Crack initiation detection in crankshaft ductile cast iron based on information entropy of acoustic emission signals under tensile loading

Abstract

This study investigated the use of acoustic emission (AE) signals for the crack initiation detection in a ductile cast iron, EN-GJS700-2. Uniform uniaxial tensile tests were performed on specimens, extracted from the web of an automobile crankshaft. Experiments were done under displacement-controlled conditions, as instructed in ISO-6892. The crack initiation detection was performed by two methods: the analysis of AE features and the analysis of the AE information entropy. The fracture surface of specimens was also examined and analyzed by the scanning electron microscopy (SEM). The results showed that while the crack initiation could be detected by analyzing AE features, this approach involved some uncertainties. However, the logarithmic cumulative Shannon entropy and the logarithmic cumulative Kullback-Leibler relative entropy curves could be used for the sufficiently accurate crack initiation detection without this problem. Therefore, the analysis of the AE information entropy could serve as a highly practical approach for real-time condition monitoring.