Tomographic Reconstruction Of Three-Dimensional Luminous Flame Temperature Field Based On Refocused Light Field Image

High-temperature combustion exists widely in the fields of aerospace,energy and power,and occurs in high-temperature equipment such as aerospace engines,internal combustion engines,power plant boilers,gas turbines,etc.Theoretical and experimental research on high-temperature combustion phenomenon will optimize the design of the combustion system and monitor the status of high temperature equipment.The flame temperature is an important item among the parameters of describing the flame state.The real-time and accurate acquisition of the flame temperature distribution is of great significance for the understanding of the combustion process and the analysis of the combustion theory.However,the flame usually burns unsteadily,the measurement methods are difficult to detect the instantaneous internal temperature field of the flame.Therefore,the accuracy of on-line monitoring of flame temperature field still needs to be discussed.To achieve this goal,the non-contact flame temperature measurement method based on radiation imaging method can be used,so as to realize the continuous online measurement of the temperature field of a large-space luminous flame.The light field camera(plenoptic camera)is a kind of non-contact optical detection equipment based on radiation imaging method.It can record the multi-angle radiative information of the subject only through a single shot,which can simplify the design and debugging of the flame measurement system.The application of the light field camera to flame measurement is a promising online temperature-field-detection method.In this thesis,a layered and a non-uniform flame light field imaging model is established based on the Monte Carlo method.According to the characteristics of different flame layers,different reconstruction methods are used to reconstruct the refocused light field image.By presetting the uniform distribution of radiative properties,the temperature field can be reconstructed.Compared with the temperature field of the input model,the threedimensional distribution of relative error can be obtained,which verifies the algorithm.The accuracy range of temperature reconstruction under unknown radiation properties is studied.It provides theoretical model,solution idea and technical support for on-line obtaining the flame temperature distribution.The main work is as follows:A light field imaging model is built.A quality evaluation system for the flame light field images is established and the images under different radiative properties are compared.By simulating the images of the flame with different attenuation coefficients,the effects of the attenuation of the light energy emitted by the flame on its light field image can be analyzed.For complex flame structure,the axisymmetric flame light field imaging model with the non-uniform distribution of the temperature and radiative properties is developed.A complete simulation of luminous flame light field radiation transmission process is also implemented.Input parameters,such as the distribution of the flame temperature and radiative properties,to the flame light field imaging model,the flame light field images can be generated,the corresponding subaperture images and refocused images can then be obtained.It provides a model basis for the reconstruction of flame temperature field using light field images.Based on the flame light field imaging model,a flame-temperature reconstruction platform is established.The three-dimensional reconstruction of the flame temperature field is processed.The simulated calibration for the temperature reconstruction platform is implemented.The point spread function is also calibrated.The wiener filter method is applied to the refocused image of different flame layers to obtain the reconstructed image.The corresponding relationship between gray scale and actual spectral radiative force is used to reconstruct the flame temperature,and then the relative error distribution is calculated.Using the flame-temperature reconstruction platform based on the light field imaging process,the temperature reconstruction algorithm is discussed.Through the rational utilization of the refocused light field image,different reconstruction methods are used to calculate the temperature field according to the characteristics of different flame layers.The nearest neighbor filtering method,the wavelet image processing method and the deconvolution algorithm are introduced.The nearest neighbor filtering method and the deconvolution algorithm are compared by the temperature reconstruction results of different flame layer,and both are integrated on a particular flame layer.Then,the nearest neighbor filtering and the deconvolution joint algorithm is put forward.Continually,using the flame-temperature reconstruction platform based on light field imaging,and the nearest neighbor filtering and the deconvolution joint algorithm,the influence of uniform and non-uniform distribution of flame radiative properties and the location of flame imaging on temperature reconstruction accuracy are analyzed.The uniform attenuation coefficient distribution is used to estimate the flame temperature under different attenuation coefficient distributions,the relative error range of the reconstructed temperature is investigated.The influence of the non-uniform attenuation coefficient distribution on the temperature reconstruction is analyzed through the regional division of the medium.Under the premise that the flame size and the effective imaging area do not match,the adaptability of the light field camera to different flame sizes is investigated and evaluated.The computational fluid dynamics simulation results are input to the light field imaging model for numerical simulation of ethylene laminar diffusion flame and the generation of its light field image.The flame light field images are then captured through the experiments,the refocused image is used to calculate temperature reconstruction results,verifying the temperature reconstruction method of the ethylene laminar diffusion flame.Then the light field image of the dimethyl ether laminar diffusion flame is measured and calculated,with soot volume fraction one order of magnitude lower than the ethylene flame.The reconstructed temperature is also acquired.The light field imaging results of the flame and its temperature reconstruction result with less soot particles are investigated.Based on the above research,a simulation model of flame light field imaging is built in this thesis,which considered the non-uniform temperature field and radiative properties field.The numerical simulation of the refocused light field imaging process of the high temperature luminous flame is realized,and the imaging characteristics of the flame with different depth of field are obtained.Since there are limitations of the nearest neighborhood deconvolution algorithm when dealing with the edge and center temperature of flame,a joint algorithm for temperature reconstruction of layered flame characteristics is proposed.Efficient on-line reconstruction of three-dimensional flame temperature field is realized.A light field tomographic temperature reconstruction strategy for luminous flame with unknown radiation properties is proposed.The correctness of the reconstruction strategy is verified by experiment and simulation,and the strategy is applied to the reconstruction of temperature field of different types of coflow laminar diffusion flame.Above research provides the simulation model and data basis for on-line measurement of flame temperature field,and the solution idea and technical support for three-dimensional reconstruction of flame temperature field.