Prediction algorithm for non-destructive on-line evaluation of mechanical hardness of electrical steels

Mechanical hardness of electrical steels is generally measured destructively. A computerised system based on a generic prediction algorithm has been developed for non-destructive determination of mechanical hardness in real time. Discrete data-bases of magnetic and mechanical hardness characteristics are compiled then subjected to a grouping analysis and quantised with respect to the desired resolution of the hardness predictions. Combinations of the magnetic parameters are used to perform a synchronous field matching process between various discrete magnetic-hardness quantised data bases. The results of the field mapping operation allows a series of specific coincidence patterns to emerge. The final Vickers Pyramidal Numbers (VPN) have uncertainty levels of +/- 5 VPN. The deduction technique has the potential to predict mechanical hardness of other continuous strip materials.