Application Of Quanttative Schlieren Method In Flame Temperature Measurement

As a non-intrusive flow visualization method, schlieren has been existing for many centuries. It is widely used in the area of combustion, fluid dynamics and heat transfer. However, there are many factors that can influence the result of quantitative schlieren. In most experiments, schlieren has been used as a qualified method. “Z” type schlieren, rainbow schlieren and background-oriented schlieren(BOS) are three typical schlieren systems. Recently, the development of computer science, especially the digital image processing technique, quantitative schlieren method seems to be possible.Based on axisymmetric assumption, quantitative schlieren measurement of flame temperature involves computed tomography technique, digital image processing, Gladstone-Dale relationship and idea gas law. Unlike “Z” type schlieren, BOS needs to calculate pixel shift of image rather than a calibration of the system to obtain deflection angle.The application of quantitative schlieren method in flame temperature measurement was studied in this paper, and related work as following:1. The involved tomographic algorithms were compared under axisymmetric assumption. They are direct integral method, two-point Abel inversion, Radon inversion and algebraic reconstruction technology. The results shows that direct integral method and Abel inversion can both achieve the least error, otherwise, Radon inversion strongly depends on the filter function and algebraic reconstruction technique depends on the number of discrete points. Three dimensional algorithms included Radon inversion and algebraic reconstruction technology did not show the advantage under axisymmetric assumption in two dimensional measurement.2. Standard photometric method was used in “Z” type schlieren for calibration purpose to obtain deflection angle. The relationship between grayscale value of flame schlieren image and deflection angle was established through a calibration lens. The refraction index was obtained by tomographic algorithms. Then the temperature of methane flame was obtained by combining Gladstone-Dale relationship and ideal gas law. The temperature reconstructed by quantitative schlieren technique was compared to thermocouple measurement and the results show the validation of this quantitative technique.3. Optical flow algorithms based on variational methods and cross-correlation methods were compared through Yosemite image sequence and particle image sequence. Horn-Schunck method and Lucas-Kanade method achieved better and smoother pixel-shift filed than cross-correlation method when using a multi-scale wavelet noise background in experiment. The set-up of BOS system was introduced in this paper. The deflection angle was obtained by calculation of pixel shift using Lucas-Kanade variational optical flow method. Then combined with Gladstone-Dale relationship and ideal gas law, the section and slice of premixed methane flame temperature profile was obtained. The sensitivity and error source of this system were also analyzed.