Near-Infrared Spectroscopy and Principal Components Regression for the Quality Analysis of Glass/Epoxy Prepreg

A near-infrared (NIR) spectroscopic technique associated with a principal components regression (PCR) algorithm has been applied for rapid and nondestructive quality determination of glass/epoxy prepreg. Abundant information of prepreg related with the resin and volatile was observed in the spectra. The raw spectra and first derivative spectra over different spectral ranges were analyzed to obtain the optimum models. The models had correlation coefficients (R) of 0.950 for the volatile content and 0.981 for the resin content. The root mean square errors of prediction (RMSEP) for the volatile content and the resin content were 0.316% and 0.666%, respectively. The NIR spectroscopy method can be used to determine the resin content and the volatile content simultaneously, and the prepreg could be analyzed once within 30 s without sample destruction.