Turbulent combustion data analysis using fractals

This paper investigates several types of data analysis, based upon fractal geometry concepts, using time series generated in turbulent combustion research. The techniques are quite general and may be used for other turbulent flows. Investigated are the generalized fractal dimension, multifractal probability density function, fractal filtration, multifractal spectrum, and fractal interpolation and its possible connection with chaotic dynamics. It is concluded that several of the techniques are useful for 1) new visual depiction of the data, 2) discrimination of portions of data traces as noise-contaminated, 3) separation of wanted and unwanted highor low-frequency events, and 4) interpolation between sparse data points in either short run time or low data acquisition rate situations.