An Industrial Gauge for Measuring The Phase Distribution of Galvanneal

Augmentation of the internal software of a commercial x-ray fluorescence gauge is shown to enable the instrument to extend its continuous on-line real-time measurements of a galvanneal coating's total elemental content to encompass similar measurements of the relative thickness of the coating's three principal metallurgical phases. The mathematical structure of this software augmentation is derived from the theory of neural networks. The empirical basis for the numerics embedded in the software's decision logic is presented. The performance of the augmented gauge is validated by comparing the gauge-implied real-time phase distribution with the phase distribution independently measured off-line on time-tagged samples drawn from the galvanneal production line where the measurement gauge had been installed. The performance validation is shown to demonstrate good agreement between the gauge and laboratory measurements and to suggest preferred approaches to be followed in future applications of the augmented gauge