Experimentally Validated Image Simulations of Tracer Particles for Laboratory-Scale X-Ray PIV

We evaluate two simulation methods for prediction of X-ray PIV (XPIV) system performance: Beer-Lambert (BL) ray tracing and Monte Carlo N-Particle (MCNP) photon tracking. These simulation tools enable the design and methodical improvement of XPIV experiments and tracer particles. Simulated and experimental data of hollow, silver-coated, glass sphere tracer particles (AGSF-33) are compared. We also compare two tracer particles. As predicted by the simulations, the AGSF-33 particles are visible with a signal-to-noise ratio greater than unity in 100ms exposure time images, demonstrating their potential as XPIV and XPTV tracers. Although BL simulations neglect scattering detections, among other simplifications, the BL approach provides a first-order estimate of the system behavior and is computationally cheap. BL enables exploration of a vast parameter space for system design. MCNP simulations, on the other hand, more accurately predict experimental images. For most applications, however, the order of magnitude greater computational expense of MCNP may not be justified by the improved performance compared to BL. Additionally, we present data showing the localizability of 10 micrometer-scale, neutrally buoyant tracer particles along with preliminary in-lab XPIV measurements.