Applications Devleopment

During FY02, we completed the following applications development activites.

Magnetic Fusion. We completed an initial implementation of an AMR code for nonideal MHD in a singly-periodic rectangular domain, suitable for investigating the fundamental physics of the internal kink mode in a resistive plasma. This included the development of a suitable second-order Godunov method for compressible MHD, with a semi-implicit coupling to viscous and resistive terms. We began the validation of this capability in collaboration with CEMM. We also began the development of an embedded boundary version of this capability suitable for representing the plasma boundary of a tokomak in cylindrical geometry.

Accelerator Modeling. We replaced the uniform-grid algorithm traditionally used in electrostatic PIC codes with an adaptive mesh refinement finite-difference Poisson solver. We also coupled this to a Shortly-Weller Cartesian grid treatment of irregular Dirichlet boundaries. The use of adaptive grids for PIC calculations is a natural extension, since grid resolution requirements typically are greatest in the neighborhood of large particle densities, particularly when particles are near irregular boundaries. We integrated this package into existing simulation codes for RF accelerators.

Combustion. In FY02, we focused our research in combustion on two target investigations: pollutant production in a steady ammonia-seeded methane diffusion flame, and a turbulence-chemistry interactions in a wrinkled premixed methane flame.

We are investigating ammonia-seeded diffusion flames in collaboration with researchers at the University of Heidelberg and the Technical University of Denmark. The goal of the work is to improve the understanding of NO pollutant formation due to fuel-bound nitrogen. Data is gathered from the experiment using laser-induced fluorescence (LIF). In the ``two-line'' analysis approach for temperature extraction, two frequencies associated with the NO molecule are excited by the laser over a two-dimensional sheet through the vertical midplane of the flame, and the resulting LIF signals are processed using the known temperature-dependence of the bands. The absolute NO concentration can be inferred from either signal if the temperature and mole fractions are known for all species that may quench the NO LIF signal. Unfortunately, the NO signal is strongly quenched by the O$_2$ molecule, so that the two-line approach depends critically on assumed chemical profiles.

In our FY02 work, we suggested a new approach that more closely intertwined simulation and experimental data interpretation. Using an adaptive axisymmtric two-dimensional low Mach number combustion model, we computed steady diffusion flame solutions corresponding to the experimental setup. After validating the computational results against the temperature profiles inferred from the experiment, we used the simulation results, combined with quantum-dynamical quenching models to generate numerical LIF images. Across the range of experimental parameters, these numerical LIF images showed exceptional agreement with the raw LIF data from the experiment. Additionally, using simulated chemical distributions directly in the quench calculations, we demonstrated very good agreement between the computed and experimentally inferred NO concentrations. We did observe some discrepancies in the NO profiles in low-temperature regions of the flame, and are continuing our collaborative investigations in this area to pinpoint the responsible weaknesses in the model.

Concurrent with this work, we are applying our three-dimensional, high-resolution numerical algorithm technology toward understanding the detailed response of a lean premixed methane flame to fuel-stream turbulence. These simulations are the largest of their type ever attempted, and were achievable through our use of locally-adaptive gridding methods coupled to the low Mach number formulation. The flame chemistry is modeled with a 19 species, 84 reaction subset of the GRIMech-1.2 methane chemistry mechanism. The calculations are carried out over the 8$\times$8$\times$16 mm domain on a hierarchical grid structure with an effective resolution of 256$\times$256$\times$512 cells.

Simulation results confirm the widely held theory that wrinkling due principally to turbulence increases the effective flame propagation speed. Under the turbulence regimes studies thus far, the increase in speed is predominately attibutable to the increased flame surface area. This suggests that the time-dependent stretch and flame curvature have little aggregate effect overall on the flame stucture. However, when we probe more closely, we find a different picture. Even for low turbulence levels, the heat release in the flame correlates strongly with flame curvature. In fact, the strength of many of the reactions in the system are correlated to flame curvature, an observation which is likely related to focusing and defocusing preferential hydrogen diffusion effects. We found that an additional effect of the turbulence was to spatially separate regions production and destruction zones for a subset of the chemical species, particularly in regions of negative curvature. As a result, the residence times, and computed molar concentrations of these species where correlated to the curvature as well.