With presentations from IVEC 2000 two years ago clearly indicating that three dimensional (3-D) TWT modeling had arrived, the IVEC 2002 presentations showed that 3-D modeling continues to reach more into the mainstream. Three of the six presented papers in this session covered the progress of the Naval Research Laboratory's fully three-dimensional electron gun and collector modeling tool, named MICHELLE. John Petillo started the session with a MICHELLE overview. The code is just about ready for a full U.S. industry release, with both 3-D and 2-D steady-state versions being available with both structured and unstructured gridding capability. A very complete secondary model for MICHELLE was presented by Norm Dionne (fourth talk). The model includes low energy secondaries as well as backscattered primaries. Eric Nelson rounded out the MICHELLE discussions by showing modeling results on a gridded gun in full 3-D (last talk). The vastly different grid sizes required throughout the model made this problem particularly challenging, yet MICHELLE converged well if given enough mesh points.

Thuc Bui from Calabazas Creek Research (CCR) presented a progress report on CCR's beam optics analysis (BOA) code (third talk) which has many features similar to MICHELLE and also includes adaptive meshing. Some test validation cases were presented, but a full 3-D gun model is still forthcoming. Karl Vaden and Tushar Ghosh (winner of the IEEE EDS 2002 Graduate Student Fellowship) each presented initial results on optimizing multistage depressed collector geometries. Ghosh (second talk) used a genetic algorithm, while Vaden (fifth talk) used a simulated annealing algorithm. Ghosh used the 3-D simulator LKOBRA for his collector solves and pushed the collector efficiency from 80% to over 90% on a particular case. Vaden used the 2-D simulator EGUN for his collector analysis and was also able to optimize a collector efficiency to beyond 90%. Both analyses are still in the early stages, and more complete results are expected.