A Flow feature detection framework for large-scale computational data based on incremental proper orthogonal decomposition and data mining

ABSTRACT A framework based on incremental proper orthogonal decomposition (iPOD) and data mining to perform large-scale computational data analysis is presented. It includes iPOD to incrementally reduce data dimensions and decouple dynamic flow structures in massive CFD data; data mining to classify and identify candidate global regions of interest (ROIs) for focused analysis; feature detection to capture key flow features and ultimate ROIs (UROIs); and targeted data storage and visualisation. Quantitative results show that iPOD is able to process large datasets that overwhelm the batch-POD, leading to 4–16× data reduction in the temporal domain. Data mining and feature detection algorithms, respectively, identify 50–70% of the spatial domain with high probability of flow feature occurrence and only 2–30% containing key flow features. The UROI and associated data can be selectively stored and visualised. In contrast to batch-POD, iPOD reduces memory usage by more than 10× and time by up to 75%.

[1]  S. Fu,et al.  Analysis of flow structures in supersonic plane mixing layers using the POD method , 2008 .

[2]  M. Pauline Baker,et al.  VisBench: A Framework for Remote Data Visualization and Analysis , 2001, International Conference on Computational Science.

[3]  Robert S. Laramee,et al.  The State of the Art in Flow Visualisation: Feature Extraction and Tracking , 2003, Comput. Graph. Forum.

[4]  Robert Haimes,et al.  Shock detection from computational fluid dynamics results , 1999 .

[5]  J. N. Sørensen,et al.  Low-Dimensional Modeling of a Driven Cavity Flow with Two Free Parameters , 2003 .

[6]  M. S. Chong,et al.  A general classification of three-dimensional flow fields , 1990 .

[7]  Richard D. Sandberg,et al.  Identification of large coherent structures in supersonic axisymmetric wakes , 2009 .

[8]  B. Haasdonk,et al.  REDUCED BASIS METHOD FOR FINITE VOLUME APPROXIMATIONS OF PARAMETRIZED LINEAR EVOLUTION EQUATIONS , 2008 .

[9]  M. Manhart,et al.  A spatiotemporal decomposition of a fully inhomogeneous turbulent flow field , 1993 .

[10]  Matthew Brand,et al.  Incremental Singular Value Decomposition of Uncertain Data with Missing Values , 2002, ECCV.

[11]  Benjamin M. Sloan,et al.  Orthogonal decomposition as a design tool: With application to a mixing impeller , 2013 .

[12]  Kun Yang,et al.  A temperature prediction-correction method for estimating surface soil heat flux from soil temperature and moisture data , 2008 .

[13]  A. Vernet,et al.  Vortex modes in the wake of an oscillating long flexible cylinder combining POD and fuzzy clustering , 2010 .

[14]  Conrad Sanderson,et al.  Armadillo: a template-based C++ library for linear algebra , 2016, J. Open Source Softw..

[15]  Raghu Machiraju,et al.  Physics-based Feature Mining for Large Data Exploration D a T a M I N I N G , 2001 .

[16]  Tomas W. Muld,et al.  Mode Decomposition on Surface-Mounted Cube , 2012 .

[17]  Utkarsh Ayachit,et al.  The ParaView Guide: A Parallel Visualization Application , 2015 .

[18]  Tomas W. Muld.  Slipstream and Flow Structures in the Near Wake of High-Speed Trains , 2012 .

[19]  Pieter G. Buning,et al.  PLOT3D user's manual , 1990 .

[20]  Paul Van Dooren,et al.  Low-rank incremental methods for computing dominant singular subspaces☆ , 2012 .

[21]  Helwig Hauser,et al.  Energy‐scale Aware Feature Extraction for Flow Visualization , 2011, Comput. Graph. Forum.

[22]  Frank Thiele,et al.  Coherent and turbulent process analysis in the flow past a circular cylinder at high Reynolds number , 2008 .

[23]  Octavian Frederich,et al.  Large-scale dynamics in the flow around a finite cylinder with a ground plate , 2011, Proceeding of Sixth International Symposium on Turbulence and Shear Flow Phenomena.

[24]  Einar Heiberg,et al.  Three-Dimensional Flow Characterization Using Vector Pattern Matching , 2003, IEEE Trans. Vis. Comput. Graph..

[25]  Pierre Kerfriden,et al.  An efficient goal‐oriented sampling strategy using reduced basis method for parametrized elastodynamic problems , 2015 .

[26]  P. Moin,et al.  Eddies, streams, and convergence zones in turbulent flows , 1988 .

[27]  William E. Lorensen,et al.  The visualization toolkit (2nd ed.): an object-oriented approach to 3D graphics , 1998 .

[28]  Nicholas J. Bisek,et al.  Plasma Control of a Turbulent Shock Boundary-Layer Interaction , 2013 .

[29]  E. Schuster,et al.  Model order reduction for high dimensional linear systems based on rank-1 incremental proper orthogonal decomposition , 2011, Proceedings of the 2011 American Control Conference.

[30]  William E. Lorensen,et al.  Marching cubes: A high resolution 3D surface construction algorithm , 1987, SIGGRAPH.

[31]  Dan S. Henningson,et al.  Flow structures around a high-speed train extracted using Proper Orthogonal Decomposition and Dynamic Mode Decomposition , 2012 .

[32]  Xiang Li,et al.  Feature extraction from massive, dynamic computational data based on proper orthogonal decomposition and feature mining , 2014, Journal of Vision.

[33]  Nicholas Bisek,et al.  High-Order Implicit Large-Eddy Simulations of Supersonic Corner Flo , 2014 .

[34]  F. Sotiropoulos,et al.  On the bimodal dynamics of the turbulent horseshoe vortex system in a wing-body junction , 2007 .