Escolar Documentos
Profissional Documentos
Cultura Documentos
H I G H L I G H T S G R A P H I C A L A B S T R A C T
art ic l e i nf o a b s t r a c t
Article history: Hydrodynamics and bio-dynamics are equally important in upflow anaerobic sludge blanket (UASB)
Received 8 April 2015 reactors, but few studies have been conducted to consider both issues when developing a mathematical
Received in revised form model. Therefore a novel ADM1-based dispersive model was developed in this study to provide a better
23 June 2015
description of UASB reactor performance by simultaneously taking into account the hydrodynamics and
Accepted 11 July 2015
bio-dynamics of the reactor. Organic overloading shock experiments were carried out in a laboratory-
Available online 26 July 2015
scale UASB reactor to validate the developed model. The results of the simulation model were in good
Keywords: agreement with experimental data. Although considerable calculation time was required, the ADM1-
Anaerobic digestion model No. 1 (ADM1) based dispersive model showed obvious advantages in predicting reactor status, which may give an early
Dispersive model
warning of reactor abnormalities. The models developed here are expected to simulate UASB reactors
Hydrodynamics
more effectively and to provide useful information for their design and operation.
Bio-dynamics
Upflow anaerobic sludge blanket (UASB) & 2015 Elsevier Ltd. All rights reserved.
reactor
1. Introduction several decades (Li et al., 2013; Schmidt and Ahring, 1996). The UASB
reactor has a special advantage in maintaining high biomass con-
Due to their obvious advantages for minimizing pollutants and centrations, which are usually developed at the bottom of the reactor
recovering energy, upflow anaerobic sludge blanket (UASB) reactors in the form of aggregated granular sludge (Ahn et al., 2002; Schmidt
have become widely used for wastewater treatment in the last and Ahring, 1996). Formation of granular sludge significantly
improves the ability of UASB reactors to treat high-concentration,
n
possibly refractory organic matter from wastewater (Li et al., 2013;
Corresponding author. Fax: þ 86 551 63607907.
nn Liu et al., 2015; Martínez et al., 2013; Mirzoyan and Gross, 2013).
Corresponding author. Fax: þ 86 551 63600203.
E-mail addresses: yangmu@ustc.edu.cn (Y. Mu), rzeng@ustc.edu.cn (R.J. Zeng). The Anaerobic Digestion Model 1 (ADM1) has been widely used
1
These authors contributed equally to this work. to simulate anaerobic digestion processes since its first release by
http://dx.doi.org/10.1016/j.ces.2015.07.016
0009-2509/& 2015 Elsevier Ltd. All rights reserved.
678 Y. Chen et al. / Chemical Engineering Science 137 (2015) 677–684
∂C
¼0z¼H ð3Þ Experiments were performed in a Plexiglas UASB reactor which
∂z
was equipped with a water jacket to maintain the reactor under
These equations show the hydrodynamic transport of the mesophilic conditions (37 1C). The internal diameter was 0.7 dm,
components. When combined with the bio-kinetic part, the and the height of the working zone was 7.5 dm, with a volume of
distributions of the soluble substrate and insoluble biomass in 2.9 L. In the three-phase separator, the internal diameter and the
the ADM1-based dispersive model are given in Eqs. (4) and (5) height were 1.2 dm and 2.5 dm respectively, with a volume of
Y. Chen et al. / Chemical Engineering Science 137 (2015) 677–684 679
Table 1 The injector and detector temperatures were 250 1C and 300 1C
Experimental conditions of organic overloading shocks conducted on the UASB respectively. Produced gas volume was recorded by a gas meter,
reactor.
and the gas contents were analyzed by a gas chromatograph
Shock run R1 R2 R3 R4 R5 R6 (Lunan model SP7890, CN) equipped with a thermal conductivity
detector and a 1.5-m stainless steel column packed with a 5 Å
Baseline (g-COD/L) 3.3 3.3 3.3 5 5 5 molecular sieve. The temperatures of the injector, detector, and
Shock run concentration (g-COD/L) 4.95 9.9 19.8 15 25 30 column were kept at 170 1C, 170 1C, and 150 1C, respectively. H2
Duration (h) 24 24 24 24 24 24
was used as the carrier gas.
2.7 L. Therefore, the total volume was 5.6 L. A gas meter was 3. Results and discussion
installed at the outlet to measure volumetric gas production.
