3-D Temperature Measurement Of Flame Based On Multi-camera Light Field Imaging Technique

Combustion is widely applied in power plant boiler,fluidized bed,engine,gas turbine and industrial production of cermet materials.Flame temperature is the research basis of the combustion mechanism and the diagnosis basis of the burning state.It's also an important basis for the optimization design of combustion devices.The reconstruction of three-dimensional(3D)temperature field for combustion flames helps to achieve the high efficiency,low pollution and intelligent utilization of energy,which is of great significance.Existing flame-temperature measurement methods and systems are complex in terms of system installation and operation,poor in real-time performance and reconstruction precision for flames with complex structure.Therefore,a 3D temperature measurement methods of flame is proposed based on multi-camera light field imaging technique in this study.A systematic investigation is carried out theoretically and experimentally.The light field imaging principle of flame radiation and the response model of CCD detection surface is studied,and the tracing method for radiations emitted from flames is then proposed.The light field imaging technique and radiative transfer model of internal flame are combined to develop the model of multi-camera radiative light field imaging.The intensity of radiative rays corresponding to radiation field images are calculated.The concept of detection matrix weight uniformity is proposed to evaluate the sampling characteristics of this multi-camera light field system.The effects of flame's physical properties and camera's configuration parameters to sampling characteristics are then analyzed.The optimized configuration scheme of multi-camera light field imaging system is realized by quantum particle swarm optimization algorithm(QPSO)and provides design basis for the temperature measurement system.The results indicate that more cameras and more dispersive there are arranged?more flame elements and better distribution of optical thickness lead to better sampling uniformity.The more complex the flame structure is,the more light field cameras are needed and the worse the sampling uniformity which can be achieved.Excessive optical thickness will result in inefficient sampling zones,therefore light field cameras should be arranged in the view of reducing detection zones,which will improve the reconstruction accuracy of the temperature field of flams.An optimized reconstruction algorithm of 3D temperature field of flames and a strategy for filtering sampling rays are proposed.The conventional 3D reconstruction algorithm is improved and the arrangement scheme of light field cameras is optimized.Numerical simulations are carried out to verify the feasibility of reconstruction method and optimum arrangement scheme of sampling characteristics.The simulation results indicate that reconstruction effect of 3D temperature field is related to the temperature distribution,the distance from cameras,and the reconstruction algorithm.The optimization algorithm has better anti-noise performance ? satisfies with non-negative conditions and has stronger reliability than LSQR algorithm.Temperature measurement system with single light field camera is unable to reconstruct the 3D temperature field of non-centrosymmetrically distributed flame accurately,therefore multiplecameras need to be introduced.The reconstruction accuracy can be significantly improved by optimizing the sampling angles based on detection uniformity.A temperature measurement system with two light field cameras after optimizing its configuration scheme is applied on non-centrosymmetrically distributed flame.The maximum relative error of reconstructed temperature is less than 0.05%,which is of good agreement with set value.It illustrates that the proposed optimized reconstruction algorithm?sampling rays filtering strategy and the arrangement scheme of light field cameras are accurate,reasonable and reliable.The experimental research of the 3D flame temperature field reconstruction methods based on multi-camera light field imaging technique is carried out.A co-flow burner with non-centrosymmetric flame nozzles is designed and fabricated.The ethylene laminar diffusion flame diffusion flame is used as object.The multi-camera light field imaging temperature measurement system is composed of two light field cameras.The fitting relationship between gray level and radiation intensity is obtained by radiation intensity calibration of the light field cameras.The spatial optical axis alignment and synchronized trigger settings of the two light field cameras are carried out.The light field images are captured from different conditions and different visual angles to reconstruct the flame 3-D temperature field.The reconstruction results using thermocouple are compared to evaluate the measurement accuracy.These two results are of great agreements and the maximum difference rate is less than 7%.The experimental results illustrate the accuracy and reliability of 3D flame temperature field measurement methods based on multi-camera light field imaging technique in this study.