RO Fouling Assessment of MBR-RO Trains using Flat-Plate MBR Membranes

There is growing interest in using membrane bioreactors (MBR) for potable reuse because of their improved effluent quality and pathogen removal compared to conventional secondary processes. Therefore, using an MBR as the main treatment process upstream of reverse osmosis (RO) is often desirable for advanced treatment facilities. Most advanced treatment studies with MBRs employ hollow fiber or tubular membranes. In this 12- month pilot study, an alternative flat-plate MBR was employed as the pretreatment method to assess downstream biological and colloidal RO fouling in a MBR-RO train. The study evaluated RO performance under varying operational parameters such as MBR and RO flux conditions, biological process conditions in the MBR, and chloramine dose in the RO feed. The RO system was also subject to known MBR membrane breaches. The two-stage RO system was operated at ~70% recovery to avoid mineral scaling so the effects of fouling driven by MBR filtrate quality could be observed. Typical RO permeate flux values used in reuse applications (10-14 gfd) were used in this study. Key monitoring parameters included MLSS, TMP, and filtrate water quality for the MBR as well as normalized RO specific flux. A novel online integrity monitor was also used to detect small colloidal particles in the RO feed using differential pressure across a filter. RO membranes were also periodically sent for autopsy to evaluate the type of fouling observed on the RO membranes prior to a clean-in-place. Normalized RO data has shown evidence of periods with both higher and lower rates of fouling. RO flux decline data was plotted against key MBR operating and filtrate quality parameters. The MBR filtrate turbidity and integrity monitor data showed notable correlations. Lowest RO fouling rates (< 0.0003 gfd/psi-day) were observed with lower filtrate turbidity and lower differential pressure from the integrity monitor. Higher fouling rates (> 0.0003 gfd/psi-day) took place at higher MBR permeate sCOD (>30 mg/L) and low chloramine levels, both of which would favor biofilm growth as a driver of fouling. However, autopsy results showed mostly organic material as the foulant with minimal biofilm coverage suggesting that colloids were the main driver of fouling with biological growth as a secondary factor. Normalized RO data was evaluated individually for stage 1 and stage 2. Conventional knowledge would suggest that colloidal fouling in RO systems is more prevalent at the first stage than in subsequent stages. However, the results from this study suggests that the RO fouling pattern is similar between stage 1 and stage 2 membranes, as shown in Figure 1. The performance of the MBR-RO pilot using flat-plate MBR membranes was compared to existing data from MBR-RO treatment processes using hollow fiber or tubular MBR membrane, which suggests that the type of MBR membrane used upstream of the RO system does not impact the overall performance of the RO system. Table 1 provides RO performance data from two reference facilities compared to the performance observed during this pilot study. This presentation will provide insights into the drivers of RO fouling as well as operational strategies for improved RO performance, and will compare the performance of this pilot study against existing process data from other advanced water treatment plant facilities. Target values for MBR filtrate turbidity, MBR tank MLSS, and feed chloramines were identified as parameters that promote low fouling rates and reasonable CIP intervals. The novel integrity monitor was shown to provide useful insight into whether colloid passage was sufficient to increase RO fouling rates.