Temperature And Concentration Measured By Holographic Interferometry For Non-premixed Flame

The combustion researches include the theoretical, numerical and experimental studies. Certainly, these studies are performed seldom through single means, these means usually combine with, validate and promote each other. This dissertation combines numerical simulations with experimental study. Basing on the result of numerical simulation, the systematic theoretical schemes are developed. The temperature and concentrations for non-premixed flames are measured by holographic interferometry (HI) with these schemes.Laser HI can be employed to measure temperature and density. It can be used to implement non-intrusive field measurement with high space and time resolution. So it is applied widely. This dissertation simulated the process that the axially-symmetric temperature fields are measured by HI. This dissertation compared the reconstructing effects of four methods for temperature fields with single peak and double peaks. The results indicate that the concentric circles method and Abel conversion method can apply rather to complex temperature field, but their reconstructed point number and positions are limited. The curve fitting method is always affected by adopted power, but can give continual distribution of temperature and produce better reconstructing result for temperature field with single peak. This dissertation promotes the traditional concentric circles method with equally partitioned space. The concentric circles are placed in the middle of two adjacent fringes so that the linear equation group obtained has the exclusive resolution and can be more easily solved. The simulation for temperature field with two peaks indicates that the accuracy and effect of reconstruction are promoted, the reconstructed points are dense where the gradient of temperature changes greatly and the fringes are dense so the points can reflect the change of temperature much more. This promotion for the traditional concentric circles method has not been found in other literatures.This dissertation simulated the ethylene-air non-premixed flame with detailed numerical model. The cross-grids are adopted for velocity fields. The SIMPLE arithmetic widely applied is employed. The models with more than one equation are examined with simulation results. The result indicates that the model of four equations much more approaches practical conditions. This dissertation proposes similarity principles of composition distributions, which are examined with numerical simulation results. The result with the principle of double areas exhibits good similarity of concentration distributions between oxygen and nitrogen, between fuel and nitrogen.This dissertation develops two sets of systematic scheme basing on the result of numerical simulation, solves the puzzle that local composition must be known when temperature of flame is measured by HI. The solution for the puzzle is the origination at home and abroad. This assumption of composition of air will bring large errors for diffusion flames. The study of methane-air diffusion flame by Xiao X. et al shows that this error reaches 34%. This investigation indicates this error arrives at 48.8% for ethylene-air diffusion flame. With one of two sets of systematic scheme, HI can not only be used to measure temperature of diffusion flame accurately so that this error can be eliminated, but also HI can be used to measure concentrations, mixture fraction etc. The idea of the first scheme, which is simple, is linear. Firstly, the variable state relationship between temperature and refractive index is developed. With this relationship temperature can be computed directly from refractive index without composition concentrations. This relationship includes two interconnected lines. The under line is fixed, originates in the state point of ambient air with the gradient of slightly less than–1 and represents the distribution outside flame surface. The upper line varies, originates in the under line. The line or extended line assembles in the state point of fuel at the exit of fuel nozzle. So the upper line can be determined and represents the distribution inside flame surface. The temperature of ethylene-air diffusion flame is measured by HI with the relationship. The result indicates that this relationship can reduce the error led to by assumption of composition of air from 48.8% under 4%. Then, the normalized state relationship is obtained between temperature and mixture fraction. Using this relationship, mixture fraction of diffusion flame is calculated from measured temperature. Finally, concentrations of main species are computed from mixture fraction with the generalized state relationships for diffusion flames.The second scheme is appreciably complex, but with it we can gain the more accurate temperature and the G-D constant. The linear relation between G-D constant and mixture fraction is obtained in the process developing the scheme. The basic thought is that G-D constant related to compositions is inferred also from refractive index indirectly through the middle variable, i.e. mixture fraction. Firstly, according to the similarity of sectional distributions, the normalized state relationship is gained between mixture fraction and refractive index for diffusion flames. The relationship exhibits linear distribution and slightly scatters inside the flame surface. Then the nearly linear state relationship is discovered between G-D constant and mixture fraction. Basing on principle of the physical optics, this linear relation between G-D constant and mixture fraction is deduced using the Lorenz-Lorentz equation and the definition of mixture fraction. The relations also apply to other flames, since there is not the limit of diffusion flame in the process deducing them. The analysis about this relation has not been reported at home and abroad. Using these developed state relationships, HI has been used to measure mixture fraction and G-D constant of diffusion flame. Finally, temperature is computed from the inferred G-D constant and measured refractive index. The result indicates that the systematic method can reduce this error brought by the assumption of composition of air from 48.8% under 1.6% which is surprisingly small. Also, concentrations of main species are inferred from obtained mixture fraction with generalized state relationships.