Enhancement of Operational Flexibility in Reverse Osmosis Membrane Processes by Concentrate Recycling

Improvements in system and process design has enabled reverse osmosis (RO) membrane technology to gain foothold in various water treatment and desalination applications. However, energy consumption and mineral scaling on membrane surfaces remain impediments to high recovery operation of RO desalination. Accordingly, research efforts have been devoted to process optimization to improve permeate water productivity, reduce energy cost, and mitigate membrane mineral scaling. In this regard, RO concentrate recycling is an effective approach for enhancing product water recovery, reducing system footprint, and lowering installation as well as operating costs of RO desalination systems and plants. Energy consumption and membrane mineral scaling propensity in RO processes are impacted primarily by the osmotic pressure magnitude and the level of supersaturation of mineral scalants at the membrane surface, respectively, which are, in turn, governed by the operational strategy. Optimization of RO processes with concentrate recycle require fundamental models of RO processes with concentrate recycle under both steady and unsteady-state operation, To date, only simple models have been proposed which are of limited applicability to practical systems given the use of various simplifying assumptions of complete energy recovery, neglect of the efficacy of concentrate flushing in unsteady-state semi-batch RO (SBRO), and omission of concentration polarization. Previous studies have not provided experimental data to corroborate conclusions regarding system performance based on oversimplifications. Moreover, SBRO operation was not assessed relative to steady state RO (SSRO) with partial recycle (SSRO-PR) which is also suitable for high recovery operation with a small footprint. Accordingly, in the present work a fundamental quantitative modeling framework was developed and implemented for the design and operation of SSRO-PR and SBRO. Modeling RO desalting, including at the limit of the thermodynamic restriction, was undertaken to evaluate the minimum energy consumption as a function of product water recovery. SBRO process analysis revealed a significant increase in the salinity range during the RO filtration period and a progressive rise in the initial filtration period salinity until the stable cycle-to-cycle operation is reached. SBRO performance is highly dependent on the efficacy of concentrate discharge and flushing during the SBRO flushing period. For the condition of concentrate flushing under ideal plug flow, energy consumption in SBRO was assessed to be lower than for single-pass RO (SPRO) operation at the same level of overall product water recovery. However, for the practical range of expected concentrate flushing efficacy energy consumption in SBRO could be significantly above that which would be attained in SPRO. Using the direct real-time membrane surface optical imaging mineral scaling was experimentally evaluated in both SBRO and SSRO-PR pilot systems, with respect to the efficacy of concentrate flushing with the undersaturated raw feed water, at given levels of supersaturation or product water recovery. The predicted RO element feed stream osmotic pressure and solution supersaturation levels (for the target mineral scalant) at the membrane surface, for a given product water recovery, were higher on average in SBRO relative to SSRO-PR as corroborated by experimental data. The experimental data revealed mineral scaling propensity, which was significantly higher in multi-cycle SBRO operation, relative to SSRO-PR at the same water recovery. However, the rates of crystal nucleation and growth were similar at when both systems are compared at the same level of average supersaturation, although the product water recovery was lower in SBRO compared to SSRO-PR.The present theoretical modeling framework, validated by experimental data, provide a fundamental approach to assessing the performance of SBRO and SSRO-PR desalting systems with respect to energy consumption and mineral scaling propensity. The presented approach provides the means necessary for optimizing these low footprint technologies for high recovery RO desalination