Five Years after Start-up of Sidestream Fermenter at the Wakarusa River WWTP: The Bad and The Good.

Introduction At WEFTEC 2019, data from one year of operation of a newly built sidestream fermenter for a biological phosphorus removal (BPR) system was presented. This paper covers the first 5 years of operation of the Wakarusa River wastewater treatment plant (WWTP) at Lawrence, Kansas. New processes are always scrutinized by the industry. Was the pilot operation a fluke? Can its performance be repeated? Is the system lightly loaded? Is the process resilient and reliable? Data from the last 5 years shows that sidestream fermentation is reliable, but there were periods of degraded performance. The Wakarusa River WWTP was designed to treat 2.5 mgd (9 500 m3 /d) annual average flow of domestic sewage. The Wakarusa facility is essentially a new train located at another site for the Kansas River WWTP. Flow to Wakarusa has averaged about 2 mgd over the 5-year period (80 percent of the design capacity). The Wakarusa Facility is a 3-stage Bardenpho process with a flow-through anaerobic zone, an anoxic zone, and an oxidation ditch for aerobic treatment plus a sidestream fermenter for enhanced biological phosphorus removal (EBPR). The raw influent to Wakarusa is fairly 'fresh' with a little VFA (12 mg/L) and a low rbCOD to TP ratio (50% of the time less than 13.5). Not enough rbCOD was available for good BPR performance. The Bad: Impact of Return Flows on EBPR Performance Figure 1 shows the effluent orthophosphate (OP) concentrations from an online analyzer, and effluent total phosphorus (TP) concentrations from operator collected data and 24-hour composites for 2019. This graph contains over 150,000 data points from the on-line analyzer. Figure 2 shows the 30,000 points spanning from Nov 2019 to Sept 2020 and it is much easier to see the periods of good effluent from the not so good effluent quality. The permit limit is 1 mg/L TP on a 365-day rolling average. The Figure 1 graph show periods, highlighted by the pink transparent boxes, of high effluent P coinciding with periods when biosolids land application was delayed by excessively wet fields. This pattern repeats each year. Solids build up in the plant MLSS (or system increased SRT) and the basin MLSS concentration rises significantly which interferes with sidestream fermentation efficiency. A process model was developed in SUMO (Dynamita) to simulate the plant performance at increasing SRT conditions. Model simulations indicated that the VFA production in the fermenter decreased as the SRT in the activated sludge basin increased while maintaining a constant flow to the fermenter. Model simulations are continued to further explore this phenomenon to determine if operational conditions can improve fermenter performance. In 2022 there was an upset between Sept and Nov in system performance. The centrifuges had been in operation for 4 years and erosion/ wear of internals resulted in poor solids capture. The net result was poor BPR performance in Sept through Nov of 2022, highlighted on Figure 1. Poor solids capture resulted in the recycling of high concentrations of BOD, TSS and TP to the plant influent (Figure 3). Once the centrifuges were repaired and solids capture was under control, the effluent OP and TP returned to very low concentrations. After 5 years of operation, no ferric chloride has been added nor has there been any supplemental addition of a carbon source. After the high inventory of MLSS was reduced, the fermenter bounced back and produced a very low effluent OP concentration. The Good: Impact of Low DO Operation on TN and TP Removal Performance and Microbial Ecology Figure 4 shows the nitrogen profile in the system during the summer of 2022. The abrupt decline in the effluent nitrate concentration was intentional. The facility has two online DO probes located in the outer curve and the inner curve of the ditch system. The DO setpoints were initially set at 0.8 mg/L and 0.5 mg/L during start-up and during the summer of 2022 plant staff lowered the DO setpoints to around 0.2-0.3 mg/L. Effluent nitrate concentrations dropped to below 1 mg N/L. The effluent phosphorus concentration remained very low (Figure 1). During the July — August 2022 timeframe, the effluent OP and TP concentrations remained very low. Near the end of this trial period, centrifuge performance began to deteriorate and excessive solids in the system interfered with good BPR. These results indicate that low DO operation did not have a negative impact on phosphorus uptake as well as nitrification. As the DO setpoints were lowered, the AOB (ammonia oxidizing bacteria) population declined and the population of Commamox increased. Recent studies demonstrated the prevalence of Comammox Nitrospira can be dominant ammonia oxidizers in a mainstream low DO nitrification reactor. The Good: Impact of Fermenter SRT on VFA Species and PAO/GAO Population A study performed by University of Kansas on fermenter SRT and operating parameters showed the impact of SRT on the quantity of VFA formed and the diversity of VFA species (Figure 5). Acetic acid is still the predominant VFA present in the fermenter effluent. Figure 6 shows the PAO versus GAO as a fraction of the overall microbial population in the anaerobic zone. Very few GAOs were present in the system. Summary The sidestream fermentation process has been reliable and consistent in producing a low effluent OP concentration while at plant loading of 80% of design capacity