CHINESE JOURNAL OF ENERGETIC MATERIALS
+Advanced Search
en
×

分享给微信好友或者朋友圈

使用微信“扫一扫”功能。
参考文献 1
周培毅, 徐更光, 张景云, 等. 改性B 炸药装药发射安全性实验研究[J]. 火炸药学报, 1999, 22(4): 34-36.
ZHOUPei‑yi, XUGeng‑guang, ZHANGJing‑yun, et al. The experimental study of lunching safety of modified composition B Explosive Charge[J]. Chinese Journal of Explosives & Propellants, 1999, 22(4): 34-36.
参考文献 2
孙家利, 王秋雨, 夏克祥, 等.高效毁伤注装药工艺技术[J]. 兵工自动化, 2012, 31(1): 16-17.
SUNJia‑li, WANGQiu‑yu, XIAKe‑xiang, et al. High efficiency damage melt‑loading process technology[J]. Ordnance Industry Automation, 1999, 22(4): 34-36.
参考文献 3
林宏亮, 朱晓红, 付强, 等. 二维气相色谱检测爆炸残留物中TNT[J]. 刑事技术, 2009, 34(3): 37-38.
LINHong‑liang, ZHUXiao‑hong, FUQiang, et al. Detection of TINT in explosive residues by two‑dimensional gas chromatography[J]. Ordnance Industry Automation, 2009, 34(3): 37-38.
参考文献 4
褚小立, 陆婉珍.近五年我国近红外光谱分析技术研究与应用进展[J]. 光谱学与光谱分析, 2014, 34(10): 2595-2999.
CHUXiao‑li, LUWan‑zhen. Rsearch and application progress of near infrared spectroscopy analytical technology in China in the past five years[J]. Spectroscopy and Spectral Analysis, 2014, 34(10): 2595-2599.
参考文献 5
温晓燕, 苏鹏飞, 刘红妮, 等. 近红外漫反射光谱法测定硝化棉含氮量的数值模拟及实验研究[J]. 火炸药学报, 2014, 37(6): 87-90.
WENxiao‑yan, SUpeng‑fei, LIUHong‑ni, et al. Numerical simulation and experimental study on the content in nitrocellulose by near infrared diffuse reflectance spectroscopy[J]. Chinese Journal of Explosives & Propellants, 2014, 37(6): 87-90.
参考文献 6
邹权, 邓国栋, 郭效德, 等. 近红外在线检测硝胺推进剂主要固体组分含量[J]. 固体火箭技术, 2008, 31(5): 62-65.
SUPeng‑fei, ZHANGgao, WENxiao‑yan, et al. Real‑time detection of the main solid components concentration in RDX‑CMDB with NIR spectral analysis[J]. Journal of Solid Rocket Technology, 2012, 35(6): 838-841.
参考文献 7
JINYe, WUZeng, LIUXue‑son. Near infrared spectroscopy in combination with chemometrics as a process analytical technology tool for on‑line quantitative monitoring of alcohol precipitation[J]. Joural of Pharmaceutial and Biomedical Analysis, 2013, 77(5): 32-39.
参考文献 8
GJB8094-2013. 改性B炸药规范[S]. 中华人民共和国国家军用标准.
GJB8094-2013.Specification for modified composition B[S]. National Military Standards of the People's Republic of China.
参考文献 9
严衍禄, 赵龙莲, 韩东海, 等.近红外光谱分析基础与应用[M]. 北京: 中国轻工业出版社, 2005: 125.
YANYan‑lu, ZHAOLong‑lan, HANDong‑hai, et al. The basis and application of near infrared spectroscopy[M]. Beijing: Chinese Light Industry Press, 2005: 125.
参考文献 10
翁诗莆.傅里叶变换红外光谱分析[M]. 北京: 化学工业出版社, 2009: 232-259.
WENGShi‑pu. Analysis of Fu Liye transform infrared spectroscopy[M]. Beijing: Chemical Industry Press, 2009: 232-259.
参考文献 11
王镇浦, 周国华, 罗国安.偏最小二乘法及其在分析化学中的应用[J]. 分析化学, 1989, 17(7): 662-669.
WANGZhen‑pu, ZHOUGuo‑hua, LUOGuo‑an. Partial least squares method and its application in analytical chemistry[J]. Aanalytical Chemistry, 1989, 17(7): 662-669.
目录 contents

