Automated Extraction of Flow Features

Journal of Visualization

While vector fields are essential to simulate a large amount of natural phenomena, the difficulty to identify patterns and predict behaviors makes the visual segmentation in simulations an attractive and powerful tool. In this paper, we present a novel user-steered segmentation framework to cope with steady as well as unsteady vector fields on fluid flow simulations. Given a discrete vector field, our approach extracts multi-valued features from the field by exploiting its streamline structures so that these features are mapped to a visual space through a multidimensional projection technique. From an easy-to-handle interface, the user can interact with the projected data so as to partition and explore the most relevant vector features in a guidance frame of the simulation. Besides navigating and visually mining structures of interest, the interactivity with the projected data also allows the user to progressively enhance the segmentation result according to his insights. Finally, to successfully deal with unsteady simulations, the segments previously annotated by the user are used as a training set for a Support Vector Machine approach that classifies the remaining frames in the flow. We attest the effectiveness and versatility of our methodology throughout a set of classical physical-inspired applications on fluid flow simulations as depicted in the experiment results section.

2011 International Conference on Virtual Reality and Visualization, 2011

Feature visualization plays an important role in visualization of complicated flows because it can highlight the feature of the flows with a simplified representation. The traditional feature visualization methods may exact some important features in flow field imprecisely due to the lack of the knowledge and the experience of the user. This paper gives a particle-based visualization system which is developed with the application of the interactive fuzzy feature extraction and interactive visual analysis theories. To obtain a more precise feature extraction, we have proposed an interactive fuzzy feature description language (FFDL) and an interactive fuzzy feature extraction algorithm. Based on the work before, we introduced the proportion ration for different rules and optimized our algorithm in practice further by communicating with specific researchers and doing lots of experiments. The further experiments show that our method can not only make full use of the ability of the user to extract the features precisely, but also reflect the uncertainty of the numerical simulation data.

Finite Elements in Analysis and Design, 1995

The objective is to present new ideas for the implementation of visualization and analysis environments. This is carried out with a software interface allowing the design and configuration of project-specific environments built around a core library that performs all the data extraction and manipulation. This includes processing of solutions for rendering as well as for extraction of flow features for active control on the solution procedure. While the first function is highly interactive, the second is usually a batch oriented process coupled to the numerical algorithm.To achieve this goal, it is necessary to interpret solution data in a generic sense. The grid, the primary variables of the problem, and the derived variables are all considered as a number of simple scalar (discrete) fields defined on the domain. Basic entities in the library then provide building blocks to create images for rendering purposes or new fields as required in the grid adaption control loop.A data analysis language has been created. Its entities and grammar are oriented towards the description of both the rendering and the analysis processes. This customized environment is saved in an user configuration file, with an easily understandable syntax, which is loaded at execution time, hence providing the proper configuration for each application.

2005

Abstract Visualization of CFD simulation data on adaptive resolution, three-dimensional grids poses several challenges. The wide range of real-world data set sizes and the geometric versatility within individual, CFD simulation models present challenges to the engineers analyzing simulation results. Users also face perceptual problems such as occlusion, visual complexity, lack of directional cues, and lack of depth cues.

IEEE Transactions on Visualization and Computer Graphics, 1995

International Journal of Numerical Analysis and Modeling, 2008

Very complex flow structures occur during separation that can appear in a wide variety of applications involving flow over a bluff body. This study examines the ability to detect the dynamic interactions of vortical structures generated from a Helmholtz instability caused by separation over bluff bodies at large Reynolds number of approximately 10 4 based on a cross stream characteristic length of the geometry. Accordingly, two configurations, a thin airfoil with flow at an angle of attack of 20 0 and a square cylinder with normally incident flow are examined. A time-resolved, three-component PIV data set is collected in a symmetry plane for the airfoil, whereas direct numerical simulations are used to obtain flow over the square cylinder. The experimental data consists of the velocity field, whereas simulations provide both velocity and pressure-gradient fields. Two different approaches analyzing vector field and tensor field topologies are considered to identify vortical structures and local, swirl regions. The vector field topology uses (1) the Γ function that maps the degree of rotation rate (or pressure-gradients) to identify local swirl regions, and (2) Entity Connection Graph (ECG) that combines the Conley theory and Morse decomposition to identify vector field topology consisting of fixed points (sources, sinks, saddles) and periodic orbits, together with separatrices (links connecting them). The tensor field feature uses (1) the λ 2 method that examines the gradient fields of velocity or pressure-gradient to identify local regions of pressure minima, and (2) tensor field feature that decomposes the velocity-gradient or pressure Hessian tensor into isotropic scaling, rotation, and anisotropic stretching parts to identify regions of high swirl. The vectorfield topology requires spatial integration of the velocity or pressure-gradient fields and represents a global descriptor of vortical structures. The tensor field feature, on the other hand, is based on gradients of the velocity of pressuregradient vectors and represents a local descriptor. A detailed comparison of these techniques is performed by applying them to velocity or pressure-based data and using spatial filtered data sets to identify the multiscale features of the flow. It is shown that various techniques provide useful information about the flow field at different scales that can be used for further analysis of many fluid engineering problems of practical interest.

2002

This thesis demonstrates a generalized method for extracting shear layers as features in viscous computational fluid dynamics flowfields. A source term indicative of shear in the flow combines with convective terms to constitute a conservation equation for a scalar shear layer detector variable. The source terms explored in this inves-tigation come from the viscous terms of the Navier-Stokes momentum equations as well as from a simple shear measure obtained from the velocity gradient tensor. A threshold value of the scalar detector variable indicates the boundary layer or wake edge, while the magnitude of the detector indicates the intensity of shear influence upon any particular point in the flow. Test cases validate the convective technique and demonstrate its rejection of false sources. Among the applications of shear layer feature extraction are solution improvement techniques like grid adaptation and tools for analysis and design such as visualization.

2003

Abstract Flow visualisation is an attractive topic in data visualisation, offering great challenges for research. Very large data sets must be processed, consisting of multivariate data at large numbers of grid points, often arranged in many time steps. Recently, the steadily increasing performance of computers again has become a driving force for new advances in flow visualisation, especially in techniques based on texturing, feature extraction, vector field clustering, and topology extraction.

2021

2005

Abstract Usually, research related software consists of individual, isolated prototypes because researchers are interested in a small proof-of-concept application for demonstration. Here we present software developed for research purposes, but which has been included into a larger, commercial visualization system. We describe the design and implementation of a flow visualization subsystem within the framework of a software package capable of modeling, simulation, and visualization of CFD simulation data.