Application Studies On Rapid Analysis Of Composite Modified Double Base Propellant By NIR Spectroscopy

Traditional methods for the main quality indexes of composite modified double base propellant(CMDB) determination were not only can’t apply to on-line analysis but also time-consuming. NIR has been widely used in areas such as agriculture, pharmacy, petrochemical industry, tobacco and food for its rapid, efficient, non-destructive, non-polluting, accurate and on-line analysis. The precise quantification of the main components of the modified double base propellant is critical to ensure the quality of the propellant, therefore, it is significant for the determination of the main components of the modified double base propellant by NIR.1、To Compare with commonly used methods of MLR, PCR and PLS, PLS is an optimal modeling method. A NIR method was investigated to determine the primary content of A and B in CMDB. The influence of Smoothing, Normalization, D1, D2, MSC and SNV different pretreatment results were compared, the results indicated that a more accurate result can be obtained with Normalization for both component A and B. The calibration model of component A was established and optimized, which the root mean square error of cross-validation (RMSECV), and the Root Mean Squared Error of Predication(RMSEP), and the relative mistake of predictive value and actual value in validation sets were0.2028,0.2822and0.0509, respectively; RMSECV and RMSEP of the calibration model of component B were0.1095and0.1286, respectively; and the relative mistake of predictive value and actual value in validation sets was0.0183. NIR has been proved to be a accurate and efficient quantitative method for the component A and B in two-components system. In addition, it is said the foundation for more complex systems research.2、In this part, the three main raw materials (A, B and C) of modified double base propellant were determined by NIR. The influence of Smoothing, Normalization, D1, D2, MSC and SNV different pretreatment results were compared, the results indicated that a more accurate result can be obtained with Normalization for component A and C, and SNV was more suitable for B. The calibration model of component A was established and optimized, which RMSECV, RMSEP and the relative mistake of predictive value and actual value in validation sets were0.1191,0.1484and0.0380, respectively; The calibration model of component B was built and optimized, the RMSECV and RMSEP were0.0838,0.0554, respectively, and the relative mistake of predictive value and actual value in validation sets was0.0085; The RMSECV and RMSEP of the model for component C were0.1690and0.1029, respectively, and the relative mistake was0.0420. Therefore, the main components in three-components system can be rapid, accurate predicted by NIR, and it is significant for online rapid determination of the main components of the modified double base propellant.