    摘要

    采用近红外光谱技术,研究了改性B炸药近红外光谱快速组分检测方法。采用偏最小二乘法(PLS)建立了主要组分梯恩梯(TNT)和黑索今(RDX)的定量模型。所建模型的交互验证决定系数(R2)分别为0.9611、0.9415,内部交互验证的均方根误差(RMSECV)分别为0.478%、0.501%,外部验证的均方根误差(RMSEP)分别为0.555%、0.521%,表明所建定量模型具有较高的准确度和精密度。用近红外法的测定值与标准分析法的参考值进行t检验的计算结果为tRDX=1.03,tTNT=0.72,均小于临界值t9=2.26(α9=0.05),这表明,近红外光谱法与标准分析法无显著性差异。

    Abstract

    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.

  • 1 引 言

    1

    B炸药是一种以梯恩梯(TNT)、黑索今(RDX)为基的高能炸药。在北约等西方国家一般采用注装B炸药的方式来提高中、大口径弹丸的爆炸威[1]。我国科学家针对B炸药在生产和应用中存在的力学性能和使用安全问题进行改进,形成了改性B炸药,已应用于多种大口径弹[2]

    改性B炸药的组成是其性能的基本保证。无论是火炸药研究单位还是装药生产企业,都需要对改性B炸药的各种组分含量进行定量分析,才能保证工艺过程的稳定性,从而保障改性B炸药最终产品的可靠性。工艺过程控制和产品质量检测方法一直沿用重量法或气相色谱法等标准方[3],这些方法所需仪器设备多、操作步骤繁琐、使用溶剂种类多、数量大、耗时费力,尤其是试样前处理,研磨样品时,存在一定的安全隐患,分析时间往往需要数小时,检测速度缓慢,难以满足改性B炸药生产和装药过程安全快速检测的要求。

    近红外光谱的波长为780~2526 nm,该区域主要是C─H,O─H,N─H和S─H等含氢基团振动光谱的倍频及合频吸收,这些吸收谱带信号丰富。近红外光谱技术最大的优势是能够实现快速、无损、实时检测,已成为近十年来石油、烟草、医药等行业生产过程中备受瞩目的在线检测技4。近年来,一些文献报道了硝化棉、混合炸药和固体推进剂等的近红外光谱技术研5,6,7,但迄今为止,尚未见有关近红外光谱技术在改性B炸药组分含量检测中应用的文献报道。

    为此,本研究针对炸药实际生产线,采用近红外光谱技术,对改性B炸药主要组分TNT和RDX快速检测方法进行研究,建立近红外光谱校正模型,并进行应用验证,论证近红外光谱法的可靠性。

  • 2 实验部分

    2
  • 2.1 样品集及参考值

    2.1

    收集改性B炸药生产线不同批次送检样品200个,设计制备样品96个,作为样品集,样品集组分含量范围内样品数分布基本均匀。按照标准方法GJB8094-2013[8],检测样品集中TNT和RDX组分含量作为建模样品的参考值。样品集各组分含量范围见表1

    表1 样品集主要组分含量范围(%)

    Table 1 Content range of main components in sample set(%)

    componentcontent rangeproduct range
    TNT35.6-46.840.8±2.0
    RDX50.8-59.955.0±2.0
    表1
                    样品集主要组分含量范围
  • 2.2 光谱测量条件

    2.2

    光谱测定仪器:BRUKER公司MPA型FT‑NIR近红外光谱仪,采样方式:积分球漫反射,扫描参数:扫描光谱范围4000~12500 cm-1;扫描次数:64次;分辨率:8 cm-1;数据形式:吸光度。

  • 2.3 光谱采集

    2.3

    试样制备:按照标准方[8]准备试样。

    光谱采集:近红外光谱仪开机后预热0.5 h,进行光谱采集。按照光谱测量条件扫描背景;取适量试样倒入样品杯,用木槌压实,样品厚度应大于3 cm。每个试样重新装样,测量3次,取平均光谱。

  • 3 结果与讨论

    3
  • 3.1 光谱测定条件的选择

    3.1

    仪器的分辨率影响样品光谱的质量,最终影响分析的准确性。根据经验,一般要求仪器分辨率应相当于样品光谱半高宽的1/5~1/10,样品在近红外谱区漫反射光谱的半高宽约为150~400 cm-1,因此理论上用15 cm-1的分辨率能够满足近红外分[9]。为了适度提高扫描速度,本实验中采用分辨率8 cm-1

