Calculation of Neutral Beam Injection into SSPX

The SSPX spheromak experiment has achieved electron temperatures of 350eV and confinement consistent with closed magnetic surfaces. In addition, there is evidence that the experiment may be up against an operational beta limit for Ohmic heating. To test this barrier, there are firm plans to add two 0.9MW Neutral Beam (NB) sources to the experiment. A question is whether the limit is due to instability. Since the deposited Ohmic power in the core is relatively small the additional power from the beams is sufficient to significantly increase the electron temperature. Here we present results of computations that will support this contention. We have developed a new NB module to calculate the orbits of the injected fast fast-ions. The previous computation made heavy use of tokamak ordering which fails for a tight-aspect-ratio device, where B{sub tor} {approx} B{sub pol}. The model calculates the deposition from the NFREYA package [1]. The neutral from the CX deposition is assumed to be ionized in place, a high-density approximation. The fast ions are then assumed to fill a constant angular momentum orbit. And finally, the fast ions immediately assume the form of a dragged down distribution. Transfer rates are then calculated from this distribution function [2]. The differential times are computed from the orbit times and the particle weights in each flux zone (the sampling bin) are proportional to the time spent in the zone. From this information the flux-surface-averaged profiles are obtained and fed into the appropriate transport equation. This procedure is clearly approximate, but accurate enough to help guide experiments. A major advantage is speed: 5000 particles can be processed in under 4s on our fastest LINUX box. This speed adds flexibility by enabling a ''large'' number of predictive studies. Similar approximations, without the accurate orbit calculation presented here, had some success comparing with experiment and TRANSP [3]. Since our procedure does not have multiple CX and relies on disparate time scales, more detailed understanding requires a ''complete'' NB package such as the NUBEAM [4] module, which follows injected fast ions along with their generations until they enter the main thermal distribution