Modeling of time-dependent gas pumping networks in the whole range of the Knudsen number: Simulation of the ITER dwell phase

Abstract

A hybrid time-dependent algorithm to simulate the transient response of gas distribution systems of arbitrary size and complexity, in the whole range of the Knudsen number, is proposed. The pressure evolution in the vessels is described by a simple macro model derived via mass conservation principals, while the pressure and mass flow rates in the pipe network are described by a micro model, consisting of the in-house steady-state gas network code “ARIADNE”, based on kinetic theory. The two models are explicitly coupled, i.e. at each time step the gas network is solved via ARIADNE and the computed node pressures and pipe flow rates are provided to the macroscale evolution equations to update the vessel pressures. The proposed methodology and code are successfully validated by solving two prototype problems and comparing the results with corresponding ones available in the literature or obtained by Molflow+. The computational effectiveness and efficiency of the proposed approach to model large size networks is demonstrated by simulating the transient response of the ITER torus primary pumping system in the dwell phase. Interesting findings for the torus effective pumping speed and pressure evolution, including the final pressure at the end of the dwell phase are provided. © 2019 Elsevier B.V.