Biological Phosphorus Removal Potential Test

Evaluation of the BPR Potential Test

Using an SBR to Validate the BPR Potential Test

The BPR potential test was validated by operating a bench-scale SBR and comparing the effluent phosphorus concentrations measured from the SBR to that predicted by the BPR potential test. Wastewater obtained from the Oakfield WWTP was used as influent. The initial PAO-containing activated sludge with 7.5% of phosphorus content in volatile suspended solids (VSS) was obtained from the Nine Springs WWTP. The wastewater sample taken on June 5, 1997, was used as influent feed for the first 9 days, and the wastewater sample taken on June 12, 1997, was used for the remainder of this experiment. The SBR was run for 18 days. The results of SBR performance are shown in Figure 13.

Figure 13. BOD5 and Phosphorus concentration in SBR with Oakfield wastewater.

The phosphorus concentration was higher in the effluent than in the influent until the 9th day due to release of phosphorus from the PAO-containing activated sludge with P/VSS content of 7.5%. Only a small fraction of phosphorus was removed in the last 9 days of the experiment. During the experiment, the P/VSS content decreased from 7.5% to 4.5%. PAOs did not seem to prosper in the SBR fed with Oakfield wastewater. The phosphorus release/uptake profile obtained using the sludge taken on the 18th day is shown in Figure 14. It can be seen that phosphorus was not removed in the aerobic zone below the initial concentration, although there was slight release of phosphorus (~2.4 mg P/L) in the anaerobic stage.

Figure 14. Phosphorus release/uptake profile in SBR with Oakfield wastewater.

Based on our BPR potential model (Equation 2), the predicted effluent phosphorus concentration was 3.9 - [7.3 ´ 0.15] - {[(5 ´ 5) + 90] - 1 ´ 93} = 2 mg TP/L (see Tables 3 and 4). The average effluent soluble phosphorus concentration in the SBR after 9 days of acclimation period was 1.7 mg P/L, which is similar to the predicted value from the BPR potential test.

The BPR potential test was validated by comparing the test result with the performance of the SBR test using the Oakfield wastewater sample. Both tests demonstrated that BPR alone would not meet the effluent phosphorus discharge permit for the Oakfield WWTP.

Effect of Sludge Characteristics on the BPR Potential Test

From the behavioral pattern of phosphorus release under anaerobic conditions in a BPR system, Wentzel et al. (1985, 1990) proposed the following hypotheses:

1. Only readily biodegradable soluble COD (S_bs) can be converted to SCFAs, which are then taken up by PAOs under the anaerobic conditions.

2. The conversion is mediated by non-PAOs in the anaerobic zone.

3. All S_bs converted is immediately taken up by PAOs, i.e., the rate of phosphorus release is controlled by the rate of conversion.

4. The rate of conversion is considered first-order with respect to non-PAO biomass concentration (X_ahn) and the readily biodegradable COD concentration (S_bs):
   
   dS_bs/dt = - K X_ahn S_bs,
   
   where K = first-order rate constant.

5. The rate of phosphorus release is assumed to be stoichiometrically related to the mass of the converted S_bs uptake:
   
   dP/dt = - C_sp (dS_bs/dt) = C_sp K X_ahn S_bs
   
   where C_sp = stoichiometric ratio between phosphorus release and converted S_bs uptake.

Based on kinetic studies of biological phosphorus release (Wentzel et al. 1985, 1990), the phosphorus release rate is determined indirectly by both the characteristics of wastewater and the active biomass. In addition, Mino et al. (1987) indicated that the ability of sludge to take up acetate anaerobically is limited by the amount of polyphosphorus stored in the cell when the phosphorus content of sludge is < 35 mg P/g VSS.

In the BPR potential test, a basic assumption is that the characteristics of PAO-containing sludge will not affect the outcome. Therefore, the effect of sludge characteristics on the BPR potential test results must be evaluated. PAO-containing sludge from different types of BPR systems (modified UCT and SBR), different wastewater treatment plants (Nine Springs and Beloit), and different sludge concentrations (MLVSS = 1,190 and 2,250 mg/L, both PAO-containing sludges obtained from the Nine Springs WWTP) were tested with wastewater from the Nine Springs WWTP.

Figure 15 shows the phosphorus release for sludge in the modified UCT and SBR BPR systems fed with the Nine Springs wastewater. Both sludges were acclimated with the Nine Springs wastewater, and both BPR systems could remove phosphorus to below 0.5 mg TP/L. The phosphorus contents in sludge of the modified UCT process and SBR were 6.2% and 5.5%, respectively. Figure 15 shows that there was no difference in phosphorus release response in the first 60 minutes. After 2 hours the difference in phosphorus release was approximately 5 mg P/L. This difference may be due to the different sludge concentration (1,340 mg VSS/L for the modified UCT-process sludge and 1,150 mg VSS/L for the SBR sludge).

Figure 15. Phosphorus release profile for PAO-containing sludge from UCT and SBR systems.