    为了选取合适的扫描次数以提高信噪比,进行以下实验:选取30个改性B炸药样品,其中22个作为校正集建立模型,8个作为验证集进行模型验证,光谱扫描次数分别为32、64、96,建立TNT组分含量的校正模型,并用决定系数(R2)和预测标准偏差(RMSEP)来评价模型,结果见表2。由表2可以看出,扫描次数的增加能改善模型的预测准确性,当增加至64次后,不再有显著性提高,因此,本实验选择64次为扫描次数。

    表2 扫描次数对模型的影响

    Table 2 The effect of scanning times on models

    scanning timesR2RMSEP / %
    320.93950.427
    640.96670.461
    960.96890.469
    表2
                    扫描次数对模型的影响
  • 3.2 改性B炸药图谱分析

    3.2

    利用近红外光谱仪采集到的改性B炸药样品集的近红外光谱叠加图如图1所示。RDX、TNT与改性B炸药的近红外光谱见图2

    图1
                            296个改性B炸药样品的近红外光谱

    图1 296个改性B炸药样品的近红外光谱

    Fig.1 NIR spectra of 296 modified Composition B samples

    图2
                            TNT、RDX和改性B样品的近红外光谱

    图2 TNT、RDX和改性B样品的近红外光谱

    Fig.2 NIR spectra of TNT, RDX and modified Composition B samples

    从图1与图2可看出,TNT和RDX分子中甲基和亚甲基中C─H键伸缩振动的二倍频分别在8935 cm-1和8854 cm-1左右,伸缩振动和弯曲振动一倍频的合频在8557 cm-1和8425 cm-1左右,伸缩振动一倍频均在8030 cm-1左右,这些倍频、合频谱峰受基体的影响,出现位移和叠加,给准确分析带来一定困难。改性B炸药样品的近红外光谱一级倍频吸光度小于0.1,二级倍频吸光度小于0.4,整体表现出复杂、重叠、变动的弱信息,因此采用简单的线性回归法很难建立TNT和RDX组分含量与改性B炸药光谱之间的线性关系。为了建立快速、简单的分析方法,需要对光谱进行预处理,并采用化学计量学方法实现从复杂、重叠、变动的光谱中提取大量有效信息,排除各种干扰因素的影响,建立定量模型。

  • 3.3 定量模型的建立与优化

    3.3

    从样品集中选择216个样品作为校正集用于建立校正模型,所选样本能够覆盖整个样品集组分含量范围,其余样品作为验证集,用于对所建立的校正模型进行外部验证。采用偏最小二乘[10]建立模型,选择交叉验证法对模型进行内部检验,通过杠杆值对光谱异常样品进行判断,依据模型内部检验和外部验证相结合,选择不同波长范围、不同预处理方法并逐步优化,确定定量模型。评价参数依次为RMSEP、RMSECV(校正集预测值与参考值的均方根误差)、R2、RMSECV与主因子数的关系图,其中RMSEP、RMSECV越小模型越好;R2越接近1,模型越好;在RMSECV与主因子数的关系图中,RMSECV的值随着主因子数的增大先快速减小,然后缓慢增大,符合这种规律的模型稳定性好。采用一阶导(1st D)、二阶导(2nd D)、矢量归一化(SNV)、多元散射校正(MSC)、最小‑最大归一化(Min & max N)、1st D+ SNV、1st D + MSC、消除常数偏移量、减去一条直线、一阶导+减去一条直线等多种光谱预处理方[11],以选择最佳光谱预处理方法,提高光谱与组分含量的相关性。以改性B炸药RDX组分为例进行模型优化,并结合其它评价参数进行综合评价,部分校正模型的评价参数见表3。从表3可以看出,模型5的R2最大,RMSECV、RMSEP最小,而且主因子数图也很好,为最优模型,因此确定为RDX组分的定量模型。

    RDX组分定量模型的光谱预处理方法为最小‑最大归一化法,光谱定量范围较窄,为7502.2~5446.4 cm-1,其中RDX组分最优模型的主因子数为8,交叉验证R2为0.9415,预测值与参考值的线性关系见3,RMSECV与主因子数的关系见4,RMSECV为0.501%,RMSEP为0.555%。采用同样方法,优化建立TNT组分定量模型,其光谱预处理方法也是最大‑最小归一化法,光谱定量范围为9338~6098 cm-1和5450~4598 cm-1,模型的主因子数为6,定量模型的R2为0.9611,预测值与标准值的线性关系见5,RMSECV与主因子数的关系见6,RMSECV为0.478%,RMSEP为0.521%。上述结果表明,定量模型提取有效信息效率高,其预测值与参考值的线性关系较好,具有很高的准确度和预测能力。

