摘要
为了解决复合含能材料中CL-20晶型定量分析问题,提出了一种基于X射线衍射(XRD)结合偏最小二乘(PLS)的定量分析方法;制备了不同晶型CL-20与DA基药混合的复合含能材料样品,采集了XRD谱图,并结合多种数据预处理(MSC、SNV、D1st、D2nd、WT)和变量选择方法(SIPLS、BIPLS、CARS)对模型性能进行优化。结果表明,MSC-CARS-PLS模型对ε-CL-20和CL-20总含量的预测性能最优,其决定系数(R_(p)^(2))分别为0.9862和0.9759,对应的均方根误差(RMSE_(p))分别为0.5901和0.5276;将该模型应用于实际生产的复合含能材料样品中,预测结果与参考值高度一致,验证了其在CL-20晶型定量分析中的准确性与实用性;说明此方法是一种快速、准确的CL-20晶型定量分析方法。
To address the challenge of quantitative analysis of CL-20 polymorphs in composite energetic materials,a quantitative analysis method based on X-ray diffraction(XRD)combined with partial least squares(PLS)was proposed.The composite energetic material samples containing different CL-20 polymorphs and diacetamide(DA)-based matrix were prepared.The XRD patterns were obtained,and the performance of models were optimized using various data preprocessing methods(MSC,SNV,D1st,D2nd,WT)and variable selection strategies(SIPLS,BIPLS,CARS).The results indicate that the MSC-CARS-PLS model exhibits the best predictive performance for bothε-CL-20 and the total CL-20 content,with coefficients of determination for prediction(R_(p)^(2))of 0.9862 and 0.9759,and corresponding root mean square errors of prediction(RMSE_(p))of 0.5901 and 0.5276,respectively.By applying the model in the composite energetic material samples from actual production,the predictions are highly consistent with reference values,verifying its accuracy and practicality for the quantitative analysis of CL-20 polymorphs.It is demonstrated that this method provides a rapid and accurate approach for the quantitative analysis of CL-20 polymorphs.
作者
周利园
王德祥
张天龙
刘国权
李华
汤宏胜
ZHOU Li-yuan;WANG De-xiang;ZHANG Tian-long;LIU Guo-quan;LI Hua;TANG Hong-sheng(School of Chemistry and Materials Science,Northwest University,Xi′an 710127,China;Xi′an Modern Chemistry Research Institute,Xi′an 710065,China)
出处
《火炸药学报》
北大核心
2025年第8期756-762,I0003,共8页
Chinese Journal of Explosives & Propellants
基金
国家自然科学基金面上项目(No.21873076)。
关键词
分析化学
偏最小二乘
PLS
X射线衍射
定量分析
CL-20
变量选择
analytical chemistry
partial least squares
X-ray diffraction
PLS
quantitative analysis
CL-20
variable selection
