Developing A Design Basis of Pathogen Removal Across MBRs In The Absence Of A Regulatory Requirement

INTRODUCTION
The Morro Bay Water Reclamation Facility (WRF) is designed to meet multiple effluent goals: 1) finished water for indirect potable reuse through groundwater recharge, and 2) discharge to the Pacific Ocean. This is first WRRFs which employs a biological nutrient removal membrane bioreactor (BNR MBR) process for wastewater treatment directly followed directly by reverse osmosis (RO) in the advanced water purification facility (AWPF). MBRs have a long and successful history of use in water resource recovery facilities (WRRFs). Despite the widespread application of MBRs for non potable reuse, there are currently no well-established potable water applications employing MBRs. Indirect Potable Reuse (IPR) regulations stipulate, among other requirements, that a WRRF should be designed (and operated) to achieve at minimum 12-log, 10-log and 10-log reduction of enteric viruses, Giardia cysts and Cryptosporidium oocysts, respectively. The attenuation of these pathogens in indirect potable reuse is achieved through the entire process flowsheet comprising, typically, of biological treatment and advanced water treatment (i.e., RO and UV AOP) processes. The pathogen removal credits which may be attributed to an MBR has been the subject of extensive research by academics and industry practitioners alike. The most extensive longitudinal review done by the Australian Water Recycling Centre of Excellence. The preliminary observations from this study resulted in a tiered approach to assigning pathogen removal credits to MBRs; with the lowest tier offered a rather conservative set of credits. The presentation will share an approach to a cogent basis of design and a validation approach for utilities to employ as part of the program development and startup/testing of a potable reuse system. The goal of this research was to evaluate the anticipated pathogen removal across the MBR based on data from other operating full scale and pilot MBR facilities. The objective was to fill key knowledge gaps related to the potential pathogen removal benefits offered by a BNR MBR process.
METHODOLOGY
An extensive data evaluation and analysis of pathogen removal data from existing full-scale and prior pilot scale facilities (Table 1.1) was conducted. Data collection and analytical methods for each reference and site were reviewed and evaluated. The data were curated based on the analytical methods, reported detection limits, and the reported filtrate turbidity. The resulting data sets were analysed using small-sample evaluation statistical techniques. Two approaches were employed to evaluate the longitudinal LRV data: Longitudinal LRV Monte-Carlo Sampling with Pair-wise Data and Influent-Filtrate Monte-Carlo Sampling with LRV Calculation. The probability distribution function parameters describing the LRV distribution for each pathogen were employed to generate predictions for LRV distributions for pathogen attenuation across the Morro Bay WRF MBR system using a Monte Carlo sampling strategy. The 5th percentile value of the Monte Carlo generated LRV distribution was used as a basis of design for the conditional LRV.
PRELIMINARY RESULTS AND DISCUSSION
A total of 40 virus LRV data from four separate BNR MBR facilities – three full scale and one pilot system – were analysed. Evaluation of the data suggests that all the LRV data for Enterovirus from the sites may be considered a single data set and all the Norovirus groups may be analysed as a single data set (both, Mann Whitney alpha = 0.05). The reported Norovirus LRVs based on the data from the four sites (N = 32) range from 2.74 to 8.16 with a mean value of 5.63 ± 0.25 (Figure 1.1). The LRVs from Enterovirus (N = 8) range from 3.56 to 8.68 with a mean value of 6.91 ± 0.72. Both data sets satisfactorily fit a normal distribution (evaluated using the Anderson Darling (AD) statistic with alpha = 0.05). A total of 24 LRV data for protozoa from three separate facilities were evaluated. The data represent operation of the respective BNR MBR systems at a range of conditions of flux, TMP and mixed liquor concentration. In addition to the data for LRVs of Giardia and Cryptosporidum, 165 LRV data were provided for attenuation of Clostridium Perfringens (C. Perfringens) spores across Zeeweed MBRs. These data are from a total of four sites. The reported Giardia LRVs based on the data from the three sites (N = 12) range from 3.30 to 5.97 with a mean value of 4.77 ± 0.85 (Figure 1.2). The reported C. Perfringens LRV data from the four sites (N = 165) range from 3.91 to 6.18. Taken together, the analyses suggest that the observations and results from the AWRCE efforts may be significantly conservative. Furthermore, evaluation of the data collection and analytical approaches suggests that there may be a significant left censoring challenge when analysing protozoa data and the actual performance of MBR systems in attenuation of protozoa may be significantly higher than reported.