Rapid Detection Technology Of Component Content And Particle Size Of Single-base Propellant

Single-base propellant is a soluble plastic propellant with nitrocellulose(NC)as the only energy component.In the production process,other additional components need to be added to meet its combustion performance,ballistic performance and stability,etc.Changes in the content of each component and particle size of single-base propellant have a great impact on its combustion performance,ballistic performance and other properties.Therefore,it is necessary to accurately detect content of each component of the single-base propellant.In order to solve the shortcomings of traditional detection methods for the content of single-base propellant components such as time-consuming detection and environmental pollution,this paper applied near-infrared spectroscopy(NIR)analysis technology to rapidly detect the content of single-base propellant components in the production process.In addition,in order to solve the disadvantages of time-consuming detection and complex operation in current detection of single-base propellant particle size,an imaging and measurement system based on a stereo microscope was used to quickly detect the size of the single-base propellant.Near-infrared spectroscopy analysis method was used to quickly detect the ethanol and water content in dehydrated NC.The modeling spectral intervals of ethanol and water models were determined by spectral analysis.According to the performance of models established by different preprocessing methods,the best spectral preprocessing method was determined.The partial least square regression(PLS)was used to establish ethanol and water models.The determination coefficient of cross-validation(R²_cv)of ethanol model is 0.9816,the root mean square error of the cross-validation(RMSECV)of ethanol model is 0.138,and R²_cv of water model is 0.9532,RMSECV of water model is 0.0526.The prediction performance of the models was validated.The determination coefficients of calibration(R_c²)and prediction(R_p²)of the ethanol model are 0.9954 and 0.9693,respectively.The root mean square errors of calibration(RMSEC)and prediction(RMSEP)are 0.0756 and 0.182,respectively.The relative predictive deviation(RPD)value is 5.72;the R_c² and R_p² of water model are 0.9257 and 0.9690,respectively,the RMSEC and RMSEP are 0.0709 and 0.0502,respectively,and RPD value is5.74.The average deviations of the near-infrared prediction values of ethanol and water are0.15%and 0.04%,respectively.The established models are stable and reliable.Near-infrared spectroscopy analysis was adopted to rapidly detect the content of stabilizer diphenylamine(DPA).The modeling spectral ranges of DPA were determined by spectral analysis.The best spectral preprocessing method was determined by comparing the performance of models established by different preprocessing methods.The PLS regression was used to establish the model.The R²_cv of the model is 0.9677 and the RMSECV of the model is 0.0195.In addition,the prediction performance of the model was evaluated.The R_c²and R_p² of DPA model are 0.9844 and 0.9668,respectively.The RMSEC and RMSEP are0.0155 and 0.0198,respectively.The RPD value is 5.5 and average deviation of the near-infrared value is 0.026%.The established model is stable and reliable.Near-infrared spectroscopy analysis was applied to quickly detect the internal and external volatiles of the single-base propellant in the pre-drying process.The modeling ranges of the internal and volatiles model were determined through spectral analysis.According to the performance of models established by different preprocessing methods,the best spectral preprocessing method was determined.The internal and external volatiles models were established using the PLS regression.The R²_cv of internal volatiles model is 0.9552,the RMSECV is 0.0232.Meanwhile,the R²_cv of external volatiles model is 0.9651,and the RMSECV is 0.117.The prediction performance of the models was confirmed by other samples.The determination coefficients of calibration(R_c²)and prediction(R_p²)of internal volatiles model are 0.9656 and 0.9766,respectively.The root mean square errors of calibration(RMSEC)and prediction(RMSEP)are 0.0219 and 0.0132,respectively,and the relative predictive deviation(RPD)value is 6.89.The R_c² and R_p² of external volatiles model are 0.9817 and0.9731,respectively.The RMSEC and RMSEP are 0.0936 and 0.0829,respectively,and RPD value is 6.35.The average deviations of the near-infrared predictive values of internal and external volatiles are 0.010%and 0.07%,respectively.The established models are stable and reliable.The system based on a stereo microscope was applied to quickly detect the size of single-base propellant particles.Firstly,perform image acquisition and object reconstruction on the pre-processed single-base propellant particles,and then the web size,aperture and particle length were marked in the measured image.After that,their size values can be automatically counted to obtain the average value as the detection result.For 4/7 graphite single-base propellant particles,the measured web size,aperture and particle length are 0.464mm,0.179mm and 2.675mm,respectively.For 5/7 graphite single-base propellant particles,the measured web size,aperture and particle length are 0.552mm,0.190mm and 2.905mm,respectively.By comparing with the industry standard of single-base propellant,it was determined that the selected batch of single-base propellant is qualified in size.Comparing the detection results of this method with the traditional ones,it can be determined that the deviations of the detection results are all low and the images collected by this system are clearer.In addition,this system has automatic data result statistics function.Therefore,it can quickly and accurately detect the size of single-base propellant particles.