Use of Dynamic Modeling for Wastewater Treatment Plant Acceptance Testing

Spokane County, Washington recently built a new wastewater treatment plant in order to connect tens of thousands of households/businesses previously served by septic tanks and leach-fields, and to facilitate further growth of the community due to future capacity limitations of the existing City of Spokane Riverside Park Facility. The new 8 mgd annual average flow facility, the Spokane County Regional Water Reclamation Facility (SCRWRF), discharges to the Spokane River and is designed to achieve one of the lowest effluent TP limits in the country at 50 μg TP/L on a seasonal average basis. This paper discusses the use of dynamic process modeling for confirming the process treatment capabilities of this facility under a DBO contract. A process model was setup in the BioWin™ software package to simulate the performance of the Spokane County RWRF during the periods of March to August 2012 with particular focus on the model calibration period of July, the model validation period on the 8th and 9th of August and the Acceptance Test during the months of April and May. This calibrated and validated model was then used to simulate plant performance under maximum month contract flow and load conditions in April and average contract flow and load conditions in May. The wastewater temperatures used in the model are what were actually experienced at the plant in April and May (average of 15.5 C), where the plant was designed based upon a minimum wastewater temperature of 17 C.In the calibrated model, it was necessary to:Decrease the rates of both the nitrifiers (autotrophs) and heterotrophs by increasing the Biowin default Kp value from 0.001 to 0.007.Further decrease denitrification rates by decreasing the anoxic reduction factor (default of 0.5) to 0.3. This was done since Biowin does not allow separate adjustment of the Kp values for Heterotrophs and autotrophs, and the impact of low phosphorus was felt to be higher on the higher metabolic rate heterotrophs.The autotrophic half saturation constant for oxygen (Koa) was increased from the Biowin default of 0.25 to 0.5. This makes the autotrophs more sensitive to oxygen levels, which is appropriate in a high MLSS system such as an MBR.It was not feasible to diurnally measure the influent characteristics during the entire acceptance test period. Thus the validation acceptance was based on a statistical analysis of the variation of effluent ammonia and nitrate (from on-line measurements) against model results for three parameters:Daily maximum permeate concentrations,The time of at which the daily maximum occurred and,The daily time-weighted average of the permeate concentration.The resultant model was found to have acceptable predictive power when the actual validation runs were compared to the model results. This model was then used with the contract maximum month and average contract loads at spring time temperatures to determine performance, while maintaining the MLSS at levels below those actually used at the plant. It was found that, without carbon, the plant was able to meet the seasonal average (April, May) ammonia requirement of 1 mg/L, but was not able to meet the TN requirement of less than 10 mg N/L. A separate spring run was then made with carbon addition (300 gpd of glycerol), which resulted in a seasonal average effluent ammonia of 1 mg/L and TN's of 8.4 mg/L in April and 8.6 mg/L in May, both less than the contract requirement of 10 mg TN/L.Based on these results the SCRWRF is expected to be able to meet the contract process requirements for springtime performance. Subsequent full-scale acceptance tests were conducted in the summer of 2013 to confirm the ability of the SCRWRF to meet the summer performance requirements shown in Table 1.