PS-P-99-050 BACKGROUND • Development of high flux (> 4 kg/m 2 hr) and high selective (> 200, 5% H 2 O-nBuOH) inorganic pervaporation membranes and modules. • Experiments: existence of strong mass and heat transfer resistances decrease the pervaporation efficiency. OBJECTIVE Increase the separation efficiency by improving the flow characteristics and design of the (lab scale) module. APPROACH Optimisation of the lab scale module by: • Chemical engineering calculations (Poster part I). • Computational Fluid Dynamic calculations (This poster). 1. understand physical transport phenomena in the module, 2. calculate pervaporation efficiency of proposed module designs. INORGANIC MEMBRANE PERVAPORATION MODULE DESIGN Optimisation by Computational Fluid Dynamics CFD MODEL • 2D and 3D modelling of feed side (concentration, temperature, flow, membrane length, and annular space). • Low Reynolds turbulence modelling (no wall function necessary). • Heat and mass transfer. • Membrane included as boundary condition on feed side: flux is function of concentration and temperature on membrane surface. • Transversal flow near inlet and outlet strongly increases local PV flux. CONCLUSIONS • Lab scale module: strong influence of temperature and concentration polarisation. • Bench scale module: need for e.g. baffles to enhance transversal flow/turbulence. • Detailed physical insight in heat and mass transport leads to better design of bench and pilot scale modules. • Local flux strongly depends on local flow conditions. 2D CFD RESULTS Flux decrease Concentration Temperature Combined due to polarisation polarisation effect at Re = 2400 27% 8% 31% at Re = 4700 n.a. n.a. 21% at Re = 9400 n.a. n.a. 14% Figure 1: Schematic view lab scale module Figure 4: PV flux on top, side and bottom of membrane surface. Figure 5: 2D/3D PV flux as function of axial distance • Increasing flux with feed flow rate, however still polarisation effects at high flow rates. 3D CFD RESULTS Figure 3: PV flux as function of axial distance Figure 2: 3D CFD mesh, approx. 300,000 computational cells Y.C. van Delft H.M. van Veen P.P.A.C. Pex Netherlands Energy Research Foundation, ECN, P.O. Box 1, 1755 ZG Petten, The Netherlands Tel. +31 224 564640, E-mail: [email protected] Work performed by: J.A. Lycklama à Nijeholt C.J.J. Beemsterboer NRG, P.O. Box 25, 1755 ZG Petten, The Netherlands E-mail: [email protected] S. Sommer B. Klinkhammer T. Melin Institut für Verfahrenstechnik, RWTH Aachen, Turmstrasse 46, Aachen, Germany This work is partially financed by the European Union and by Novem, the Dutch organisation for Energy and Environment.