Rapid and nondestructive prediction of component content is the key to improve industrial production efficiency. However, limited data sets also result in low generalization capabilities of the model, and it is time-consuming to obtain a large amount of content reference values and costly. Here, near infrared (NIR) spectroscopy technique combined with deep convolutional generated countermeasure network (DCGAN) was used to predict the trinitrotoluene (TNT) content of the melt -cast explosive. DCGAN was used to simultaneously extend its spectral data and content data. After several iterations, fake data were produced, which was very similar to the experimental data. The partial least squares (PLS) regression model was established and the performance was compared before and after data enhancement. The results showed that this method not only improved the performance of regression model, but also solved the problem of requiring large number of training data.

为实现凤香型原酒总酸和总酯含量的快速分析,及时为基酒入库分级提供数据参考.利用中红外光谱技术对原酒样品进行光谱扫描,并结合传统分析方法,最终建立了原酒总酸和总酯指标的快速分析模型.结果表明,所建模型的各线性关系较好,且模型内部参数RMSEC(均方差)和RMSEP(预测均方差)值均较小且相近,模型的质量较好;另取验证样品对模型进行外部验证,通过与传统分析方法的检测结果进行比较,结果表明,总酸和总酯指标模型预测的平均相对误差分别为2.1％和3.2％,同时两种方法的检测值之间不存在显著性差异(P>0.05),说明模型具有较好的预测能力,可用于实际生产分析检测.

In order to quickly and accurately detect and analyze the intermediates and impurities produced in the synthesis of Hexanitrohexaazaisowurtzitane （CL-20）， control the purity or quality of CL-20， and ensure its sensitivity and detonation performance， In this study， nuclear magnetic resonance （NMR） and ultra-high performance liquid chromatography-Quadrupole time-of-flight mass spectrometry （UHPLC-QTOF-MS） were used to rapidly and efficiently analyze the components and impurities in the typical synthesis process of CL-20.The results showed that the impurity in HBIW was 1， 3-dibenzyl imidazole， the impurity in tetraacetyl dibenzyl hexazepane （TADB） was low acylated triacetyl tribenzyl hexazepane （TATB）， and the impurity in tetraacetyl hexazepane （TAIW） was incomplete TADB. The impurities in CL-20 are not fully nitrated monoacetyl pentanitrohexazazine （MPIW） and diacetyl tetranitrohexazazine （DATN）. This work will provide technical support for the quality control of CL-20 without affectingits detonation performance.

Organoaluminum compounds have been employed as catalysts in many reactions. However, due to their high reactivity of the Al-C bond towards water or air, the accurate analysis and detection of organoaluminum compounds are still challenging. Here we demonstrate the possibility of rapidly and precisely detecting the purity and impurities of triethylaluminium (TEAL). An indirect GC-MS method via alcoholysis reaction using a designed alcoholysis reactor was adopted to detect and separate the main components and impurities of the synthesized TEAL. Previously, the nuclear magnetic resonance (NMR) and the inductively coupled plasma-mass spectrometer (ICP-MS) methods were also selected to detect the main compounds in the synthesized TEAL. The impurities in TEAL was identified as n-butyl aluminum, and the purity of synthesized TEAL is 97.47 % with high accuracy (ca. 99 %). This work suggests a new strategy to detect the purity and impurities of some kind of hazardous compounds.

建立一种凤香型基酒密度和粘度指标的快速分析模型,以期为基酒真实性判别提供一种方法参考.利用近红外光谱技术对样品进行光谱扫描,并结合常规分析方法,通过相关建模软件和算法最终建立了其定量分析模型.结果表明,所建模型的线性关系较好,且各模型内部参数RMSEC和RMSECV值均较小,说明模型的质量较好;另取验证样品对模型进行外部验证,通过与常规分析方法的检测结果进行比较,各模型预测的平均相对误差均小于0.2％,同时两种方法的检测值之间均不存在显著性差异(P>0.05),说明模型具有较好的预测能力.

