depth profile estimation will be significant. Although several methods have been proposed to estimate the depth profile with reasonable accuracy, these mentioned methods are unable to determine the error rate of estimation process. In this paper, to eliminate mentioned drawbacks, a first and second fuzzy approximation methodology is proposed to estimate the surface crack depth profile of an arbitrary crack in metallic structures based on alternating current field measurement signals. In the proposed method, it is feasible to achieve a definite error rate to determine the surface crack depth profile under evaluation. Next, to achieve the most reasonable accuracy and the least computational complexity simultaneously, a solution will be proposed for the method presented in this paper. Finally, the results of several common methods are compared with the method presented in this paper in analytical and experimental forms.

I. Introduction

One of the most major concerns in metallic structural elements is the initiation and growth of fatigue cracks. Fatigue cracking is caused by cyclic stresses, which are below the ultimate tensile stress or even the yield stress of the material [1]. Undetected active fatigue cracks in structural elements under service loads can result in unstable fracture, as well as they can cause final failure of the structure under consideration. Consequently, distinguishing the named cracks is one of the most fundamental sources in metallic structural elements [2], [3].