Multi-stage nanofiltration for brine concentration: experimental and modelling study

P. Pasqualin, P.A. Davies*

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

91 Downloads (Pure)

Abstract

Multi-stage nanofiltration (NF) has been proposed in several configurations for brine concentration applications where reverse osmosis (RO) is impractical due to high osmotic pressures. For example, the nNF-RO configuration, comprising several NF stages followed by a RO stage, has been proposed for liquid desiccant air-conditioning (LDAC). The RO-nNF configuration, comprising a RO stage followed by several NF stages, has been proposed for zero liquid discharge (ZLD). Previous theoretical investigations of nNF-RO and RO-nNF were based on idealised assumptions. Here, a two-stage system has been constructed and tested. By accounting for energy and pressure losses, more accurate models were developed. Experimental and model results agreed within 11%. The models were extended to four-stage systems, representing 3NF-RO and RO-3NF, and applied to two cases: LDAC using MgCl2 for greenhouses and ZLD using NaCl brine approaching crystallisation. A range of feed concentrations, energy recovery devices, and solution temperatures were modelled. The four-stage systems have SECel of 45–140 and 4–40 kWhel/m3 in LDAC and ZLD, respectively. In comparison, their thermal alternatives require 120 and 25 kWhel/m3, respectively. Thus, compared to thermal brine concentration, multistage NF is consistently more energy efficient in LDAC and more efficient at feed concentrations below 90 g/L in ZLD.
Original languageEnglish
Article number116928
JournalDesalination
Volume566
Early online date22 Aug 2023
DOIs
Publication statusPublished - 15 Nov 2023

Keywords

  • Multi-stage systems
  • Water recovery
  • Nanofiltration
  • High brine concentration
  • Zero liquid discharge
  • Sensitivity analysis

Fingerprint

Dive into the research topics of 'Multi-stage nanofiltration for brine concentration: experimental and modelling study'. Together they form a unique fingerprint.

Cite this