Three-dimensional Temperature Measurement Of Flame Based On Light Field Imaging

Combustion is widely applied in daily activities and industries such as steel,metallurgy,power plants,and aerospace.The investigation of the three-dimensional temperature(3D)field reconstruction of flames would support many studies such as the investigation of the nature of combustion chemistry,adjusting the combustion mode,optimizing the combustion process,and controlling combustion pollutants,and therefore,it is significant and valuable in science and practical applications.Existing flame-temperature measurement methods and systems are complex in terms of system setup and installation,and also limited in terms of 3D radiation sampling and reconstruction.Therefore,3D temperature measurement methods for the flames are proposed based on the light field imaging technique in this study.A systematic investigation is carried out theoretically and experimentally.Tracing methods for the radiation emitted from flames based on light field imaging are proposed.The principles of calculating the radiative transfer process inside the flame and the intensity of radiative rays are introduced.The technique of light field imaging and radiative transfer model are combined to develop the mathematical model of radiative light field imaging.Light field images of flames are calculated for different types of light field cameras and different radiative properties of flames.The concepts of effective pixels and utilization rate of pixels are proposed for the light field images of flames to evaluate the capacity of radiation sampling.The refocused images of flames at different depths are also calculated and compared.The simulatioin results indicate that the light field camera can be applied to sample the 3D radiation information of the flame.A slight difference can be observed among the refocused images at different depths inside the flame,which is due to the limited volume of the flame.The strategy and method to reconstruct the 3D flame temperature field based on the light field imaging is proposed.Linear optimization algorithms of LSQR(least squares via QR factorization,LSQR)and NNLS(non-negative least squares)and non-linear optimization algorithms of Levenberg-Marquardt are described to solve radiative transfer equations in the 3D flame temperature field reconstruction bead on the light field imaging.On this basis,a hybrid algorithm of LMBC(Levenberg-Marquardt method with Boundary Constraint)and NNLS is proposed for the unknown absorption coefficient in radiative transfer equations.The LMBC-NNLS algorithm is applied to reconstruct the temperature and absorption coefficient of the flame simultaneously.In numerical simulations,four cases of different distributions of temperature and absorption coefficient are set.For the four cases,the light field images of the flames are calculated and the 3D flame temperatures are reconstructed.The results indicate that the LMBC-NNLS algorithm is applicable to the simultaneous reconstruction of the temperature and absorption coefficient.The relative errors of the four cases are all less than 0.1.The LMBC-NNLS algorithm is accurate and reliable.A single pixel is considered as one sampling unit.On this basis,the concepts of sampling region(SR),sampling angle per unit(SAPU),and sampling angle(SA)are proposed to evaluate the directional accuracy of the radiation sampling.The concept of the sampling ray is defined.The distributions of the sampling rays for different distances between the microlens array and photosensor and different microlens arrays(single focal length and multiple focal lengths)are compared inside the flame.The light field flame images and reconstruction results of different parameters are analyzed and compared.Compared to light field camera 2.0,the distributions of sampling rays of light field camera 1.0 are less homogeneous and therefore are disadvantageous to the temperature reconstruction of flames.Compared to the microlens array with multiple focal lengths,the distributions of sampling rays are more homogeneous,and therefore,the reconstruction results are better for the microlens array with a single focal length.The measurement system of the 3D flame temperature field bead on light field imaging is developed.The basic performance of the elements in the measurement system is evaluated.A novel calibration method for the focused light field camera is proposed.The calibration model is developed based on the light field images.The specific parameters to be calibrated are further calculated according to the characteristic of the matching F-numbers.The calibration results are compared and validated using total focused images.The experiments for intensity calibration are investigated and the errors of the intensity calibration are calculated.The reprojection errors of the geometric calibration are below 7 pixels.The method to calibrate the focused light field camera is feasible.The experimental evaluation and practical application of the 3D flame temperature field reconstruction bead on light field imaging are carried out.A co-flow burner and combustion system are designed and fabricated.The measurement system of 3D flame temperature field bead on light field imaging is utilized to capture the light field images of flames.The 3D distributions of the temperature and absorption coefficients of ethylene flames are reconstructed using the proposed hybrid LMBC-NNLS algorithm.The reconstruction results using the LMBC-NNLS algorithm and thermocouple results are compared.A good agreement is observed between the two results,and the maximum difference is 150.9 K.The accuracy and reliability of the method to reconstruct the 3D flame temperature field bead on light field imaging are validated.The main error sources of the reconstruction results are analyzed to evaluate the accuracy.The practical application of the 3D temperature field reconstruction of biomass flames on a fluidized bed is investigated.The results demonstrate that the measurement system bead on the light field imaging can be utilized for the temperature field reconstruction of biomass flames.Novel investigation methods were provided for visualization experiments on the combustion of biomass or coal particles.