(539d) Membrane Transport and Process Modelling in Osn: A Multiscale Approach

Membrane transport and process modelling
in OSN: a multiscale approach

Patrizia Marchetti
¹ , Binchu Shi ^1,2, Dimitar Peshev ³, Andrew
G. Livingston ^1,2 *

¹
Department of Chemical Engineering, Imperial College London, Exhibition Road,
London SW7 2AZ, United Kingdom

²
Evonik Membrane Extraction Technology Limited, Unit 6 Greenford Park, Ockham
Drive, Greenford, London, UB6 0AZ, United Kingdom

³
Department of Chemical Engineering, University of Chemical Technology and
Metallurgy, 8, Kl. Ohridsky Blvd, Sofia 1756, Bulgaria

* a.livingston@imperial.ac.uk

In
Organic Solvent Nanofiltration (OSN), a fundamental understanding of the basic
separation mechanism is useful to advance the field and wisely tune new materials
and processes. The development of membrane processes usually involves several
stages, starting from feasibility tests at membrane scale, passing through fluid
dynamics and mass transfer in membrane modules and finishing with large
industrial scale process design (Figure 1) [1,2].

Figure 1. Modelling levels for
the development of a membrane process [1].

Substantial
research has focused on the description of the transport mechanism through
membranes (Figure 1(a)) aiming at: parameter estimation, when experimental
data for standard solute/solvent systems are available, and it is desired to
estimate relevant parameters for as yet uncharacterized systems; and prediction,
when the model parameters for the solvent-solute are available, and modelling
can be applied to describe performance of a different solute/solvent system [1,3].
Most of these studies have used flat sheet membranes.
On the other hand, there is lack of studies on the fluid dynamics and mass
transfer characteristics in OSN spiral-wound modules (Figure 1(b)), and attempts
to implement fundamental transport models on a process level (Figure
1(c)) are
few, and are based on simple, non-predictive membrane transport models.

This
paper reports a systematic comparison of different transport models using
selected experimental data for various solutes, solvents and membranes [3].
Solution-diffusion, pore-flow and transient models were used to perform
regression of experimental data, and prediction, based on the regressed model
parameters. It was observed that solution-diffusion
models described permeation through glassy membranes better than pore-flow
models, against both positive and negative rejection data. Afterwards,
the classical solution-diffusion model combined with the film theory was
applied to model the performance of different rubber-coated spiral-wound
modules (1.8?x12?, 2.5?x40? and 4.0?x40?) under various pressures and retentate
flowrates. Correlations for the fluid dynamics and mass transfer in the modules
were derived by regression of experimental data. Finally, OSN
unit operations at membrane and module levels were made available in the Aspen
Plus process modelling environment, to streamline process design using the
so-called ?OSN Designer? [2]. Matlab routines for the relevant
models were interfaced to Aspen Plus by means of CAPE OPEN and custom OSN unit
operations for common batch and steady-state membrane processes were developed.
The suitability of this technique was demonstrated by comparing the model
simulation to the experimental data for both ideal and non-ideal solutions.

References

[1] P. Marchetti,
M.F. Jimenez Solomon, G. Szekely, A. Livingston, Chem. Rev. 114 (2014)
10735

[2] D. Peshev,
A. Livingston, Chem. Eng. Sci. 104 (2013) 975

[3] P.
Marchetti, A. Livingston, J. Membr. Sci. 476 (2015) 530