Views from Intelligent Light
Brad Whitlock

DOE Invites Intelligent Light to Present In Situ with VisIt, Libsim, and FieldView

 

The Department of Energy hosts an annual meeting called the Computer Graphics Forum which brings together leading visualization experts who carry out DOE-supported research. Experts from National Laboratories, Department of Defense Research Institutions, Universities and select companies are invited to present updates on their research.  Intelligent Light was invited for a special vendor participation session and gave a talk called “Promoting In Situ with VisIt, Libsim, and FieldView”.

 

Topics of interests selected by the DOE this year include: computer procurement, status updates for software packages, and research for in situ and parallel programming on advanced HPC systems.

 

Advanced HPC systems have special challenges as they are increasingly heterogeneous architectures (often consisting of CPUs plus accelerators such as GPUs) with deep memory hierarchies. Several talks focused on new programming paradigms that are being created to develop large code bases that are both portable and efficient on heterogeneous architectures.

 

In situ was also a prominent research topic. In situ brings data analysis and visualization into solvers as they run, enabling them to extract information from the resident data so that more concentrated data can be written out. Saving smaller, more concentrated data is important because HPC systems have far higher compute capacity than I/O bandwidth and storage needed to store full results.

 

Related: DOE Awards Follow-On Grant for FieldView / VisIt Integration

Earl Duque

Wind Leaders Addressing Future Data Needs – Atmosphere to Electrons Initiative

I had the honor and pleasure to participate in the Atmosphere to Electrons Workshop hosted by the Department of Energy, Office of Energy Efficiency and Renewable Energy.  The focus of the initiative is on the use of computational simulation to improve understanding and performance predictions from the microscale to the mesoscale.

 

Penn_Schmitz_windturbine_F300x159

FieldView image published in paper: “Turbulence Transport Phenomena in the Wakes of Wind Turbines”, Earl Duque, Intelligent Light; Pankaj Jha and Jessica Bashioum and Sven Schmitz, The Pennsylvania State University

The event brought together leaders from the wind energy community including National Labs, Universities and Industry. The purpose was to map out the direction for simulating the performance of a wind turbine farm; capturing the temporal and spatial scales from meso-scale (kilometer and hours) down to the airfoil boundary layer scales (micron and milliseconds). Morning and afternoon sessions began with a topical plenary talk followed by working groups focused on the computation and modeling needs at different scales such as Park Scale, Turbine Scale and Airfoil Scale.

Wind Farm - FieldView image as published in "Wind Farm Simulations Using a Full Rotor Model for Wind Turbines", J. Sitaraman, D. Mavriplis, E. Duque AIAA Paper 2014-1086

Wind Farm – FieldView image as published in “Wind Farm Simulations Using a Full Rotor Model for Wind Turbines”, J. Sitaraman, D. Mavriplis, E. Duque
AIAA Paper 2014-1086

 

For me, it was clear that it will be essential to include in-situ data analysis methods and file I/O standards in order to work with the tremendous volumes of data that will be created and processed. This was recognized by many at the meeting.  The use of in-situ methods with FieldView and VisIt offers solutions to those grappling with the current data analysis bottlenecks.

 

FieldView image published in paper: "Turbulence Transport Phenomena in the Wakes of Wind Turbines", Earl Duque, Intelligent Light; Pankaj Jha and Jessica Bashioum and Sven Schmitz, The Pennsylvania State University

With the high-caliber people from government, academia, and industry converging on this challenging problem, the A2E initiative is making progress toward vast improvements in the understanding of the complex physics of wind flowing into and through wind farms.  DOE sees the potential to improve wind farm efficiency by 20% while drastically reducing operating costs for wind energy producers.

Related Research Papers:


Yves-Marie Lefebvre

ANSYS® Fluent® 16.0 Enables FieldView HPC Capabilities

FieldView Parallel Export from ANSYS® Fluent® 16.0

 

ANSYS Fluent 16.0 provides the capability to export results to FV-UNS while maintaining the partitioning established for a FLUENT parallel solution.  The file is fully compatible with FieldView Parallel allowing for the substantial time savings from parallel processing.  Every FieldView license will support at least 8 processor cores.  FieldView Parallel will readily scale to 64 processors and beyond.  The export files can be automatically generated during the solver run.

 

Surface Export – From the ANSYS Fluent TUI a new option called “fieldview-unstruct-surfaces” creates a surface only FV-UNS export, where the user selects the exported surfaces from a list of boundary conditions and post-processing surfaces (iso-surfaces, planes…).

  • Surface export for Parallel FieldView – A surface export made from a parallel session of Fluent will maintain the partitioning of the surfaces, and the resulting file can thus be read in parallel by FieldView.
  • Interior surfaces may be selected for export in Fluent 16.0.  This is a longstanding request and is useful in applications like integrating flow rate or seeding streamlines on control surfaces.  Please contact our support team for information about how to use this new feature.

FieldView and FLUENT – Power for Unsteady Simulations

 

Unsteady simulations are often post-processed and reviewed using FieldView XDB files which are a fraction of the size of volume data.  Exporting surface data, reading into FieldView Parallel and then having FieldView post-process and produce XDB files can be entirely automated.  This workflow delivers lightweight XDB files for review in FieldView or by the free XDBview reader. When exporting from ANSYS FLUENT, the resulting file can be read into FieldView in parallel, dramatically reducing the read-in time.

Choose the right export for your data:  Quick Start – Working with FLUENT16 data in FieldView (PDF)