Research On Real-time Flow Visualization Based On Background Oriented Schlieren

Flow visualization is an important means to solve practical engineering problems,and it has wide applications in meteorology analysis,aerospace and other fields.Various flow visualization and measurement experiments can make people understand the complex fluid flow phenomenon,explore the physical mechanism and motion law,and provide scientific basis for the establishment of new concepts and mathematical models.Background Oriented Schlieren(BOS)has become the mainstream streaming flow visualization technology because of its simple installation,low cost,large field of view and suitable for the complex environment measurement.Real-time flow visualization based on BOS is studied in this paper,the main contents of which are as follows:1)The influence factors of BOS on imaging quality is analyzed.The best performance optical imaging parameters are set for camera's defocus problem.A Gaussian speckle image is designed based on the condition that speckle image do not overlap.The quality of speckle image is evaluated by four speckle image quality evaluation indexes.The results show that the quality of the speckle image is good,which is convenient for the accurate measurement of the subsequent image point displacement.2)The advantages and disadvantages of common background image processing algorithms are summarized.Digital Image Correlation(DIC)algorithm is selected as the subsequent image point displacement measurement algorithm.In view of the problems such as underexposure or overexposure and motion blur caused by the complex environment,several correlation coefficient functions are analyzed.Normalization Sum-Squared Difference(NSSD)correlation with strong anti-jamming performance is determined.Cuckoo SearchParticle Swarm Optimization(CS-P)algorithm is selected because of higher accuracy and stronger noise resistance.The better sub-pixel matching method,Gradient-Curved Surface Fitting,is given.Compared with other algorithms,the improved Gradient-Curved Surface Fitting method combines the advantages of the two methods well,it greatly improves the algorithm accuracy and realizes the high precision real-time sub-pixel matching.3)The structure of the real-time flow visualization system is analyzed.The optical imaging module is selected.The software interface of the real-time flow visualization is designed.The real-time flow visualization system is built based on the BOS.A series of flow field experiments is completed based on the system.The influence of several system parameters on the sensitivity of the system is verified by experiments,optimal optical performance parameters is set,and the correctness of the previous theoretical analysis is confirmed.The validity of the image point displacement measurement algorithm is verified.In view of the random error caused by environmental interference between the reference image and the measured image of traditional BOS,it is proved that the method of selecting the reference image as a disturbed image is better than the method of non-disturbing image.The performance difference is compared between the natural background and the speckle background.The selection criteria of the natural background are given.The foundation for the study of complex flow field is laid.The flow field experiment of hair dryer and candle flame is carried out,which shows that the performance of the flow visualization system is good,and the real-time flow visualization measurement with high precision is possible.The physical parameters when the drone hovers are analyzed and visualized,and the results show that some flow field parameter changes can be displayed in real-time and can be used to monitor the flow field changes.