为实现凤香型原酒三大酯(乙酸乙酯、乳酸乙酯和己酸乙酯)含量的快速分析,及时地为基酒入库分级提供数据参考.利用中红外光谱技术对原酒样品进行光谱扫描,并结合传统分析方法,最终建立了原酒三大酯指标的快速分析模型.结果表明,所建模型的各线性关系较好,R2均在0.95以上,且模型内部参数RMSEC和RMSECV值均较小且相近,模型的质量较好;另取验证样品对模型进行外部验证,通过与传统分析方法的检测结果进行比较,结果表明,三大酯指标预测值和真实值的相关系数均大于0.95,模型预测的平均相对误差均小于5.0％,同时两种方法的检测值之间不存在显著性差异(P>0.05),说明模型具有较好的预测能力,可用于实际生产分析检测.

为实现高粱淀粉含量的快速分析,提高样品日常分析效率,利用近红外光谱技术对样品进行光谱扫描,并结合常规分析方法,最终建立了高粱淀粉含量的快检模型.结果表明,所建模型的各线性关系较好,R2均在0.90以上,且模型内部参数RMSEC、RMSECV和RMSEP值均较小且相近,分别为0.207、0.389和0.211,模型的质量较好.另取验证样品对模型进行外部验证,通过与常规分析方法的检测结果进行比较,模型预测的平均相对误差为0.7％,同时两种方法的检测值之间不存在显著性差异(P＞0.05),说明模型具有较好的预测能力,可用于实际生产分析检测.

为实现酿酒辅料稻壳水分的快速分析,提高样品日常的检测效率.采用近红外光谱技术对稻壳样品进行光谱扫描,并结合常规分析方法,最终建立了稻壳水分指标的快检模型.结果表明,所建模型的各线性关系较好,R2均在0.95以上,且模型内部参数RMSEC、RMSECV和RMSEP值均较小且相近,分别为0.113,0.142,0.149,模型的质量较好;另取验证样品对模型进行外部验证,通过与常规分析方法进行比较,模型预测的平均相对误差为1.4％,同时2种方法的检测值之间不存在显著性差异(p＞0.05),说明模型具有较好的预测能力,可用于样品日常分析检测.

The detection and separation of impurities are crucial for the quality control of triethylboron. While the purity of triethylboron is the key factor for ballistic performance, impact accuracy, and storage safety. This work will explore a modified gas chromatography-mass spectrometer method to detect and separate the impurities in synthesized triethylboron. They are identified as CH3CH2OCH2CH3, (CH3CH2)(3)B, (CH3CH2O)(3)B, CH3CH2CH2CH2B(CH3CH2)(2) and (CH3CH2)(3)O3B3. Further, optimized GC conditions are demonstrated to quantify the synthesized triethylboron accurately and determine its purity. The purity of triethylboron is 96.39 % via a normalization method with high accuracy (RSD=0.05 %). The optimized GC conditions were as follows: a hydrogen flame ionization detector, nitrogen as carrier gas whose pathway was installed with a water trap and an oxygen trap, an HP-5 capillary column (30 mx0.32 mmx0.25 mu m) with the column temperature of 60 degrees C, an inlet temperature of 120 degrees C, the detector temperature of 150 degrees C, and an injection volume of 0.6 mu L. This work demonstrates a simple and precise method to analyze the synthesized triethylboron and its impurities.

为实现原酒乙酸乙酯含量的快速分析,及时地为原酒入库分级提供数据参考.利用近红外光谱技术对原酒样品进行光谱扫描,并结合传统分析方法,最终建立了原酒乙酸乙酯指标的快速分析模型.结果 表明,所建模型的各线性关系较好,R2均在0.95以上,且模型内部参数RMSEC、RMSECV和RMSEP值均较小且相近,分别为7.87、9.22和9.42,模型的质量较好;另取验证样品对模型进行外部验证,通过与传统分析方法的检测结果进行比较,模型预测的平均相对误差为2.8％,同时两种方法的检测值之间不存在显著性差异(P＞0.05),说明模型具有较好的预测能力,可用于实际生产分析检测.

