Research On Determination Of Radii Of Spherical Flames And Computer Realization

Due to wide range of experimental conditions, outwardly propagating sphericalflames were widely in researches on the extraction of laminar burning velocities,Markstein lengths and other important parameters. A large number of flame imageswere obtained by using high-speed schlieren camera system in experiments. And thenlaminar burning parameters can be determined by the corresponding relation betweenflame radii and propagating time. Efficiency and accuracy should be mainlyconcerned in this data processing. Instead of artificial processing, radii determinationby computer can significantly improve the efficiency. However, to guarantee theefficiency, several aspects should be involved and the choosing of available radii isparticularly important. Actually, the choosing of available radii was mainly affectedby flame cracks and instabilities, for which the application of fractal theory canprovide meaningful analysis. Therefore, the present study aimed to investigate twoparts: the one was the new approach of radii determination of spherical flames; theother was the calculation of fractal dimension D3of spherical flames to provide someguidance for available radii choosing. The main conclusions of the present study aresummarized as follows:(1) Gray images with little noise were acquired through graying, framedifference and median filtering. Binary edge were obtained by using Canny edgedetection and ray arc approach. The results of comparing square approach andRandomized Circle Detection(RCD) approach showed that flame radii determined byRCD approach were little higher than artificial results, radii determined by squareapproach were very close(the error was within1mm) to artificial results.(2) The result of the comparison of box-counting approach, circle approach andcaliper approach showed that box-counting approach had better resonableness andconsistency. The flame cracks and instabilities were investigated by analyzing D3.The conclusions resulted from D3was in good consistency with other relatedconclusions. Flame instability analysis based on D3had therefore feasibility.(3) The software Spherical Flame Processor(SFP) was developed under theenvironment of Visual Studio2008+OpenCV2.3.1. A large number of schlieren images of spherical flames with different resolution can be well processed every timeand the calculating results of the relation of time-number-radii-D3can be exported toa excel file. Moreover, the errors of SFP resulted from theoretic errors andexperimental errors: theoretic errors included noise and algorithm, while experimentalerrors included schlieren performance, image resolution and the cleaning of quartzglasses.