Guozhong Shi, Yaqin Zhang, Hai Huang, Yingfeng Lai, Lin Zhang
Nanofiltration (NF) has emerged as a key technology for high-salinity brine treatment, providing an effective front-end for water reuse and salt valorization toward zero liquid discharge (ZLD). However, NF performance in practice is often constrained by the inherent trade-off between water recovery and ion selectivity and further degraded by concentration polarization and feed-quality fluctuations, which increase energy consumption and compromise stable salt fractionation. To address this issue, this study develops and optimizes an integrated NF system consisting of three sub-units arranged in a two-stage, two-pass recirculating configuration to enable cooperative Cl − /SO 4 2− separation and high water recovery. Here, NF system design is integrated with a mechanistic membrane-separation model, namely, the concentration-polarization coupled with Donnan steric pore model (CP–DSPM). An operation–performance design framework is established for membrane allocation and pressure regulation, and verified by both modeling and experiments over a wide range of feed compositions. The integrated system selectively enriched SO₄ 2− in the concentrate (98.5–99.7% for SO₄ 2− rejection) while facilitating the transport of Cl − into the permeate (−13.7% to 1.1% for Cl − rejection), together with an overall water recovery exceeding 80%. Furthermore, supported by both theoretical modeling and experimental verification, increasing the feed c(Cl − )/c(SO₄ 2− ) was found to not only enhance the NaCl recovery but also reduce the specific energy consumption (SEC), demonstrating the effectiveness of the CP-DSPM-guided framework in coupling separation selectivity with energy efficiency. Overall, this work provides a mechanistically grounded and engineering-feasible pathway to translates membrane-level transport model into robust system-level NF performance, which supports future adaptive NF operation through real-time sensing and predictive optimization.