The components in polymer-bonded explosive X, including cyclotetramethylene-tetranitramine, paraffin, and polytetrafluoroethylene, were determined using near-infrared (NIR) spectroscopy. Using partial least squares as the multivariate calibration method, quantitative calibration models for components in X were verified internally and externally. The possible combinations of eight general spectral pretreatment methods and different bands of the scanning spectral region (12,500-4000 cm-1) were established. The models were analyzed, evaluated, and optimized via the fitting effect. The data were combined with the mathematical meaning of the model spectral pretreatment methods and the chemical significance of the modeling spectral bands. Prediction performance offered optimal quantitative calibration models. Paired bilateral Student's t tests show that there is no significant difference between the values obtained by NIR and chemical analysis methods, and the NIR method has good accuracy. Moreover, the precision of the NIR method is better than that of the chemical method, and the analysis time is reduced from 2 days to a few minutes.

为了实现高粱水分指标含量的快速分析,并使分析结果能够及时有效地反馈到生产车间,以指导车间日常生产.采用近红外光谱技术对高粱样品进行光谱扫描,并结合传统分析方法,最终建立了高粱水分指标的快速分析模型.结果表明,所建模型的线性关系较好,R2为0.9558,且模型内部参数RMSEC和RMSECV值均较小,分别为0.280和0.319,模型的质量较好;另取验证样品对模型的预测能力进行验证,通过与传统分析方法的分析结果进行比较,模型预测的平均相对误差小于2.0％,说明模型具有较好的预测能力,可用于实际生产分析检测,提高样品分析效率.

为实现曲粮大麦、小麦以及豌豆水分的快速分析,提高样品日常的检测效率.采用近红外光谱技术对曲粮样品进行光谱扫描,并结合传统分析方法,最终分别建立了大麦、小麦以及豌豆水分指标的快速分析模型.结果表明,所建模型的线性关系均较好,R2分别为0.9900、0.9942、0.9952,且各模型内部参数RMSEC和RMSECV值均较小,分别为0.226、0.196、0.142和0.250、0.209、0.156,各模型质量均较好;另取验证样品对各模型的预测能力进行验证,通过与传统分析方法的结果进行比较,各模型预测的平均相对误差均小于2.0％,说明所建曲粮模型均具有较好的预测能力,可用于样品日常分析检测.

提出了一种RDX非自身标样定量分析方法,通过气相色谱-原子发射检测器(GC-AED)实现了无RDX标准物质情况下对RDX的高灵敏度定量检测,有效解决了无RDX标准物质带来的定量检测难题.结果 表明:最佳定量元素谱线为C(193 nm),最佳外标物为DBP.以DBP为外标物,通过C(193 nm)定量测定RDX的线性范围为0.07～ 2.02 mg/mL.以3倍信噪比计算检测限,得到RDX的方法检测限为0.06 μg/mL.RDX在线性范围内高、中、低浓度的加标回收率分别为97.45％、101.37％和87.71％,相对标准偏差分别为0.33％、0.81％和0.36％.该方法具有较好的准确度和精密度,在无RDX标准物质的情况下亦能对RDX进行定量分析.

The effect of acid concentration in nitrocellulose (NC) on thermal stability of NC for civil use was studied using differential scanning calorimeter (DSC). The thermal explosion temperature (Tb) and kinetic parameters of NC thermal decomposition were calculated using different methods based on the thermal decomposition behaviors of NC in high pressure-gold plated stainless steel crucible at different heating rates. The results show that acid content in NC reduces the stability of NC remarkably, including reducing the characteristic decomposition temperature such as peak temperature (Tp), onset temperature (Tonset), initial temperature (T0) and critical temperature of thermal explosion (Tb). The big difference between NC with different acid content is that the relative quantity of activation energy and the change trend of activation energy. The most interesting thing is that NC with more acid content has the lower characteristic temperature but higher calculated activation energy.