    表3 RDX模型优化参数

    Table 3 Optimization parameters of RDX model

    modelpreprocessing method

    quantitative range

    / cm-1

    main

    factor

    R2RMSECV / %RMSEP / %main factor number graphcomprehensive evaluation
    11st D9338.3-4597.890.92470.5550.606very good
    21st D+ SNV7502.2-5446.480.93630.5230.610good
    3SNV7502.2-5446.480.93570.5260.600very good
    41st D

    9338.3-7490.7

    6102.1-5446.4

    80.92720.5470.565very good
    5Min & max N7502.2-5446.480.94150.5010.555very goodbest
    61st D+ SNV9338.3-4597.870.92560.5530.624very good
    71st D9338.3-4597.890.93740.5100.574very goodbetter
    81st D +SNV7502.2-5446.470.94940.4660.639good
    91st D +SNV7502.2-5446.490.90310.6420.641poor
    101st D

    9338.3-7490.7

    6102.1-5446.4

    90.89950.6530.572very goodbetter
    11SNV7502.2-5446.490.90650.6300.653very good
    12Min & max N7502.2-5446.490.89970.6520.576very goodbetter
    131st D +SNV9338.3-5446.4100.84730.7980.604very good
    图3
                            RDX模型交叉验证预测值与参考值的关系

    图3 RDX模型交叉验证预测值与参考值的关系

    Fig.3 Relationship between the predicted cross validation values of the RDX model and their reference values

    图4
                            RDX模型RMSECV与主因子数的关系

    图4 RDX模型RMSECV与主因子数的关系

    Fig.4 Relationship between RMSECV and the main factors of the RDX model

    图5
                            TNT模型交叉验证预测值与参考值的关系

    图5 TNT模型交叉验证预测值与参考值的关系

    Fig.5 Relationship between the predicted cross validation values of the TNT model and their reference values

    图6
                            TNT模型RMSECV与主因子数的关系

    图6 TNT模型RMSECV与主因子数的关系

    Fig.6 Relationship between RMSECV and the main factors of the TNT model

  • 3.4 应用验证

    3.4

    分别采用标准方[8]和所建近红外定量模型,对随机选取的10个不同批次改性B炸药样品进行测定,结果见表4。为进一步确定两种方法的准确度是否存在显著性差异,采用t‑对子双边检验。由表4的数据计算得到:tRDX=1.03,tTNT=0.72,而临界值t9=2.26(α9=0.05),t<t9,故在置信度为95%时,两种方法不存在显著性差异,说明近红外光谱法与标准方法的准确度一致。

    表4 近红外法测定值与参考值

    Table 4 Determination values by near infrared method and their reference values%

    serial numberRDXTNT
    determination valuereference valuedifference valuedetermination valuereference valuedifference value
    156.0555.94-0.1139.8440.170.33
    255.5155.30-0.2140.3040.510.21
    355.2255.720.5040.7140.30-0.41
    455.8356.080.2540.1040.07-0.03
    555.3955.870.4840.4540.18-0.27
    654.8054.990.1941.0941.160.07
    755.4255.540.1240.3240.660.34
    854.5954.910.3141.1441.160.02
    954.9154.88-0.0340.8241.070.25
    1055.2354.78-0.4540.7840.860.08
  • 4 结 论

    4

    基于改性B炸药主要组分的近红外光谱检测方法,通过模型优化,建立了与标准方法准确度一致的两种主要组分TNT和RDX的定量模型,所建TNT和RDX定量模型R2分别为0.9415、0.9611,RMSECV分别为0.478%、0.501%,RMSEP依次为0.521%、0.555%,表明模型具有很好的预测能力。在显著水平为0.05时,近红外模型预测值与参考值无显著性差异,因此改性B炸药组分的近红外光谱检测法可以达到实际应用要求。

  • 参考文献

    • 1

      周培毅, 徐更光, 张景云, 等. 改性B 炸药装药发射安全性实验研究[J]. 火炸药学报, 1999, 22(4): 34-36.