Figure 16 shows the phosphorus response of PAO-containing sludges obtained from different WWTPs mixed with wastewater from the Nine Springs WWTP. The Beloit treatment plant has an average design flow of about 7 MGD. The plant receives significant industrial loadings, primarily from the food processing and organic chemical industries. The Beloit WWTP is currently removing phosphorus from 17 to 3 mg TP/L by using an anaerobic/anoxic selector followed by aeration basins. The Nine Springs WWTP is currently removing phosphorus to below 0.5 mg TP/L using a variation of the UCT process. From Figure 16, the sludges obtained from the Beloit WWTP released phosphorus faster than those from the Nine Springs WWTP in the first 90 minutes of operation, but slower after that. The faster phosphorus release of the Beloit sludge in the first 90 minutes might be due to its higher sludge concentration. The sludge concentrations for the Nine Springs sludge and Beloit sludge during this test were 1,150 and 2,580 mg/L, respectively. The lower total amount of released phosphorus of the Beloit sludge may be due to its lower phosphorus content. The phosphorus contents in the Nine Springs sludge and the Beloit sludge were 7.3% and 4.3%, respectively.

Figure 16. Phosphorus release profile for PAO-containing sludge from the Nine Springs and Beloit Wastewater Treatment Plants.

In order to evaluate the effect of sludge concentration on the BPR potential test, two batch reactors containing the same wastewater and PAO-containing sludge obtained from the Nine Springs WWTP at different sludge concentrations were conducted. Figure 17 shows the phosphorus release profiles of these two batch tests. Phosphorus was found to be released faster at a higher sludge concentration than at a lower sludge concentration. The total amount of phosphorus released was slightly greater at a higher sludge concentration at the end of test, as well. However, the total phosphorus release is anticipated to be the same when the test is conducted longer than 2 hours.

Figure 17. Phosphorus release profile for different PAO-containing sludge concentrations (both sludges from the Nine Springs Wastewater Treatment Plant).

The P release vs. time plot can be modeled empirically as follows (Wentzel et al. 1985):

(3)

where

P_t = phosphorus concentration at time t, mg P/L;
P_max = maximum potential phosphorus concentration, mg P/L;
P₀ = initial phosphorus concentration, mg/L;
P_max - P₀ = maximum potential release of phosphorus, mg P/L; and
k = first order rate constant.

A nonlinear regression program written with Fortran language (Chen 1996) was used to estimate empirical model parameters, P_max and k. This program basically followed the Gauss-Newton method (Bates and Watts 1988) using a linear approximation to the expection function to iteratively improve an initial guess q0 for q (parameters) until there is no change. This program can also plot the approximate 95% inference regions for parameters automatically (Bates and Watts 1988). The accuracy of estimated results from this program was comparable with the results estimated from XLISP-STAT (Tierney 1990) and SYSTAT® (SPSS, Inc. 1996) commercially available statistical programs.¹

Table 5 summarizes the estimated P_max and k values by using Chen’s program to fit Equation 3 with experimental results of these six batch tests.

The fitted values, P_t, can be obtained from estimated parameters (Table 5) and Equation 3. An approximate 95% confidence interval of fitted values can also be obtained (Bates and Watts 1988, Seber and Wild 1989). The confidence intervals of fitted values at 2 hours for each experiment can be constructed, and the fitted values of each test can be compared. Table 6 summarizes the calculation results of fitted values and the approximate 95% confidence intervals. From the comparison of the Pt values at 2 hours, no significant difference between the different sludges was found in all three tests. This implied that the effect of sludge characteristics on the BPR potential test will be insignificant if the total phosphorus release were measured at 2 hours. However, from Figure 17, the effect of sludge concentration on biological phosphorus release will be considered to be significant if the total phosphorus release were measured at 1.5 or 1 hour. Therefore, the BPR potential test must be conducted at the real MLVSS concentration used at the wastewater treatment plant of interest.

Table 6. Calculation results of fitted values and approximate 95% confidence intervals (C.I.).

	Fitted value Pt at time = 2 hrs	Approximate C.I.
Test 1
Modified UCT sludge	49.3	49.3 ± 2.9
SBR sludge	44.8	44.8 ± 3.1
Test 2
Nine Springs sludge	41.3	41.3 ± 2.9
Beloit Sludge	37.9	37.9 ± 2.7
Test 3
Low MLVSS sludge	41.9	41.9 ± 3.1
High MLVSS sludge	45.7	45.7 ± 2.3

Figure 17. Phosphorus release profile for different PAO-containing sludge concentrations (both sludges from the Nine Springs Wastewater Plant).

Effect of Wastewater Storage on the BPR Potential Test

Since properties of a wastewater may change during transportation and storage before the BPR potential test is conducted in a laboratory, an experiment was conducted to evaluate the effect of storage time at 4°C on the BPR potential test. Profiles of the phosphorus release in the BPR potential test with three different storage times are shown in Figure 18.

Figure 18. Effect of wastewater storage duration on BPR potential test.

It can be seen that the phosphorus release was not significantly affected when the wastewater was tested within 24 hours. Therefore, a wastewater sample of interest can be tested without significantly affecting the BPR potential test within 24 hours if it is preserved at 4°C.

¹ Mention of specific products does not constitute goverment endorsement.

More information on this topic: Gerry Novotny