A Magnetic Nondestructive Evaluation Method to Simultaneously Determine Chemical Composition and Heat Treatment Characteristics of Plain Carbon Steels: A Statistical Modeling Approach

摘要

In this paper, combination of multiple regression method, logit model and discriminant analysis is proposed to extract underlying relationships among metallurgical characteristics of steel samples with different carbon contents subjected to various heat-treating cycles and magnetic properties. To investigate the ability of the proposed nondestructive evaluation (NDE) method to determine microstructural features of plain carbon steels, five groups of samples (with different carbon contents from 0.22 to 0.71 wt.%) were heat-treated in various cycles (annealing, normalizing (rapid and slow cooling in air), quenching and quench/tempering in the range of 200-600 degrees C). As a nondestructive characterization method, magnetic hysteresis loop technique was applied to all samples and four magnetic parameters including maximum flux density, coercivity, hysteresis loss and maximum differential permeability have been measured. The correlations among these parameters and phase fraction (for equilibrium heat-treated samples) and tempering temperature (for non-equilibrium heat-treated samples) as well as carbon content of all samples have been evaluated. The results show that the combination of statistical modeling and magnetic hysteresis loop technique enables the proposed NDE system to determine carbon content and phase fraction of equilibrium microstructures or tempering temperature of non-equilibrium ones of a plain carbon steel with unknown carbon content and heat treatment background.