Establishment of a NIR-based methodology for tracking the blend homogeneity of HTPB propellant slurry in the mixing process

The hydroxyl-terminated polybutadiene (HTPB) propellants with high level of solid loadings from 80 wt% to 90 wt% consist of aluminum (Al) powder, ammonium perchlorate (AP) and HTPB. The Al/AP/HTPB adhesive system full of solid grains appears high viscosity against flow. Therefore, the mixing is a crucial procedure in the production as it directly affects the structural integrity of the finished product. This work focused on the feasibility of tracking the blend homogeneity of Al/AP/HTPB adhesive system in the mixing process through using the near-infrared (NIR) spectroscopy and orthogonal partial least squares discrimination analysis (OPLS-DA). The OPLS-DA classification models were created by variable selection, spectral pretreatment and latent variables (LVs) optimization. It had been demonstrated that the developed models presented an excellent predictability with the root mean square error of cross-validation (RMSECV) for slurries in Ⅰ, Ⅱ groups of 0.1261 and 0.0789, respectively. Meanwhile, the well-fitted models for slurries in Ⅰ, Ⅱ groups with the squared correlation coefficient (R2) of 0.806 and 0.980, exhibited separately an acceptable predictive capability with the predictive squared correlation coefficient (Q2) > 0.5. Furthermore, Euclidean distance and move block standard deviation (MBSD) as reference methods were used to validate the predictive performance of the developed models with respect to the blend homogeneity of HTPB propellant slurry. The experimental results showed that the terminal time for each batch of slurry reaching to ideal uniformity predicted by Euclidean distance/MBSD and OPLS-DA were both at 26–30 min. Therefore, it had been proved that the method we proposed was a potential tool to monitor the variation of the uniform state of HTPB propellant slurry in the mixing process.