| AVS 65th International Symposium & Exhibition | |
| Vacuum Technology Division | Monday Sessions |
| Session VT-MoA |
| Session: | Pumping and Outgassing |
| Presenter: | Santiago Ochoa Guaman, Karlsruhe Institute of Technology, Germany |
| Authors: | S.L. Ochoa Guaman, Karlsruhe Institute of Technology, Germany T. Giegerich, Karlsruhe Institute of Technology, Germany C. Dahlke, HERMETIC-Pumpen GmbH, Germany C. Day, Karlsruhe Institut of Technology (KIT), Germany |
| Correspondent: | Click to Email |
In the European fusion reactor (DEMO) development program, continuous working vacuum pumps are foreseen to pump the reactor. The pumps have to process large amounts of tritium, a radioactive and chemically very active gas. In a systems engineering approach, a pumping solution based on liquid ring pumps (LRPs) and diffusion pumps has been identified. Mercury is the only fluid perfectly tritium compatible and will be applied as working fluid.
LRPs exist for around 130 years and several mathematical approaches have been developed for its 1D modelling. Diagrams and tables also have been produce from experiments for fluid densities between 800 kg/m3 to 1200 kg/m3 but mostly for water as working fluid and with air as pumped gas. Modern 3D simulation tools have not been applied so far for analyzing these pumps. Thus, in order to design and analyze the operating behavior of LRPs with mercury as working fluid, it is necessary to design a special code for the prediction of the thermodynamic and operational behavior of LRPs operating with high density working fluids.
This great challenge is presented in this work, starting with the development of a simulation code based on three already existing methods. For its benchmarking against literature data, water as working fluid and air as pumped gas will be used. In a follow-up step, the code will be run considering mercury as working fluid. These results will be discussed against experimental results produced at the THESEUS vacuum pump test facility at KIT.
In the second part of this work, a two-phase three-dimensional CFD model will be performed using a more detailed pump geometry. Goal of this activity is to achieve a more accurate description of the pump performance without the use of empirical parameters. This requires extensive modelling and high computational effort. The status of this task will be presented in this paper and first results will be shown and benchmarked against experiments and the code.