      ZHOU Pei‑yi, XU Geng‑guang, ZHANG Jing‑yun, et al. The experimental study of lunching safety of modified composition B Explosive Charge[J]. Chinese Journal of Explosives & Propellants, 1999, 22(4): 34-36.

    • 2

      孙家利, 王秋雨, 夏克祥, 等.高效毁伤注装药工艺技术[J]. 兵工自动化, 2012, 31(1): 16-17.

      SUN Jia‑li, WANG Qiu‑yu, XIA Ke‑xiang, et al. High efficiency damage melt‑loading process technology[J]. Ordnance Industry Automation, 1999, 22(4): 34-36.

    • 3

      林宏亮, 朱晓红, 付强, 等. 二维气相色谱检测爆炸残留物中TNT[J]. 刑事技术, 2009, 34(3): 37-38.

      LIN Hong‑liang, ZHU Xiao‑hong, FU Qiang, et al. Detection of TINT in explosive residues by two‑dimensional gas chromatography[J]. Ordnance Industry Automation, 2009, 34(3): 37-38.

    • 4

      褚小立, 陆婉珍.近五年我国近红外光谱分析技术研究与应用进展[J]. 光谱学与光谱分析, 2014, 34(10): 2595-2999.

      CHU Xiao‑li, LU Wan‑zhen. Rsearch and application progress of near infrared spectroscopy analytical technology in China in the past five years[J]. Spectroscopy and Spectral Analysis, 2014, 34(10): 2595-2599.

    • 5

      温晓燕, 苏鹏飞, 刘红妮, 等. 近红外漫反射光谱法测定硝化棉含氮量的数值模拟及实验研究[J]. 火炸药学报, 2014, 37(6): 87-90.

      WEN xiao‑yan, SU peng‑fei, LIU Hong‑ni, et al. Numerical simulation and experimental study on the content in nitrocellulose by near infrared diffuse reflectance spectroscopy[J]. Chinese Journal of Explosives & Propellants, 2014, 37(6): 87-90.

    • 6

      邹权, 邓国栋, 郭效德, 等. 近红外在线检测硝胺推进剂主要固体组分含量[J]. 固体火箭技术, 2008, 31(5): 62-65.

      SU Peng‑fei, ZHANG gao, WEN xiao‑yan, et al. Real‑time detection of the main solid components concentration in RDX‑CMDB with NIR spectral analysis[J]. Journal of Solid Rocket Technology, 2012, 35(6): 838-841.

    • 7

      JIN Ye, WU Zeng, LIU Xue‑son. Near infrared spectroscopy in combination with chemometrics as a process analytical technology tool for on‑line quantitative monitoring of alcohol precipitation[J]. Joural of Pharmaceutial and Biomedical Analysis, 2013, 77(5): 32-39.

    • 8

      GJB8094-2013. 改性B炸药规范[S]. 中华人民共和国国家军用标准.

      GJB8094-2013.Specification for modified composition B[S]. National Military Standards of the People's Republic of China.

    • 9

      严衍禄, 赵龙莲, 韩东海, 等.近红外光谱分析基础与应用[M]. 北京: 中国轻工业出版社, 2005: 125.

      YAN Yan‑lu, ZHAO Long‑lan, HAN Dong‑hai, et al. The basis and application of near infrared spectroscopy[M]. Beijing: Chinese Light Industry Press, 2005: 125.

    • 10

      翁诗莆.傅里叶变换红外光谱分析[M]. 北京: 化学工业出版社, 2009: 232-259.

      WENG Shi‑pu. Analysis of Fu Liye transform infrared spectroscopy[M]. Beijing: Chemical Industry Press, 2009: 232-259.

    • 11

      王镇浦, 周国华, 罗国安.偏最小二乘法及其在分析化学中的应用[J]. 分析化学, 1989, 17(7): 662-669.

      WANG Zhen‑pu, ZHOU Guo‑hua, LUO Guo‑an. Partial least squares method and its application in analytical chemistry[J]. Aanalytical Chemistry, 1989, 17(7): 662-669.

