Numerical simulation of two-dimensional kettle reboiler shell side thermal–hydraulics with swell level and liquid mass inventory prediction

Abstract

Simulation and analyses of two-phase flows across tube bundles is important for design and safety analyses of various types of steam generators and kettle reboilers. The information about the shell side thermal–hydraulics in this thermal equipment should include the liquid and vapour two-phase flow velocity fields, the void fraction distribution, as well as the swell level position and the liquid mass inventory. The two-fluid model of boiling two-phase flow around tubes in the bundle is applied for the simulation of the kettle reboiler shell side thermal–hydraulics. The tube bundle is modelled as a porous medium. Transfer processes at the vapour–liquid interfaces and on the tube walls are predicted with closure laws. The model is numerically solved by the “in-house” CFD code. The applied modelling method is validated against measured data of pressure drops in refrigerant R113 and n-pentane two-phase flows across tube rows in the bundle of a thin slice kettle reboiler, which are available in the open literature. The swell level position on the shell side is calculated solely by solving of the two-phase flow governing equations and with the application of an appropriate closure law for the vapour–liquid drag force, which enables the prediction of the liquid separation due to gravity from the upward flowing two-phase mixture. This is an improvement on the thermal–hydraulic modelling and numerical simulation of the kettle reboiler since the previous numerical simulations from the open literature have been performed with a priori specified swell level position and arbitrary boundary conditions for the velocity (or pressure) and void fraction boundary conditions at the swell level. The prediction of the swell level position also enables the calculation of the liquid mass inventory on the shell side, which gives insight into the kettle reboiler operating conditions and is crucial for the reliable prediction of the tube bundle dry-out during incidents of liquid feeding stoppage. There is a one-to-one correspondence between the liquid mass inventory and the swell level position. The presented simulation method enables iterative prediction of the liquid mass inventory for the specified swell level position and vice versa. In addition, the correlations for the liquid–vapour interfacial drag coefficient, which have shown previously fairly good predictions in cases of water–steam and water–air two-phase flows, are extended for the general application to other fluids.