Acoustic emission during martensitic transformation and welding.

摘要

The aim of this research was to study and analyze acoustic emission (AE) signals generated during martensitic transformation and to develop techniques to use AE to monitor welding so that measures can be taken to prevent martensite-induced cracking.;An earlier mathematical relation of acoustic emission during athermal martensite formation to the associated Gibbs free energy changes was extended. The new form was found to adequately describe AE measurement made on several steels. The intensity of the AE signals is proportional to the temperature derivative of the fraction of martensite transformed, the cooling rate and the specimen volume. It is also a function of the carbon content of the steel. Values of the martensite-start temperature determined from the measurements of AE intensity agreed well with those in the literature.;To incorporate the dynamic characteristics of the process in the analysis, a model was further developed for the dynamic displacement (AE signal) from the transformation strains and the growth process of martensitic transformation in an elastic half-space, using Green's function. The AE signal amplitude as found to be inversely proportional to the distance between the martensite source and the sensor, and also the duration of formation of a martensite plate. It also depends on the orientation of the martensite plate. The AE signal frequency bandwidth increases as the duration of plate formation decreases.;Finally, experiments were done to monitor martensitic transformation signals generated during welding of 4340 steel and classification of the AE signal generated by martensite formation from signals associated with porous and normal welds. A frequency-based pattern recognition technique using linear discriminant functions was implemented with successful classification. Using a binary decision strategy and independent data testing, classification rates of 100% and 99.1% between normal welding signals and signals from martensite formation/porous weld, and 96.4% and 76.0% between martensite formation and porous weld signals, respectively, were achieved using the eight best features.