The reactor was fed at a flow rate of 0.2 L/h with solutions A 3.1. Reactor performances
and B (1:1). The composition of solution A was as follows (in mg/
L): KCl 5.2, MgCl2 6H2O 7.2, CaCl2 2H2O 1, NH4Cl 0.4, Na2SO4 0.08, At steady state, the effluent pH of the UASB reactor was
MnCl2 4H2O 0.8, CoCl2 6H2O 1.2, H3BO3 0.2, CuCl2 2H2O 1.1, maintained at 7.27 0.3, and the gas produced by the reactor was
Na2MoO4 2H2O 0.1, ZnSO4 7H2O 3.2, FeSO4 7H2O 3.2, evaluated at a rate of 0.30–0.37 L CH4/g COD with CH4 concentra-
NiCl2 6H2O 0.5, NaHCO3 3.5. Solution B was glucose, and its tion in the 53–59% range.
concentration is given in Table 1. In R1 and R2 runs, only acetate was detected at the effluent.
The UASB reactor was inoculated with granule sludge from an Other VFAs were either below the detection baseline or within the
expanded granular sludge-bed reactor treating starch wastewater ability of the organisms to degrade them. Propionate, butyrate, and
(Huayi Ltd, China). Seed sludge was inoculated at 37 1C and iso-butyric acid were detected only when the reactor endured
sparged with nitrogen gas (4 99.99%) for 20 min before use. several organic shocks. Under shock loads, the effluent propionate
Following start-up, reactor operation was switched from batch to and acetate concentrations increased instantly. In R3, R4, R5 and R6,
continuous mode at an HRT (Hydraulic Retention Time) of 28 h propionate accumulated more heavily than acetate, whereas buty-
and an organic loading concentration of 3.3 g COD/L. The SRT rate and iso-butyric acid increased only slightly (Figs. 2–4). A
(Solids Retention Time) of the reactor was kept at one month. significant increase in discharged gas was observed, but the CH4
Six runs of organic-overloading shock tests were conducted on percentage decreased in the gas phase. Bubbles were generated in
the UASB reactor and are summarized in Table 1. Both HRT and greater amounts during shock phases. A fraction of the granules
SRT remained the same during the shocks. The reactor was were driven upwards along with the ascending bubbles, making the
operated for at least four weeks to restore it to baseline conditions sludge blanket expand slightly. In the first four runs, the reactor
before the next run. The duration of all shock tests was set to 24 h. managed to recover to steady state within three days. In Runs 5 and
Samples were collected at the outlet and analyzed for concentra- 6, the reactor took more than 10 days to restore itself. High loading
tions of volatile fatty acids (VFAs). shocks in this study demonstrated that propionate accumulates
more easily than acetate and butyrate (Figs. 2 and 3). This is the case
2.4. Tracer study because propionate is much more difficult to degrade due to its
positive Gibbs free energy change for conversion (Stams, 1994).
A pulse-response study was carried out to elucidate the Others have also pointed out that a pH over 7 selects a microbial
hydrodynamics of the UASB reactor. Li þ was chosen as the tracer community that produces mainly propionate rather than acetate
because of its ease of analysis and its lack of chemical reactivity and butyrate (Angenent et al., 2002). Without instant removal of
with the substances in the reactor. A solution containing 480 mg VFAs, propionate also accumulates quickly, which subsequently
Li2SO4 (10 mg/L Li þ on average in the whole reactor) was injected inhibits bacterial activity and prevents propionate from degrading
at time t¼0 in the influent at the bottom of the reactor. Sample to acetate (Björnsson et al., 1997; Pullammanappallil et al., 2001).
(1 mL) from the top, middle, and bottom sampling ports was The propionate/acetate ratio was below 1 when suffering in the
collected every two hours and was analyzed for the Li þ concen- low organic loading shocks (such as R1, R2). However, the ratio
tration. The dead volume of the reactor was estimated using an was over 1 when suffering in the high organic loading shocks
equation provided in a previous study (Ren et al., 2008). (such as R3 to R6). The maximum ratio was 7 in R4. Li et al. (2014)
suggested that the ratio of propionate to acetate higher than
2.5. Analyses 1.4 indicated impending digester failure, it was the similar results
to our study. So, it was concluded that the ratio of propionate to
Samples were first centrifuged at 10,000 rpm for 5 min and acetate was an early warning indicator.
