Maximum growth and decay rates of autotrophic biomass to simulate nitrogen removal at 10 ° C with municipal activated sludge plants

The present study aims at determining most likely values for the maximum growth rate (μA, max) and the endogenous decay rate (bA) of nitrifiers for activated sludge processes treating municipal wastewater operated at low temperature (10°C). The work used nitrification rate data measured on 10 full-scale plants and 2 pilot plants fed with domestic sewage. This set of data was combined with a modelling and a theoretical approach. The unified values (μA, max = 0.45·d and bA = 0.13·d -1) were obtained at 10°C for the kinetic parameters of the autotrophic biomass in the SRT range 10 to 50 d. In addition, the factors affecting the expected nitrification rate (rv, nit) were established by a theoretical approach and confirmed by experimental results. For a given SRT, a linear relationship with the nitrogen volumetric loading rate was shown. The COD/TKN ratio of the influent on the nitrification rate was demonstrated. Finally, an operational tool for the verification of the nitrification rate in the design procedure of activated sludge processes is proposed.


Introduction
Activated sludge functioning with intermittent aeration is the main process used to remove nitrogen from municipal wastewaters in France.The size of the aeration tank is usually determined by the minimum sludge retention time (SRT) at 10°C.The nitrification rate is, however, useful to calculate the duration of oxygen presence required to nitrify a given amount of nitrogen and to reach an ammonia discharge objective.Reported values in the literature at 10°C range between 1.0 to 4.4 mg N nit •gMLVSS -1 •h -1 (Palis and Irvine, 1985;Oleszkiewicz and Berquist, 1988;McCartney and Oleszkiewicz, 1990;Burica et al., 1996) due to differences in the organic and the nitrogen loading rates or in the COD/TKN and COD/TSS ratios of the influents (Al-Sa 'ed, 1988;Thiem and Alkhatib, 1988).For some time now it is recommended to use nitrification rate values expressed in mgN nitrified •ℓ -1 •h -1 instead of mg N nit •gMLVSS -1 •h -1 that can be determined through modelling.
The Activated Sludge Model No. 1 (ASM 1 (Henze et al., 1987)) has been widely used for about 20 years to assess, predict or optimise the nitrification capacity of biological nutrient removal (BNR) wastewater treatment plants (WWTP).For given influent characteristics and with chosen parameter values for the heterotrophic and the nitrifying biomass, the ASM 1 model calculates the nitrification rates of the mixed liquor, and the nitrogen concentrations in the effluent.However, this tool still requires bioprocess experience and time to be used correctly for design purposes.In particular, the use of validated parameter values for the conditions of the simulations is necessary (e.g.similar sludge retention time (SRT) and/or loading rate).These values are particularly important when the processes are simulated at limiting conditions.
Recent studies have shown that too low a value has been used systematically for the endogenous decay rate of nitrifiers (see default values of 0.04•d -1 at 20°C mentioned in Henze (1987)).This default value requires a specific value for the maximum autotrophic growth rate which depends on the applied SRT (Yuan et al., 1999;Dold, 2002;Lee and Oleszkiewicz, 2002).New values obtained from recent specific batch-test protocols under different sludges at 20°C were stable around 0.19•d -1 (Fillos et al., 2000;Stensel et al., 2002;Dold et al., 2005;Marrs et al., 2004;Lopez et al., 2006).For a temperature of 10°C, very few b A values can be found in the literature.Nevertheless, as Nevertheless, as these results come from several different laboratory experiments, the need to integrate these new parameter values has not been clearly demonstrated with on-site data.
The present study aims at determining most likely values for the maximum growth rate and the endogenous decay rate of nitrifiers at low temperature (10°C) for municipal activated sludge processes.The nitrification rate measurements in a pilot plant fed with domestic wastewater, and those from 10 full-scale BNR WWTPs are studied with both ASM 1 simulations and a theoretical approach.A practical design tool is then proposed to predict the maximum nitrification rate which can be expected at 10°C.

