摘要
为实时检测尿素法生产HMX过程中HMX的含量,以生产过程为研究对象,采用偏最小二乘法建立了近红外光谱技术实时检测HMX含量的定量校正模型。通过实验对定量校正模型进行了验证。结果表明,该模型可以用于HMX含量的实时检测。通过用间隔偏最小二乘法初选建模区间,将初选区间及其组合与不同光谱预处理方法优选组合建模,得到最佳建模区间及其匹配的光谱预处理方法。结果表明,近红外光谱技术和光纤探头测试方式可用于实时检测HMX合成过程中HMX的含量。本研究优选出的最佳建模区间为5 000~4 750cm-1,最佳预处理方法为(1st+平滑)预处理法。
For real-time detection of HMX content in the process of HMX produced by urea method,using the production process as the research object,aquantitative calibration model to detect HMX content in real time with a near infrared reflectance spectroscopy(NIRS)was established by the partial least squares method.The quantitative model validation shows that the model can be used for real-time detection of HMX content through selecting the modeling range primarily by an interval partial least squares method and optimizing the assembly modeling by primary interval and its combinations with different spectral pretreatment method.The best modeling interval and its matching spectrum pretreatment method were obtained through selecting the modeling range primarily by an interval partial least squares method and optimizing the assembly modeling by primary interval and its combinations with different spectral pretreatment method.The results indicate that the NIRS technology and optical fiber probe method can be used to detect HMX content in real-time in the synthesis process of HMX.The best modeling range selected in this study is 5 000-4 750cm-1 and the best pretreatment method is the pretreatment method of 1st+smooth.
出处
《火炸药学报》
EI
CAS
CSCD
北大核心
2014年第1期66-69,共4页
Chinese Journal of Explosives & Propellants
基金
火炸药基础科研计量项目(NO.J092011C002)
国家国防科技工业局"十一五"科研计划项目(NO.Z092009T002)
关键词
分析化学
近红外光谱
模型优选
实时检测
HMX
analytical chemistry
near infrared reflectance spectroscopy
model optimization
real-time detection
HMX
