Improving droplet sizing methodology for spray dynamics investigation

Abstract

Spray modelling is one of the most useful techniques to characterize engine performance, efficiency and emissions. The size of droplets is one of the key variables that govern the efficiency of combustion of the liquid fuel. This study aims to develop an interactive tool using MATLAB codes that identifies the droplets and their sizes from the image taken with the long distance microscope in the spray chamber setup. In this developed method, firstly the background of the image was removed and then image processing techniques, dilation and erosion, were applied to the image file to refine the image files. Subsequently, circle detection method based on the Hough Transform algorithm with the function of imfindcircles was implemented. This function of the program allows the user to identify size droplets from the image files. A statistical study was conducted with the results automated from the MATLAB program using a different set of threshold values of black and white contrast. The results showed an optimal range for the threshold (black and white) values between 40 and 70. This optimal threshold range was established based on consideration of the correct and incorrect identification of the droplets. The results indicated that the program has the ability to identify the droplet providing size and numbers. The MATLAB program was developed using MATLAB compiler and can be used at different workstations.

Multi-particle suspension in a laminar flow agitated by a Rushton turbine

“…The raw images obtained by the two cameras were processed and analyzed by a code in the matlab environment (version: R2017, MathWorks, USA) to determine the centroids of each particle as a function of time. In this study, the matlab function imfindcircles (the MathWorks Inc., 2017; Woo et al, 2016) based on a Circular Hough Transform (CHT) (Davies, 2005) algorithm was used to find all the circles in the images (Atherton and Kerbyson, 1999). The analysis process required a number of input parameters: the range of pixels over the radius of a particle was given first.…”