Particle Visualization

Abstract

Particle simulations are powerful tools for understanding the complex phenomena associated with many areas of physics research, including confined plasma and high energy particle beams. Analyzing the data, however, presents a challenge due to the large quantity of particles, variables, and time steps. In this paper we describe a data exploration system that visualizes time-varying, multivariate point-based data from gyrokinetic particle simulations. By utilizing two modes of interaction (physical space and variable space) the system allows scientists to explore collections of densely packed particles and discover interesting aspects of the data. We employ a parallel-coordinate interface for interactively selecting particles in multivariate space. In this manner, particles with deeper connections can be separated from the rest of the data. From the results of this system, we are able to identify features of interest, such as the location and motion of particles that become trapped in turbulent plasma flow.

Full Paper:
An Integrated Exploration Approach to Visualizing Multivariate Particle Data
Chad Jones, Kwan-Liu Ma, Stephane Ethier, and Wei-Li Lee
Computing in Science & Engineering, to appear 2008.

Physical Rendering

Particle Rendering

Particles are shown at a single instance in time using GPU shaded spherical glyphs. The 3D appearance helps to distinguish individual particles in a dense collection where the values are not locally continuious. The color and opacity are mapped using a single variable chosen from the set of variables and a simple transfer function editor.

Particles rendered using semi-transparency.

On the left, the particles are rendered with full opacity colored by parallel velocity (blue for negative, white for zero, and orange for positive). On the right, we use semitransparency to reduce the visibility of near zero valued particles.

Pathline Rendering

To help see time dependent changes of both value and location, the application can render illuminated, colored lines for the trajectories of selected particles. The line color is set by the value of the particle at each location along the path by using the same scalar mapping that is defined by the transfer function.

Parallel Coordinates

A parrellel coordinates plot is used to show relationships between the multiple variables and provide an interface for particle selection. The more particles that exhibit certain value pairs, the brighter the plot will be in the area. Thus, one can see which areas represent have high intensity and which have low intensity of connections.

Particle Selection

To allow for interaction, each axis can be "brushed" vis a mouse click and drag to select and deselect particles. The selected particle lines are drawn in red, which helps highlight the individual values for that particle. Each axis can be toggled between either the union or intersection operator. An example of the steps involved in using the parallel coordinates is shown below.

	One million particles colored by parallel velocity and no selections made. Now we would like observe the distribution of particles based on parallel velocity and statistical weigtht...
	The coloring has been changed to show the distance from the center of the tokamak, such that blue is close and orange is far. A selection is made on the 3rd axis (parallel velocity) such that only particles with a high positive value are shown.
	Here particles are selected from the 4th axis (statistical weight) such that the particles have only a high postive statistical weight.
	Now the selection can be restricted by making both axes intersecting. Now only particles that are both high parallel velocity and high stastical weight are shown. As we can see, the particles tend towards the outer rim of the tokamak.

Selection Locking

Since the data is time varying, the selection you make at one time step might change at a different time step. Two types of locking mechanisms are given: particle lock and range lock. With particle lock, the particles that are selected at one time step are kept as the focus particles regardless of how their values change. With range lock, the ranges that were used at the locked time step are strictly enforced, i.e. a new collection of particles may be calculated at each time step to see what new particles exhibit those specific value ranges.

First time step with only inner and outer particles.

Time Step 1: Particles are selected via the parallel coordinates such that they are close to the center and far from the center.

After tracing the same particles to timestep 10.

Pathlines of those particles between time steps 1 to 20.

Time Step 10: On the left, range lock is used to generate a new collection of particles with the same property as before. On the right, particle lock lets us observe how the original selection of particles has changed.

Time Steps 1 through 20: Instead of discrete comparisons, pathline rendering allows the entire trajectory between time steps 1 and 20 to be observed.

Trapped Particles

Through the integrated interface, we were able to easily capture a set of particles that represent so-called trapped particles. Instead of the continuous clockwise or counterclockwise trajectories, these particles turn around and change direction. The image and video below show particles that change direction and become trapped at the same point in time. In addition, the particles were chosen such that the deccelarete into this state quickly and accelerate quickly afterwards.

As another example, we have selected trapped particles that exhibit the trapped state over a longer period of time. It is a smaller collection, but the particles in this collection are continuously changing direction for a long period of time.