Measurements of flame-front curvature based on Fourier transformation

An optimized image processing method for calculating the local curvature of two-dimensional digitized flame-front contours is proposed. This contour-smoothing method is based on Fourier transformation of the entire contour and is different from conventional methods in which a chosen number of neighbouring pixel points is locally fitted to an ad hoc smooth function along the contour on a point-by-point basis. The maximum curvature that can be measured with the proposed method is determined by the inner cut-off scale from a fractal length measure of the flame-front contour and is limited by image spatial resolution only. Curvature measurements are also made in highly stretched turbulent premixed methane–air jet flames. The measured probability density functions of local flame curvature are consistent with the generalized form of a symmetric Gaussian distribution suggested by Bradley et al. (Combustion and Flame, 135, 503–523 (2003)). Advantages and limitations of the proposed method are also discussed.