Research On Visualization Technology Of Vortex Characterization Of SPH Fluid

Flow feature visualization technology only extracts flow feature to obtain useful information,thus compressing the huge fluid data.It has been widely applied in medicine,aerospace and other fields.As one of the most important features in fluid,feature extraction of vortex has gradually become a critical content of flow feature visualization.At present,there are mainly Eulerian grid method and Lagrangian particle method to simulate fluid.Smoothed particle hydrodynamics(SPH)is one type of Lagrangian particle method.It discretizes the fluid into several particles and calculates the flow state of the whole fluid through the interaction between particles.However,due to the influence of the moving of the reference,there are some problems in SPH fluid feature extraction,such as feature loss.Therefore,it is of great significance to study the visualization method of vortex feature which can extract SPH fluid efficiently and accurately.At present,there is no general definition to describe the vortex feature of fluid.As a result,judgment of vortex is different,and the accuracy of vortex feature extraction also needs to be further improved.In this paper,we carried out research on the feature extraction accuracy and application scope existing in the current characterization of SPH fluid,and propose a predictor-corrector method based on rotation invariance to extract vortex feature more accurately.Firstly,aiming at the low precision of feature extraction,this paper proposes a method to extract vortex seed point in rotation invariant frame.The method calculates rotation invariant Jacobian to extract seed point and eliminate the precision degradation caused by reference frame movement.Secondly,due to the rotation invariant method can only extract vortex feature which perform constant rotation motion,in this paper,a specific axis is set to expand the application of rotation invariant method.Finally,Karman Vortex Street Data and other fluid scenes are simulated to compared the proposed method and the original predictor-corrector method.The experimental results show that,the predictor-corrector method based on rotation invariance in this paper can extract vortex feature more accruately.In addition,for multi-axis scenes,the proposed method segments the data set and extracts vortex feature respectively.The extraction results show that,the proposed method can apply rotation invariant method to general scenes,and ensure a certain accuracy.