温晓燕

机 构:西安近代化学研究所, 陕西 西安 710065

Affiliation:Xi′an Modern Chemistry Research Institute, Xi′an 710065, China

邮 箱:wenxy204@163.com

作者简介:温晓燕(1974-),女,高级工程师,主要从事火炸药理化分析研究工作. e‑mail:wenxy204@163.com

苏鹏飞

机 构:西安近代化学研究所, 陕西 西安 710065

Affiliation:Xi′an Modern Chemistry Research Institute, Xi′an 710065, China

潘清

机 构:西安近代化学研究所, 陕西 西安 710065

Affiliation:Xi′an Modern Chemistry Research Institute, Xi′an 710065, China

董小虎

机 构:西安近代化学研究所, 陕西 西安 710065

Affiliation:Xi′an Modern Chemistry Research Institute, Xi′an 710065, China

杨建明

机 构:

1. 西安近代化学研究所, 陕西 西安 710065

2. 氟氮化工资源高效开发与利用国家重点实验室, 陕西 西安 710065

Affiliation:

1. Xi′an Modern Chemistry Research Institute, Xi′an 710065, China

2. State Key Laboratory of Fluorine & Nitrogen Chemicals, Xi′an 710065, China

componentcontent rangeproduct range
TNT35.6-46.840.8±2.0
RDX50.8-59.955.0±2.0
scanning timesR2RMSEP / %
320.93950.427
640.96670.461
960.96890.469
html/hnclen/CJEM2018120/media/285a7c9d-012c-4962-afba-290c96edde00-image001.png
html/hnclen/CJEM2018120/alternativeImage/285a7c9d-012c-4962-afba-290c96edde00-F002.jpg
modelpreprocessing method

quantitative range

/ cm-1

main

factor

R2RMSECV / %RMSEP / %main factor number graphcomprehensive evaluation
11st D9338.3-4597.890.92470.5550.606very good
21st D+ SNV7502.2-5446.480.93630.5230.610good
3SNV7502.2-5446.480.93570.5260.600very good
41st D

9338.3-7490.7

6102.1-5446.4

80.92720.5470.565very good
5Min & max N7502.2-5446.480.94150.5010.555very goodbest
61st D+ SNV9338.3-4597.870.92560.5530.624very good
71st D9338.3-4597.890.93740.5100.574very goodbetter
81st D +SNV7502.2-5446.470.94940.4660.639good
91st D +SNV7502.2-5446.490.90310.6420.641poor
101st D

9338.3-7490.7

6102.1-5446.4

90.89950.6530.572very goodbetter
11SNV7502.2-5446.490.90650.6300.653very good
12Min & max N7502.2-5446.490.89970.6520.576very goodbetter
131st D +SNV9338.3-5446.4100.84730.7980.604very good
html/hnclen/CJEM2018120/media/285a7c9d-012c-4962-afba-290c96edde00-image003.png
html/hnclen/CJEM2018120/media/285a7c9d-012c-4962-afba-290c96edde00-image004.png
html/hnclen/CJEM2018120/media/285a7c9d-012c-4962-afba-290c96edde00-image006.png
html/hnclen/CJEM2018120/media/285a7c9d-012c-4962-afba-290c96edde00-image005.png
serial numberRDXTNT
determination valuereference valuedifference valuedetermination valuereference valuedifference value
156.0555.94-0.1139.8440.170.33
255.5155.30-0.2140.3040.510.21
355.2255.720.5040.7140.30-0.41
455.8356.080.2540.1040.07-0.03
555.3955.870.4840.4540.18-0.27
654.8054.990.1941.0941.160.07
755.4255.540.1240.3240.660.34
854.5954.910.3141.1441.160.02
954.9154.88-0.0340.8241.070.25
1055.2354.78-0.4540.7840.860.08

表1 样品集主要组分含量范围(%)

Table 1 Content range of main components in sample set(%)

表2 扫描次数对模型的影响

Table 2 The effect of scanning times on models

图1 296个改性B炸药样品的近红外光谱

Fig.1 NIR spectra of 296 modified Composition B samples

图2 TNT、RDX和改性B样品的近红外光谱

Fig.2 NIR spectra of TNT, RDX and modified Composition B samples

表3 RDX模型优化参数

Table 3 Optimization parameters of RDX model

图3 RDX模型交叉验证预测值与参考值的关系

Fig.3 Relationship between the predicted cross validation values of the RDX model and their reference values

图4 RDX模型RMSECV与主因子数的关系

Fig.4 Relationship between RMSECV and the main factors of the RDX model

图5 TNT模型交叉验证预测值与参考值的关系

Fig.5 Relationship between the predicted cross validation values of the TNT model and their reference values

图6 TNT模型RMSECV与主因子数的关系

Fig.6 Relationship between RMSECV and the main factors of the TNT model

表4 近红外法测定值与参考值

Table 4 Determination values by near infrared method and their reference values%

image /

无注解

无注解

无注解

无注解

无注解

无注解

无注解

无注解

无注解

无注解

  • 参考文献

    • 1

      周培毅, 徐更光, 张景云, 等. 改性B 炸药装药发射安全性实验研究[J]. 火炸药学报, 1999, 22(4): 34-36.