气相色谱–原子发射检测器(GC–AED)是一种特色鲜明的联用技术,兼具气相色谱(GC)的分离能力和原子发射检测器(AED)的元素检测能力,能够进行多元素检测的同时还具有元素的等摩尔效应.介绍了GC–AED的原理、组成和特点,重点综述了GC–AED在元素选择性检测、组分定量分析及元素比例式推断共三个方面的应用概况,并对其应用前景进行了展望.

为了快速检测西凤酒大曲水分、酸度以及淀粉含量指标,使分析结果及时反馈到生产车间,为车间生产提供数据参考.采用近红外光谱技术结合偏最小二乘法(PLS)处理大曲样品近红外图谱,分别建立了西凤酒大曲水分、酸度以及淀粉指标的定量分析模型.结果表明,各模型的线性关系整体较好,且模型参数RMSEC和RMSECV值较小,说明模型的质量较好;另取验证样品对模型预测能力进行验证,通过与常规分析方法检测结果进行比较,各模型预测的平均相对误差均小于3.0％,说明模型具有较好的预测能力,可用于日常生产分析检测.

A rapid component detection method of modified Composition B was studied by near infrared spectrum technology. The quantitative models of main components TNT and RDX were established by partial least squares（PLS）method. Results show that the coefficients of determination of cross‑validation （R2）of established models are 0.9611, 0.9415, respectively. The root mean square errors of cross‑validation (RMSECV) are 0.478%, 0.501%, respectively, and the root mean square errors of prediction (RMSEP) of external‑validation are 555%, 521%, respectively, showing that the established quantitative model has high accuracy and precision. The calculated results of t‑test performed by determined values of near infrared method and reference values of the standard analysis method are tRDX=1.03, tTNT=0.72, respectively. Their t values are smaller than the critical value t9=2.26 (α9=0.05), revealing that there is no significant difference between the near infrared spectroscopy and the standard analysis method.

采用气相色谱-原子发射光谱对TNT进行定量测定,建立了一种非含能非自身标样定量测定方法.以二号中定剂为外标物,通过二号中定剂的C元素标准曲线定量TNT中的C元素.结果 表明:用二号中定剂作外标物定量TNT的C元素线性范围是22.43 ～637.50μg/mL,相关系数(r)为0.999 9.以3倍信噪比计算检测限,得到TNT的C元素方法检测限为0.015 μg/mL.TNT在高、中、低浓度下的回收率为94.6％～101.3％,相对标准偏差为0.3％～0.7％.该方法具有较好的准确度和精密度,在自身标准物质缺乏的情况下能够对TNT进行准确定量.

为了在标准物质缺乏的情况下能够准确定量3,4-二硝基呋咱基氧化呋咱(DNTF),采用气相色谱-原子发射光谱(GC-AED)联用技术,建立了一种DNTF的非含能非自身标样定量测定方法;通过对比DNTF和5种非含能材料(二苯胺、苯甲酸、乙酰苯胺、DBS、DBP)的C元素标准曲线,选择相对定量响应因子最接近1的非含能材料作外标物定量DNTF的C元素,再结合分子式对DNTF进行定量.结果表明,苯甲酸的相对定量响应因子最接近1,用苯甲酸的C元素标准曲线定量DNTF的C元素,得到DNTF的C元素线性范围为17.43～426.92μg/mL,相关系数(r)为0.9992;以3倍信噪比计算检测限,得到DNTF的C元素方法检测限为0.0096μg/mL;相同条件下,DNTF在线性范围内高、中、低浓度下的加标回收率分别为104.9％,103.7％和109.4％,相对标准偏差分别为3.3％、1.1％和1.5％;该方法具有较好的准确度和精密度,可以实现在自身标准物质缺乏的情况下用非含能材料作外标物对DNTF的准确定量.