IDL Allows Los Alamos National Lab to Visualize Complex Atomic Physics Data

Customer Challenge

The Atomic and Optical Theory (T-4) group at Los Alamos National Laboratory generates gigabytes of data using a number of opacity codes that calculate the absorption of radiation by materials under various physical conditions. Physicists on the team needed a way to organize data from the output code so it could be easily interpreted, allowing them to diagnose errors in the data generation and subsequently improve the physics models.

Solution Achieved

To help the team solve their data visualization challenge, Leslie Welser-Sherrill built what is now called OVID (Opacity Visualization in Detail), a visualization tool designed to investigate the various physical quantities of the output from opacity codes. Welser-Sherrill was brought into the group as a graduate student, tasked with building a tool for T-4, which needed a more interactive way to visualize their Fortran-generated datasets. The group had previously relied on manually extracting data from complex files and plotting it using simple plotting routines.

Welser-Sherrill, who had become proficient in IDL^® while studying physics as a graduate student, began by working closely with the physicists in T-4 to identify the types of visualizations that were required for analysis, and ultimately the functionality that would be useful in a visualization tool. From there, she created a graphical user interface that allows physicists to study 3-D rotating surfaces and 2-D images of opacities, densities, and pressures. She chose IDL based on its programming ease and advanced capabilities. “In my personal experience, no other programming language can match IDL in its versatility of scientific data visualization. The T-4 group recognized that IDL was the tool for the job to deal with their massive data sets,” she said.

OVID provides the visualization capabilities the T-4 group needs to complete their analyses, and it allows the user to examine multiple spectra and ion properties via seven tabbed windows with automated functionality:

• The main window acts as the driving engine, where data sets can be selected and viewed in a 3-D rotating object graphics plot
• Data from two codes can be compared, in the form of side-by-side rotating 3-D surfaces and a corresponding percent difference surface in the main window
• The surface from the main tab can be viewed as a 2-D image
• 2-D plots of information can be displayed in a survey environment, with six small windows to view the parameter space
• Particularly interesting 2-D plots can be zoomed into for closer study
• Total and fractional ion populations associated with individual table points can be viewed
• Spectral contributions from individual table points can be studied in detail

Output from any opacity code can be visualized by adhering to a specific format, and using a special extensible markup language (XML) package. In addition, the application is easily distributed to anyone with an IDL runtime license via the IDL Virtual Machine runtime utility, allowing users to take advantage of OVID without needing to compile and interact with the code.

“That was really useful because I could build a sophisticated tool, and the users didn’t have to worry about the code or the platform they were running on,” said Welser-Sherrill. “All they need to do is click on the executable, load in a dataset, and it’s automatically formatted in a way that the tool can read.”

OVID has been fully documented in a help file which details all aspects of the code, from setting up data according to OVID’s specified format and creating XML files to interacting with each window of the graphical user interface.

Benefits

• Non-IDL programmers can now easily interpret complex data using the distributed application
• Team members do not have to rely on others to produce intuitive representations of their data
• A windows-based environment makes it easy to perform multiple types of analyses

For more information, please contact Leslie Welser-Sherrill at lwelser@lanl.gov.