|Alternative Title||Research of the Three-dimensional Combustion Diagnostics Based on Chemiluminescence|
|Thesis Advisor||余西龙 ; 李飞|
|Place of Conferral||北京|
|Keyword||化学发光 3d-ctc技术 重建算法 火焰结构 旋流燃烧|
基于化学自发光的燃烧诊断技术对于发动机燃烧诊断非常重要。由于发动机燃烧室流场较为复杂，具有不稳定、湍流、三维等特性，绝大多数激光诊断技术无法用于精确地描述整个燃烧流场。而火焰的化学自发光与燃烧反应本身密切相关，基于化学发光的三维计算层析 (3D Computed Tomography of Chemiluminescence: 3D-CTC) 技术利用了化学反应产生的激发态自由基成像，结合CT (Computed Tomography) 技术，能够反演获得火焰三维结构和放热率的三维浓度分布。该技术不需要激光源，结构较为简单，成本低且操作性强，适合三维火焰结构以及燃烧的非稳态流场特性研究。对燃气轮机、冲压发动机等恶劣实验环境的复杂燃烧场诊断优势明显，有很强应用前景。
首先，对甲烷预混火焰中的化学发光组份（OH*，CH*，C2*，CO2*）反应机理进行了初步探究。对不同当量比（φ=0.7 - 1.33）下，甲烷-空气预混火焰进行了辐射量化测量，并与一维火焰数值模拟对比。实验中采用波长分辨的光学收集系统，实现对多组份自由基的同时定量；数值模拟中系统分析了现有OH*，CH*的各反应通道和常数，并对比了放热分布与激发态物质（OH*，CH*，C2*，CO2*）的相互关系。研究结果表明，在甲烷-空气层流火焰中，OH*，CH*最合适标识放热率，C2*次之，而CO2*与放热率分布几乎无相关性。在自发光反应机理方面，利用模拟和实验结果，定量对比了不同反应通道的OH*和CH*生成量，评估了两自发光组份的主要生成反应路径。
然后，针对于发动机燃烧室等严苛的诊断环境，提出了大视场角下的焦外模糊成像模型和反卷积处理方案，并与传统的3D-CTC成像模型进行对比，验证实验表明，对于近距离大尺寸火焰的重建，考虑了成像的模糊效应后能够更好的反映实际火焰的结构信息。而且，针对内流火焰诊断中可能存在的遮挡等视角受限情况，进行了成像模型优化，分析了局部不完全成像对重建效果的影响程度。此后，从3D-CTC的层析重建算法着手，引入了基于BP神经网络的三维重建算法，与之前所用ART（Algebraic Reconstruction Technique）算法相比尽管效率较低却准确度高、适应性强，更有前景。
最后，利用发展的3D-CTC技术开展火焰诊断，对燃气轮机模型燃烧室进行了放热率的时空特性研究。主要包括旋流火焰的放热率在时间上的热声振荡现象和空间的三维形态转变两方面。在燃烧形态转变方面，利用3D-CTC技术测量了全局当量比约0.65，总流速2.9 m/s至18.3 m/s的四个工况下旋流燃烧的CH*发光三维分布强度。以此表征放热率的三维分布，实现对旋流火焰放热空间形态转变的研究。实验是通过8个视角下的火焰二维成像来得到其三维CH*分布信息。为验证重建保真度，将重建的三维结果进行二维可视化，并与高速摄影下的二维时均结果进行对比，结果表明重建误差在5%以内。研究中，分析了不同雷诺数下放热率的空间变化规律，所有的实验工况下都显示了V形的附着火焰形态和沿着喷嘴方向的内部回流区。随着雷诺数的增加主要放热区域扩展明显且最大放热率随着流向移动。在旋流燃烧的热声震荡方面，利用CH*的二维高速摄影，对旋流燃烧的放热率不稳定性进行研究，发现放热率的振荡频率随着雷诺数的增大逐渐增加
Combustion diagnostics which based on the chemiluminescence is very important of the study of engine. Because of the turbulent and three-dimensional properties, most laser diagnostic techniques cannot be used to accurately describe the entire combustion flow field. 3D Computed Tomography of Chemiluminescence (3D-CTC) which is based on the chemiluminescence of flames combined with the Computed Tomography techniques can get the three dimensional structure of flames and the distribution of heat release rate. It is simplicity, low cost and without the need for lasers to be subjected to the study of 3D flame structure and unsteady flow field of combustion. It has obvious advantages in the diagnosis of complex combustion fields in harsh conditions such as gas turbines and ramjet engines, and has a strong application prospect.
