Rapid and nondestructive analysis of quality of prepreg cloth by near-infrared spectroscopy

Abstract A near-infrared spectroscopy analysis technique has been developed to measure the resin content, the volatile content and the pre-curing degree of resin during the manufacture of prepreg cloth by solution impregnation process. A principal components regression (PCR) algorithm was used to model the relation between the components content and the spectra. The calibration results had correlation coefficient of 0.9440 for the resin content (root mean square error of calibration, RMSEC = 0.490), of 0.9417 for the volatile content (RMSEC = 0.249) and of 0.9401 for the pre-curing degree (RMSEC = 0.186), respectively. The method was applied in the product line to analyze the quality of prepreg cloth directly, and the measurement process was accomplished in less than 2 min without sample destruction. The results showed that the maximal error were −0.70%, −0.546% and 0.378% for the resin content, the volatile content and the pre-curing degree, respectively. The study indicates that NIR spectroscopic method is a good alternative for the quality analysis of prepreg cloth.