乙腈广泛应用于医药、 化工等领域, 而乙腈属于易燃易爆化学品, 其引发的火灾事故具有极大的危害。 研究乙腈燃烧的温度场与浓度场、 火焰辐射光谱以探究其火灾污染特性具有重要实用价值。 首先采用平面激光诱导荧光技术（PLIF）与Fluent数值模拟方法, 获取了5 cm尺度乙腈池火燃烧产物NO在20、 40、 60和80 s时刻的空间浓度值, 并结合CFD与FDS仿真模拟获取了不同时刻下乙腈燃烧温度场与浓度场信息。 其次, 采用所获取的乙腈火焰温度场和浓度场数据（将火焰划分为6个热力学平衡区域）, 并基于HITRAN数据库内高温气体分子吸收系数与火焰总体辐射传输方程构建了乙腈火焰光谱辐射模型。 再次, 将所得乙腈浓度场与温度场数据代入火焰光谱辐射模型, 模型模拟计算结果与相同条件下乙腈火焰光谱实测数据进行对比, 以验证模型精度, 然后再与Radcal模型进行精度对比。 最后, 利用自行构建的火焰光谱辐射模型对燃烧特征污染产物NO进行了浓度反演。 结果表明: （1）5 cm尺度乙腈池火火焰温度范围为400～1 000 K, 在池火上方60～80 mm区域温度较高, 最高温度为945 K。 （2）在20、 40、 60和80 s时刻下5 cm乙腈池火燃烧产物NO的体积分数为0.005%～0.025 5%, H2O的体积分数为0.034 5%～0.062 5%, CO2的体积分数为0.055 5%～0.085 5%。 （3）自行构建了乙腈火焰光谱辐射模型, 模型模拟值与实测值对比得出, 燃烧产物中CO2特征峰处准确度最小为86.8%, 最大为88.7%; NO特征峰处准确度最小为79.6%, 最大为84.9%; H2O特征峰处准确度最小为84.6%, 最大为89.1%。 与Radcal模型计算的光谱辐射值进行对比, 自行构建的模型计算精度提升约10%。 （4）在5.62～5.66 μm主导波段, 乙腈燃烧特征产物NO在20、 40、 60和80 s时刻下的浓度反演精度分别为76.9%、 78.5%、 94.7%和81.3%。 此研究可为探测大尺度乙腈类化学品火灾的燃烧场信息以及遥感定量反演燃烧污染产物浓度提供基础与参考。

Acetonitrile, widely used in pharmaceutical and chemical industries, is a flammable and explosive chemical which can cause fire accidents with great harm. It is of great practical value to explore the fire pollution characteristics of acetonitrile combustion by studying its temperature and concentration fields and flame radiation spectra. In this paper, the spatial concentration values of NO, a product of acetonitrile pool fire, at the 20, 40, 60, and 80 s on a 5 cm-diameter scale were obtained using Planar Laser Induced Fluorescence (PLIF) and Fluent numerical simulation methods, and the temperature and concentration fields of acetonitrile combustion at different times were obtained by combining CFD and FDS simulations. Secondly, data from the temperature field and concentration field of the acetonitrile flame (the flame was divided into six thermodynamic equilibrium regions) were used to construct an acetonitrile flame spectral radiation model based on absorption coefficients of high-temperature gas molecules and overall radiative transfer equation of the flame in HITRAN database. Again, data from the concentration field and temperature field of the acetonitrile flame were substituted into the flame spectral radiation model, and the model simulation results were compared with the measured acetonitrile flame spectral data under the same conditions to verify the model accuracy and compare with the Radcal model. Finally, the concentration inversion of NO, a characteristic pollution product of combustion, was performed using the self-built flame spectral radiation model. The results showed that: (1) the flame temperature range of 5 cm-diameter acetonitrile pool fires was 400～1 000 K, and the temperature was higher in the region of 60～80 mm above the pool fire with the highest temperature of 945 K; (2) the volume fractions of combustion products of 5 cm-diameter acetonitrile pool fire at 20, 40, 60 and 80 s moments were 0.005%～0.025 5% for NO, 0.034 5%～0.062 5% for H2O, and 0.055 5%～0.085 5% for CO2; (3) an acetonitrile flame spectral radiation model was built by ourselves, and comparison between the model simulation value and the actual measured value showed that: in combustion products, the CO2 characteristic peak accuracy was 86.8% min and 88.7% max; NO was 79.6% min and 84.9% max; and H2O was 84.6% min and 89.1% max. Compared with spectral radiation values calculated by the Radcal model, the calculation accuracy of our model was improved by about 10%; (4) the inversion accuracy of the concentration of NO, the characteristic product of acetonitrile combustion, in dominant band of 5.62～5.66 μm at moments of the 20, 40, 60, and 80 s was 76.9%, 78.5%, 94.7%, and 81.3%, respectively. This study can provide a basis and reference for detecting combustion field information of large-scale acetonitrile chemical fires and the quantitative inversion of combustion pollution product concentrations by remote sensing.