KnowledgeAssisted Visualization of Turbulent Combustion Simulations - PowerPoint PPT Presentation

1 / 16
About This Presentation
Title:

KnowledgeAssisted Visualization of Turbulent Combustion Simulations

Description:

... of Turbulent Combustion Simulations. Chaoli Wang, Hongfeng Yu, Kwan-Liu Ma ... Observe variable relationships close to the mixfrac surface ... – PowerPoint PPT presentation

Number of Views:26
Avg rating:3.0/5.0
Slides: 17
Provided by: wcl51
Category:

less

Transcript and Presenter's Notes

Title: KnowledgeAssisted Visualization of Turbulent Combustion Simulations


1
Knowledge-Assisted Visualization of Turbulent
Combustion Simulations
  • Chaoli Wang, Hongfeng Yu, Kwan-Liu Ma

2
Turbulent combustion simulations
  • Direct numerical simulations
  • Time-varying, multivariate data
  • 800 686 217, 450MB
  • 53 time steps
  • 4 variables mixfrac, chi, HO2, and OH
  • 93GB in total

3
Visualization-specific task
  • Scientific interests
  • Observe variable relationships close to the
    mixfrac surface
  • Bring out more the lower values of other variables

The mixed rendering of the mixfrac (isovalue
0.2) and the HO2 variables
4
Challenges
  • Data ranges of other variables close to the
    surface are unknown
  • Only value-based transfer function may bring out
    undesired visualization contents
  • Lack of control over the amount of information
    shown around the surface

5
Our solution
6
Transfer function specification
7
d 0.02
Visualization result
d 0.05
8
d 0.10
Visualization result
d 0.20
9
Video demos
10
Summary
  • Knowledge-assisted visualization
  • Domain knowledge isovalue, ranges of interests
  • Derived knowledge distance volume and partial
    histogram
  • Importance-driven visualization
  • Future work
  • Time-varying, multivariate data compression
  • Utilize domain knowledge
  • Visualization-specific compression

11
Thank you!
12
Extra slides
13
Visualization-specific compression
  • Regions of interest are around the given surface
  • Data precision can thus vary according to the
    distance to the surface
  • Our solution
  • Non-uniform quantization
  • Space-time coherence utilization
  • Decompression on the fly using graphics hardware

14
Compression result
  • Compression ratio 20x
  • Each voxel only uses less than 2 bits per
    variable on the average
  • Advantages
  • Reduce data transfer among disk to main memory,
    and main memory to video memory
  • Fast offline compression and online decompression
  • Preserve fine details near the surface and
    maintain the overall image quality

15
Compression result
original
compressed (20x)
16
Compression result
original
compressed (20x)
Write a Comment
User Comments (0)
About PowerShow.com