电子鼻在危险爆炸物检测中的应用研究

Research on Explosives Detection by Electronic Nose

利用由18个纳米氧化锌厚膜气敏传感器组成的阵列对硝铵、矿山炸药、苦味酸、2,4二硝基甲苯(DNT)这4种典型爆炸物样品进行了测量.采用动、静态相结合的采样方法考察了传感器阵列的检测能力,在动态实验中通过提取不同的特征值并利用主元分析(PCA)和聚类分析(CA)方法对数据进行了分析和识别.静态实验结果表明传感器阵列在不同浓度上对4种典型爆炸物均有不同程度的响应,该方法能检测到硝铵、矿山炸药、苦味酸的浓度低至3.34μg/L,DNT为83.3μg/L;动态实验结果表明提取极值为特征值对阵列进行PCA、CA分析,可使4种典型爆炸物在毫克级上能完全区分.以上结果说明电子鼻技术在危险爆炸物检测中是一种很有发展前途的实用技术.

Four kinds of representative explosives were measured by the gas sensor arrays, which were composed of eighteen doped nano-ZnO thick film sensors. Using static and dynamic sampling methods to exam the detection ability of the sensor arrays. Principal Component Analysis （PCA） and Cluster Analysis （CA） were used in the data analysis and pattern recognition. Static sampling method shows that the sensor arrays are sensitive to the four explosives with different concentrations. The results show that the detection concentration of NH4NO3, mineral explosive and picric acid are low to 3.34μg/L and that of DNT is low to 83. 3μg/L under the laboratory conditions; Dynamic sampling method shows that all the samples can be identified completely in milligram level when extracting extremum as the feature to PCA and CA analysis. This work indicates the potential applications of the electronic nose for analyzing and identifying the explosives.