then filtered with a 0.45 μm micro-filter membrane before analy-
sis. Li þ was determined by inductively coupled plasma optical 3.2. Hydrodynamic characteristics of the UASB reactor
emission spectroscopy (ICP OES Optima 7300 DV) at a wavelength
of 670 nm. The ICP operational parameters were as follows: Based on experimental data from tracer studies (Ren et al.,
40 MHz generator frequency; 1.1 kW radio-frequency power; 2008; Zeng et al., 2005), the estimated dead volume could be
15 L/min argon plasma flow rate; 0.5 L/min auxiliary argon flow ignored because it occupied only 2.22% of the whole reactor. The
rate; 0.8 L/min nebulizer argon flow rate; and 1.4 mL/min sample simulation results of the dispersive model of UASB reactor hydro-
flow rate. A 2.4 mm central tube internal-diameter torch was used. dynamics are shown in Fig. 5. The simulation results matched the
Concentrations of VFAs were measured by a gas chromatograph experimental data very well, with a standard deviation of 0.26,
(Agilent 7890, CA) equipped with a flame ionization detector and a and the dispersion coefficient derived from this model was
10 m 0.53 mm HP-FFAP fused-silica capillary column after being 1.2 dm2/h. Two upflow velocities, with the smaller one in the
acidified with 3% (v/v) formic acid. The configurations of the separator part, instead of one uniform velocity in the whole
column operating temperature was first 70 1C for 3 min, then reactor might contribute to the quality of the curve fitting.
rising at a rate of 10 1C/min to 180 1C, and then holding for The multi-CSTR model, another common model for describing
4.5 min; this procedure was then repeated for the same sample. hydraulic phenomena (Ren et al., 2009, 2008), was also developed
680 Y. Chen et al. / Chemical Engineering Science 137 (2015) 677–684
Fig. 3. Effluent propionate variations of the UASB reactor in the last four runs.
Y. Chen et al. / Chemical Engineering Science 137 (2015) 677–684 681
Fig. 4. Effluent butyrate and isobutyric acid variations of the UASB reactor in the last four runs.
Hill, D., Cobb, S., Bolte, J., 1987. Using volatile fatty acid relationships to predict Peña, M.R., Mara, D.D., Avella, G.P., 2006. Dispersion and treatment performance
anaerobic digester failure. Trans. ASAE 30, 496–501. analysis of an UASB reactor under different hydraulic loading rates. Water Res.
López, I., Borzacconi, L., 2009. Modelling a full scale UASB reactor using a COD 40, 445–452.
global balance approach and state observers. Chem. Eng. J. 146, 1–5. Pullammanappallil, P.C., Chynoweth, D.P., Lyberatos, G., Svoronos, S.A., 2001. Stable
Li, J., Yu, L., Yu, D., Wang, D., Zhang, P., Ji, Z., 2014. Performance and granulation in performance of anaerobic digestion in the presence of a high concentration of
an upflow anaerobic sludge blanket (UASB) reactor treating saline sulfate propionic acid. Bioresour. Technol. 78, 165–169.
wastewater. Biodegradation 25, 127–136. Ren, T.-T., Mu, Y., Ni, B.-J., Yu, H.-Q., 2009. Hydrodynamics of upflow anaerobic
Li, L., He, Q., Wei, Y., He, Q., Peng, X., 2014. Early warning indicators for monitoring sludge blanket reactors. AIChE J. 55, 516–528.
the process failure of anaerobic digestion system of food waste. Bioresour. Ren, T.-T., Mu, Y., Yu, H.-Q., Harada, H., Li, Y.-Y., 2008. Dispersion analysis of an
Technol. 171, 491–494. acidogenic UASB reactor. Chem. Eng. J. 142, 182–189.
Liu, Y., Zhang, Y., Ni, B.-J., 2015. Zero valent iron simultaneously enhances methane Rodríguez-Gómez, R., Renman, G., Moreno, L., Liu, L., 2014. A model to describe the
production and sulfate reduction in anaerobic granular sludge reactors. Water performance of the UASB reactor. Biodegradation 25, 239–251.
Schmidt, J.E., Ahring, B.K., 1996. Granular sludge formation in upflow anaerobic
Res. 75, 292–300.
sludge blanket (UASB) reactors. Biotechnol. