Abstract
This work is focused on calculating entropy measures for signals in order to identify Portevin-Le Châtelier (PLC) effect types. The PLC effect is a phenomenon occurring in metals, in particular steel and aluminum alloys, within a certain range of strain rates and temperatures. It is characterized by serrations (repetitive changes from hardening to softening) visible in a load-displacement diagram and associated strain rate bands moving through a sample. Three main PLC types are distinguished: A, B and C. Type A occurs in low temperature and for high strain rate, strain rate bands then propagate continuously. Type B occurs for medium temperature and strain rate, the bands then have a hopping character. Type C occurs in high temperature and for low strain rate, the bands then nucleate in a random manner. The entropy analysis is used as a way to distinguish the types. The so-called Sample Entropy, Sample Entropy 2d and Multiscale Sample Entropy are measures utilized in signal analysis to look for patterns in data. Sample Entropy takes into consideration only force values which need to be sampled at equal intervals. Sample Entropy 2D, on the other hand, also accounts for the distances between points. Multiscale Sample Entropy extends the standard approach by analyzing the signal across multiple time scales. For computations, experimental results in the form of load-displacement diagrams for tensile tests performed on bone-shape samples are used. The experimental tests have been performed in room temperature for three strain rates. The band types are first identified based on DIC data by band movement observation. It is found that for a high strain rate we observe type A, for a medium strain rate first type A and then type B and for a low strain rate type C. The Sample Entropy and Sample Entropy 2d measures for type C are low and for type A are high. Different behavior of those two types is also visible for higher time scales. It is also found that to assess type B of PLC effect more experiments are needed.