For inland reverse osmosis (RO) systems, brine disposal is often difficult and costly, and there is significant interest in increasing recovery and reducing brine volumes. The City of Buckeye, AZ owns brackish wells that have trace amounts of arsenic, nitrate, and various metals, and is considering RO treatment for these wells. Desktop modeling predicted that sulfate and silica compounds would govern scaling, and that solubility limits would theoretically allow recoveries greater than 90%. In this study, a pilot-scale system was used to understand the scaling regime and investigate how conventional vs. advanced flow configuration affected achievable recovery.The pilot-scale system included two RO trains in parallel. One train was a conventional three-stage array of 6:4:2 pressure vessels with 3 elements each and a booster pump at each stage. The other train was a closed-circuit RO (CCRO) system with a single 3 element stage and a concentrate recirculation loop. Permeate flux was 9-11 gfd, and the trains took flow from a well with water on the higher end of the local TDS range (averaging around 2,700 mg/L). In the first phase of testing, recovery was increased incrementally from 80% to 92% (for the conventional system) and 80% to 95% (for the CCRO system) over one month. One element was taken from each system for autopsy, after which the conventional system was run at 92% recovery and the CCRO system at 93-94% recovery continuously for one month. Vitec 7400 antiscalant was used without acid addition, because lower pH would increase the risk of silica scaling.The conventional system showed stable normalized flux at every recovery up to 92%, at which point declining flux due to scaling was observed. The CCRO system achieved stable normalized flux up to 94% recovery, at which point scaling was observed. The slightly higher recovery achieved by the CCRO system appears to be attributable to its higher cross-flow velocity under all conditions. Flux declined in the conventional RO system for cross-flow velocities less than 0.080 ft/s, while the CCRO system reached approximately the theoretical scaling limit. However, at higher recovery, the stronger concentrate in the loop increased concentrations of dissolved constituents in the CCRO permeate, while the conventional system consistently achieved excellent removal of TDS, arsenic, nitrate, and metals. The membrane autopsies confirmed that sulfate compounds and silica were present in the scale.This study showed that, through understanding and control of the scaling chemistry as well as the flow conditions in the pressure vessel array, aggressive recovery rates of 90% or more could be achieved in both a conventional and a CCRO system. This finding is important to the viability of RO treatment in inland regions where brine volume must be minimized.
This presentation is available to AMTA Members only.
- Michael Adelman
- Stantec Consulting Services, Inc.
- AMTA/AWWA Membrane Master Class Webinar II - Part 1, Online
- AMTA/AWWA Membrane Master Class Webinar Series II
- Reverse Osmosis, Scaling, Cross-Flow Velocity