Methods for nitrification rate measurements
The maximum nitrification rate (r V, max nit , Eq. ( 1)) expressed in mgN nitrified •ℓ -1 •h -1 , is reached when the dissolved oxygen and the ammonia concentration are high enough compared to the halfsaturation coefficients, and when no inhibition conditions are applied.It can be directly measured by a specific experiment in a batch reactor, or estimated on-site by a mass balance on nitrogen species of daily flow proportional composite samples:

• Batch test reactor
The maximum nitrification rate is measured on sludge under non-limiting conditions: initial [NH 4 -N] over 15 mgN•ℓ -1 , DO over 4 mgO 2 •ℓ -1 .pH is monitored to be in the range 7 to 8 (see Fig. 1).In the mixed liquor, the nitrate concentration is measured every 10 min for 1 h.The slope of nitrate build-up (nitrate production rate: NPR) gives the maximum nitrification rate defined in Eq. ( 1): (1)
Considering that the assimilated nitrogen mass (ϕ Assim ) represents 5% of the removed BOD 5 mass (Henze et al., 1996), the amount of nitrogen nitrified (ϕ Nitrified ) during a 24 h period is obtained from the input TKN load (ϕ In ) and discharge (ϕ Out ).The actual nitrification rate is deduced dividing ϕ Nitrified by the aerobic time (∆t) and the volume of the aeration tank (V) as shown in Eq. ( 2): (2) When the ammonia concentration is above 5 mg N•ℓ -1 in the effluent and DO is above 3 mg•ℓ -1 , the obtained value is close to the maximum rate (Choubert and Racault, 2000).For ammonia concentration below 5 mg N•ℓ -1 in the effluent, the result obtained is approximately 1.5 times lower than the maximum rate (Oleszkiewicz and Berquist, 1988).The use of the Monod function and half-saturation constant (K NH = 1 mg N•ℓ -1 ) is necessary to obtain the maximum nitrification rate.

Nitrification rates database
Nitrification rate measurements have been performed on pilot plants and on full-scale WWTPs:

• Data from targeted pilot-plant experiment
The pilot plant consisted of a 115 ℓ tank that was completely mixed and intermittently aerated (9 cycles•d -1 ).It was connected to a 45 ℓ clarifier with a scraper, and a sludge recycling loop.The devices were controlled by a timer switch and operated in a temperature-controlled room at 10°C.Domestic sewage was B v (N) Nitrogen volumetric loading rate = daily applied nitrogen mass / volume of biological tanks (aerated + anoxic + anaerobic zones) f AT Fraction of the total sludge contained in biological tanks (intermittently aerated tank + anoxic + anaerobic zones) % F/M ratio Food to micro-organism ratio kg BOD 5 •(kgMLVSS Actual nitrification rate including oxygen and ammonia limitation mg

Figure 1 Diagram of the batch reactor
r -collected three times per week at a municipal WWTP and kept at 4°C.As it was sampled at 10:00 am which corresponds to the daily nitrogen peak load at the full-scale WWTP, the COD/TKN ratio of the influent was about 6.5.However, typical COD/TKN values for a daily influent composite sample are usually closer to 10.6 (Pons et al., 2004).
Two 6-month experiments were carried out with: 1 st series: Three predefined organic loading rate stages from F/M = 0.09 to 0.14 kg BOD 5 kg MLVSS -1 •d -1 applied for 6 weeks.The corresponding volumetric nitrogen loading rate (B v (N)) was in the range 160 to 220 mgN•ℓ -1 •d -1 and the SRT from 15 to 10 d.To calibrate the ASM 1 , two detailed sampling procedures were carried out in the aeration tank of the pilot plant to monitor the nitrogen concentrations every 10 min for three on/off aeration cycles over 8 h (Choubert et al., 2005); 2 nd series: The plant was stabilised at an average organic loading rate of 0.07 kg BOD 5 .kgMLVSS - •d -1 with a 20 d SRT.Different nitrogen volumetric loading rates were applied from 30 to 140 mgN•ℓ -1 •d -1 .The nitrification rates were estimated with the nitrogen mass balance method.

