利用高光谱火焰图像同时在线测量三维温度场颗粒浓度场和气体浓度场

In the large scale combustion devices like power station boilers, three-dimensional temperature field, the particle concentration field and the gas concentration field can provide accurate information for combustion adjustment and optimization of control. However, how to measure them is the difficult problem in the combustion measurement techniques. In particular, three-dimensional measurement of gas concentrations currently often used the laser absorption spectroscopy, where lasers and detectors were needed to rotate around the measurement object. Therefore this method can not be applied in the industrial field. The hyperspectral image is a kind of the image including the spatial and spectral information with the huge amounts of data source. A series of two-dimensional image in the continuous wavelength can be obtained by hyperspectral detector. This technique has been used in remote sensing and agriculture industries, but never used for the measurement of the flame. This project will establish a hyperspectral flame imaging system. The spatial dimension and spectral dimension of radiation image information of the flame can be obained via the hyperspectral flame imaging system. The model and algorithm of the radiative transfer inverse problem will be formed based on the flame radiation transfer process and optical imaging principle. The method to calculate the three-dimensional temperature, particle concentration and the concentration of gas according to the flame hyperspectral image will be studied in detail. The flame three-dimensional temperature field, soot concentration field and the oxygen concentration field will be reconstructed in an experimental flame and some validation method will be discussed. The biggest advantage of this approach is the use of the flame self-radiative signal under continuous-wavelength (instead of the external laser) for reconstruction. In the practical application, only several detectors are needed to be arranged in the wall of the boiler to realized three-dimensional reconstruction, with good application prospective.

电站锅炉等大空间燃烧设备内部三维温度场、颗粒浓度场和气体浓度场的重建可以为燃烧调整和优化控制提供精确的信息，也是燃烧测量技术的难点问题。特别是对于气体浓度的三维测量，目前采用的激光吸收光谱方法由于需要激光器和探测器绕测量对象旋转，因此无法在工业现场应用。高光谱图像数据是图谱合一的海量数据源，能够得到一系列波段的二维图像，被应用于遥感和农业等行业，但从未用于火焰的测量。本项目通过建立高光谱火焰成像系统，获取火焰在空间维和光谱维的辐射图像信息，建立根据火焰辐射传递过程和光学成像原理进行辐射传递逆问题求解的模型和算法，研究利用火焰高光谱图像计算温度、颗粒浓度和气体浓度的方法，实现火焰三维温度场、烟黑浓度场和氧气浓度场的同时重建。这种方法的最大优点是利用火焰自身在连续波长下的辐射信号（而不采用外加激光）进行重建，在实际应用时仅需在壁面布置适量的探测器即可实现三维重建，具有很好的应用前景。

本项目通过建立高光谱火焰成像系统，获取火焰在650~1100nm和1200~2450 nm内的光谱图像。建立根据火焰辐射传递过程和光学成像原理进行辐射传递逆问题求解的模型和算法，对于背景辐射、自吸收和火焰组分的散射作用可忽略的火焰，建立相应的简化模型。比较多种求解不完全投影下病态方程的算法，选择阻尼QR分解最小二乘算法作为重建算法。采用Mie理论计算颗粒的辐射特性，得到烟黑、炭粒和飞灰颗粒的简化辐射特性计算公式。分析气体吸收谱带位置，采用统计窄带和逐线计算两种模型计算气体吸收系数。利用气体吸收的选择性，颗粒吸收在不同波长下变化较小，通过计算分离出颗粒吸收作用和气体吸收作用。最终利用火焰高光谱图像计算出温度、颗粒浓度和气体浓度，实现火焰三维温度场、烟黑浓度场和水分浓度场的同时重建。数值模拟结果显示，重建模型对对称火焰和非对称火焰均适用。对于乙烯层流扩散火焰，成功重建出其三维温度场、烟黑浓度场和水分浓度场。重建结果显示火焰外焰是燃烧最剧烈的部位，火焰温度、颗粒浓度和水浓度在外焰高于内焰。重建温度与热电偶测量温度最大相差5%。分别利用激光法和高光谱图像法测量得到的平均烟黑体积浓度变化趋势一致，绝对偏差在0.3~0.7ppm之间。高光谱图像法的最大优点是利用火焰自身在连续波长下的辐射信号（而不采用外加激光）进行重建，在实际应用时仅需在壁面布置适量的探测器即可实现三维重建，具有很好的应用前景。通过4年的研究，掌握了采用火焰高光谱成像同时重建温度场、颗粒浓度场和组分场的模型和算法，建立了火焰高光谱成像测试系统。执行过程中已发表论文20篇（其中SCI杂志10篇），获得省部级一等奖1项，培养研究生11人（毕业5人，在读6人）。已按照计划完成研究目标。