Visualizing Particle Trajectories

Terabyte output from three-dimensional simulations motivates researchers to use parallel tools to extract physics information. Accelerator scientists collaborate with developers of VisIt parallel visualization tools to provide three-dimensional images for analysis of the wake and laser (figure 4, p16). Particle bunches must be identified and the particles must be tracked through the simulation to understand where they are injected and where accelerator properties can be modified to increase beam quality (figure 6). Clustering and machine learning techniques are being used to automatically identify beam formation and quality in the accelerator. The particles are then tracked using the FastBit indexing tool, which is under development to allow real-time exploration of particle trajectories and associated fields. These projects are in collaboration with the SciDAC VACET center, the NERSC and LBNL visualization and analytics groups, and the Tech-X VizSchema data access team.

D. Ushizima (left) and O. Rübel (right), LBNL Figure 6. Formation of high-quality particle beams is automatically identified (white dots, left) by clustering techniques . Tracks through the simulation of accelerated particles, obtained by FastBit, are overlain on a snapshot of the wake particles (gray) showing injection from the sides of the wake and oscillation.