Optimisation of image processing parameters for flame image velocimetry (FIV) measurement in a single-cylinder, small-bore optical diesel engine

High-speed soot luminosity movies are widely used to visualise flame development in optical diesel engines thanks to its simple setup and relatively low cost. Recent studies demonstrated the high-speed soot luminosity movies are not only effective in showing the overall distribution and temporal evolution of sooting flames but also flow fields within the flame through the application of combustion (or flame) image velocimetry. The present study aims to improve this imaging technique by systematically evaluating key image processing parameters based on high-speed soot luminosity movies obtained from a single-cylinder, small-bore optical diesel engine. The raw soot luminosity movies are processed using PIVlab - a Matlab-based open-source code widely used for particle image velocimetry (PIV) applications. The images are pre-processed using the Contrast-Limited Adaptive Histogram Equalization (CLAHE) filter before the velocity vectors are determined through the multi-pass Discrete Fourier Transform (DFT) approach. This is then post processed to find the optimised filter size, interrogation window size, step size and velocity limit for minimal interpolation counts. As a result, a four-step DFT approach with interrogation windows set at 32-32-16-16 pixels, CLAHE filter size at 16 pixels and velocity limit at 3 pixels/frame is determined as optimised. The derived flow fields show that the internal flame pattern change, as demonstrated by previous studies, can be used to obtain the flow vectors within the flame successfully. From a selected example movie, the processed images show the evolution of multiple vortex structures as a result of strong jet-wall and jet-to-jet interactions as well as the swirl effects.