This dissertation carries out a systematic research on the 3D combustion diagnostics based on the chemiluminescence: a preliminary research of chemiluminescence mechanisms is conducted by the quantitative measurements and numerical modeling of chemiluminescence in laminar methane-air premixed flames. A multi-directional imaging system for 3D-CTC was developed which uses only one CCD camera combined customized fibers to reconstruct an steady flame that makes the 3D-CTC technique more convenient. For the application aim for engine diagnostics, an out-of-focus imaging model is proposed over a large field angle which taken the bokeh effect into account, also an improved algorithm based on BP neural network is proposed in the method of 3D reconstruction. An application of combustion diagnostics for the gas turbine model combustor based on the improved 3D-CTC technique is executed that the effect of the fuel-air flow velocity on heat release rate of swirling non-premixed methane flames are analyzed.
Firstly, the reaction mechanisms of chemiluminescence species (OH*，CH*，C2*，CO2*) was preliminary studied. the emission of chemiluminescence were quantitative measured on the methane-air premixed flames by the wavelength-resolved detection system with optical-path-corrected emission calibration method. Numerical methods based on the one-dimensional flames simulations which cover the experimental equivalence ratios from 0.7 to 1.33 combined with various rate coefficients for species OH* and CH* were then conducted. The numerical results of the heat release rate and distributions of chemiluminescence shows that the OH* and CH* are good indicators than C2*, while the CO2* distribution has the worst correlation with the heat release. Third, quantitative comparisons were made for the formation of OH* and CH* based on the simulations and experimental results, and then the main formation channels were evaluated for these two species.
At the same time, a simple experiment on the flat flame burner was conducted due to a principle of 3D-CTC which using the traditional parallel projection mode. Reconstruction accuracy was analyzed at different viewing angle and imaging noise. For the limitation of parallel projection, some reconstruction experiments were made which using the lens clear-imaging model, in those experiments an improved system were utilized that by the combination of one CCD camera and customized fibers, the corresponding coordinates calibration method was also proposed, and a good reconstruction accuracy was found by the the analysis of the size and shape of the reconstruction results.
Then, because of the practical purpose of 3D-CTC technique, the bokeh effect of large field angle was analyzed by the utilization of out-of-focus imaging model and deconvolution models. After that, focusing on the tomographic reconstruction algorithm, a reconstruction method based on BP neural network was introduced in the 3D-CTC technique. Compared with the ART (Algebraic Reconstruction Technique) algorithm the neural network algorithm has some unique advantages although high time costing.
Finally, investigations of a gas turbine model combustor flames were taken by using the 3D-CTC technique. Both the temporal phenomenon of thermos-acoustic and the spatial flame structure transition were analyzed that, in the flame structure transition, the improved 3D-CTC technique was utilized for the measurement of 3D emissions of CH* and taken as qualitative indicators of the heat release rate under four velocity conditions (2.9-18.3 m/s) with an overall equivalence ratio of about 0.65. 8 multi-directional CH* images were used as inputs to compute the 3D distribution of CH* intensities. In order to verify the reconstruction fidelity, the two-dimensional visualization of the reconstructed shape is compared with the time-averaged projection under high-speed photography, and the results show that the reconstruction error is within 5%. In this study, the transitions of heat release area with Reynolds number were analyzed, and for all conditions, pronounced extension of inner recirculation zone (irz) along the nozzle is observed under the attached (V-shaped) swirl stabilized flames. During the increase of Reynolds number, the heat release zone changing obviously along the nozzle radial and axis direction, and the largest heat release plane moves forward significantly. In addition, an intensified high speed camera was adopted for the heat release dynamics study. Large oscillations in the flame zone appeared that significantly affected the total heat release oscillations, and the fluctuate frequency of increases with the increase of the Reynolds number.
|王宽亮. 基于燃烧化学自发光的三维诊断技术研究[D]. 北京. 中国科学院大学,2019.|
|Files in This Item:|
|博士论文-王宽亮.pdf（7049KB）||学位论文||开放获取||CC BY-NC-SA||Application Full Text|
|Recommend this item|
|Export to Endnote|
|Similar articles in Google Scholar|
|Similar articles in Baidu academic|
|Similar articles in Bing Scholar|
Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.