      ZHOU Pei‑yi, XU Geng‑guang, ZHANG Jing‑yun, et al. The experimental study of lunching safety of modified composition B Explosive Charge[J]. Chinese Journal of Explosives & Propellants, 1999, 22(4): 34-36.

    • 2

      孙家利, 王秋雨, 夏克祥, 等.高效毁伤注装药工艺技术[J]. 兵工自动化, 2012, 31(1): 16-17.

      SUN Jia‑li, WANG Qiu‑yu, XIA Ke‑xiang, et al. High efficiency damage melt‑loading process technology[J]. Ordnance Industry Automation, 1999, 22(4): 34-36.

    • 3

      林宏亮, 朱晓红, 付强, 等. 二维气相色谱检测爆炸残留物中TNT[J]. 刑事技术, 2009, 34(3): 37-38.

      LIN Hong‑liang, ZHU Xiao‑hong, FU Qiang, et al. Detection of TINT in explosive residues by two‑dimensional gas chromatography[J]. Ordnance Industry Automation, 2009, 34(3): 37-38.

    • 4

      褚小立, 陆婉珍.近五年我国近红外光谱分析技术研究与应用进展[J]. 光谱学与光谱分析, 2014, 34(10): 2595-2999.

      CHU Xiao‑li, LU Wan‑zhen. Rsearch and application progress of near infrared spectroscopy analytical technology in China in the past five years[J]. Spectroscopy and Spectral Analysis, 2014, 34(10): 2595-2599.

    • 5

      温晓燕, 苏鹏飞, 刘红妮, 等. 近红外漫反射光谱法测定硝化棉含氮量的数值模拟及实验研究[J]. 火炸药学报, 2014, 37(6): 87-90.

      WEN xiao‑yan, SU peng‑fei, LIU Hong‑ni, et al. Numerical simulation and experimental study on the content in nitrocellulose by near infrared diffuse reflectance spectroscopy[J]. Chinese Journal of Explosives & Propellants, 2014, 37(6): 87-90.

    • 6

      邹权, 邓国栋, 郭效德, 等. 近红外在线检测硝胺推进剂主要固体组分含量[J]. 固体火箭技术, 2008, 31(5): 62-65.

      SU Peng‑fei, ZHANG gao, WEN xiao‑yan, et al. Real‑time detection of the main solid components concentration in RDX‑CMDB with NIR spectral analysis[J]. Journal of Solid Rocket Technology, 2012, 35(6): 838-841.

    • 7

      JIN Ye, WU Zeng, LIU Xue‑son. Near infrared spectroscopy in combination with chemometrics as a process analytical technology tool for on‑line quantitative monitoring of alcohol precipitation[J]. Joural of Pharmaceutial and Biomedical Analysis, 2013, 77(5): 32-39.

    • 8

      GJB8094-2013. 改性B炸药规范[S]. 中华人民共和国国家军用标准.

      GJB8094-2013.Specification for modified composition B[S]. National Military Standards of the People's Republic of China.

    • 9

      严衍禄, 赵龙莲, 韩东海, 等.近红外光谱分析基础与应用[M]. 北京: 中国轻工业出版社, 2005: 125.

      YAN Yan‑lu, ZHAO Long‑lan, HAN Dong‑hai, et al. The basis and application of near infrared spectroscopy[M]. Beijing: Chinese Light Industry Press, 2005: 125.

    • 10

      翁诗莆.傅里叶变换红外光谱分析[M]. 北京: 化学工业出版社, 2009: 232-259.

      WENG Shi‑pu. Analysis of Fu Liye transform infrared spectroscopy[M]. Beijing: Chemical Industry Press, 2009: 232-259.

    • 11

      王镇浦, 周国华, 罗国安.偏最小二乘法及其在分析化学中的应用[J]. 分析化学, 1989, 17(7): 662-669.

      WANG Zhen‑pu, ZHOU Guo‑hua, LUO Guo‑an. Partial least squares method and its application in analytical chemistry[J]. Aanalytical Chemistry, 1989, 17(7): 662-669.