Theoretical expression of nitrification rate
To determine the prevailing parameters of nitrification rates in an activated sludge process, a theoretical approach was followed.The relation obtained was used to examine the nitrification rates database.The amount of autotrophic biomass kept in an activated sludge system results from their specific net growth (∆MX B,A / MX B,A ) during the duration "∆t".Equation (1) (Nowak et al., 1994) expresses the difference between the mass of bacteria produced by nitrification (Y A .ϕ Nnit ) and the mass lost by decay (b A .∆t) and by sludge wastage (∆t/SRT): (3) For steady-state conditions (∆MX B,A = 0), Eq. ( 3) gives the mass of nitrifiers (MX B,A [g COD]) contained in a WWTP shown in Eq. ( 4).Using the fraction (f AT ) of the total sludge contained in the biological reactors (intermittently aerated + anoxic + anaerobic zones + clarifier), and the volumetric nitrogen loading nitrified (B v (N nitrified )), the mass of nitrifiers (MX B,A ) can be converted into their concentration (X B,A [mg COD•ℓ -1 ]) in the biological reactor (Eq.( 5)). (4) (5) Considering η TKN and η BOD5 , the efficiencies of TKN and BOD 5 removal, then the fraction of nitrogen available for nitrification (λ) is defined as Eq. ( 6): (6) From Eqs. ( 5) and ( 6) the maximum nitrification rate r V, max nit in mgN nitrified •ℓ -1 •h -1 as a function of λ, f AT , SRT, µ A, max , b A , B v (N) is expressed in Eq. ( 7): (7) For a given SRT, Eq. ( 7) expresses a linear relationship between r v, max nit and B v (N).It also demonstrates that SRT influences the value of r v, max nit .Taking into account the oxygen limitation with a Monod type equation, the expression of the actual nitrification rate (r v, nit ) in the reactor becomes: (8) The operating conditions [SRT and B v (N)] of the 10 WWTPs of the database were used to calculate the corresponding r v, nit with Eq. ( 8).The following values were considered: K O,A = 0.4 mg O 2 •ℓ -1 (Henze et al., 1987), DO = 3 mg O 2 •ℓ -1 (concentration when aeration is on), f AT = 85 ± 10% estimated values with current values operating conditions like COD/BOD 5 = 2.5 ± 0.3, η TKN = 97%, η BOD5 = 99%, COD/TKN ratio of the influent (λ = 0.76 ± 0.05 if COD/TKN = 10.5 ; λ = 0.85 ± 0.05 for COD/TKN = 6), 4 ± 1 g MLSS•ℓ -1 for the sludge concentration in the aeration tank with 70 ± 5% for VSS fraction.

Results
The nitrification rate database has been analysed with both ASM 1 simulations and Eq. ( 8) (theoretical approach) to determine the more relevant value of maximum growth rate (µ A, max ) and endogenous decay rate (b A ) of nitrifiers at 10°C. Figure 2 illustrates the calibration and on-site validation strategies.
ASM 1 has been calibrated on the data of the detailed sampling carried out during the 1 st series of pilot-plant experiment.Most default parameter values proposed by Henze (1987) have been used.Only 2 parameters have been modified: the maxi- of the autotrophic bacteria.With high ammonia concentration in the effluent (Fig. 3) and with low ammonia concentration (Fig. 4), the fit between observed and simulated concentrations could be obtained for several pairs (µ A, max ; b A ) = (0.22 ; 0.02) or (0.30 ; 0.06) or (0.45 ; 0.13) or (0.60 ; 0.20) or (0.7 ; 0.25) (simulated ammonia and nitrates curves are represented for each pair).
To obtain the most likely pair for (µ A, max ; b A ), the results of the 2 nd series were studied (SRT = 20 d).The calculated values (Eq.( 8)) obtained with the 5 previous pairs (µ A, max ; b A ) were compared to the linear relation (r v, nit = 0.647 * B v (N), r 2 = 0.77) obtained with adjustments on observed nitrification rates.

Conclusions
The parameters prevailing on nitrification capacity of activated sludge plants have been investigated at 10°C from full-scale measurements and long-term pilot-scale experiments.This study has confirmed that a large number of combinations can be used for the autotrophic biomass parameters (µ A, max ; b A ) to simulate the concentration of the nitrogen species in an aeration tank.But, only one provides a correct fit with nitrification rate for a wide domain of validity of SRT and nitrogen loading rate conditions.The unique pair at 10°C is (µ A, max = 0.45•d -1 ; b A = 0.13•d -1 ).These values obtained on continuous reactors confirm those reported in the literature with batch-test procedures.
The maximum nitrification rate is demonstrated to be proportional to the nitrogen volumetric loading rate when the SRT is constant.The operating conditions such as the COD/TKN 76 ratio of the influent and the sludge retention time are also shown to determine the value of the maximum nitrification rate.With a typical domestic influent (COD/TKN = 10.5), the r v, nit at 10°C increases from 1 to 5 mgN nit •ℓ -1 •h -1 when B v (N) increases from 10 to 140, whereas for nitrogen-enriched influent (COD/TKN = 6), r v, nit reaches 9 mgN nit •ℓ -1 •h -1 .

Figure 2
Figure 2Applied strategy to determine µ A, max and b A at 10°C from both pilot-and full-scale data

Figure 5 Figure 5
Figure 3Pilot plant, 1 st series, with high nH 4 -n

Figure 6
Figure 6Observed and calculated nitrification rates as a function of the nitrogen volumetric loading rate and the influent COD/TKN ratio (for each B v (n) corresponds to a particular SrT)