Project 1 Proposal - Spencer

Overview

For my first project, I would like to visualize multiphase flow data in the Yurt. The owner of these data is Zhicheng Wang, a member of George Karniadakis' group. The data are genererated using Nektar++, a framework that is used to solve a wide variety of partial differential equations (PDEs). In particular, these datasets contain information about the position of particles as long as various measures of flow between particles (like vorticity, tensor invariants, etc.). These data are generally visualized using Paraview, so a major portion of my project will be incorporating the necessary features of Paraview into the Yurt so that I can create the visualizations that are desired. The final goal of the project will be to establish a pipeline so that these data can be automatically processed and fed into the Yurt after which some Paraview presets of scripts will be defined to create the various views that will be useful and interesting to those with the data.

Figure 1: An isosurface of Q (the second tensor invariant, a measure of how much flow is happening), colored by pressure.

Calendar

  • 2/11: Create visualizations for each dataset in paraview. Document and screenshot these, share with class.

  • 2/13: Check in with collaborators to ensure good visualizations of these datasets based pm feedback from class and projects. Modify visualizations of these datasets to accord with this. Investigate Paraview in the Yurt, to see if it is feasible to get it up and running for this project, and talk with Yurt people to get advice as needed.

  • 2/20: Try to get Paraview working in the Yurt for toy datasets. Work with other people in class using paraview to figure out which features work, which are missing, and which we will need.

  • 2/25: Create data processing pipeline to automate any tasks that need to be done pre-visualization. If this is simple. create a workflow and documentation describing what needs to be done.

  • 2/27: Visualize data in the Yurt.

  • 3/03: Writeup everytrhing, do Yurt demo, etc. Come up with conclusions as to the usefulness of VR in viewing this kind of data, and whether or not it provides further unknown information about the data not evident from prior visualizations.

Accompanying Deliverables

  • 2/11: Pictures and videos of the data being visualized in paraview, along with text describing what each of them actually is describing.

  • 2/13: A written record of the meeting in my journal, based on feedback. This will include what worked, what didn't, and the specific direction to take for future visualizations. Same for the evaluation of paraview.

  • 2/20: Uploaded videos of Paraview working in the Yurt, along with documentation of how to do so. Failing this, documentation of what didn't work, why, and what needs to be done to fix that.

  • 2/25: The Paraview/other scripts that create the visualizations

  • 2/27: Self-explanatory

  • 3/03: A detailed writeup on the wiki of the whole process for multiphase flow data, an addition to the wiki for Paraview (likely written with other students), documentation for any code I may have written or processes needed to view and process this data, and an evaluation as to the successes of the project and areas for future improvement

Class Activities

  • Further work in Paraview, specifically visualizing the datasets I am working with

  • Yurt demo?

  • Some kind of background simulation work to give other students an idea of how the dataset is generated -- maybe a Nektar++ demo? Unclear if this will be too technical/not useful.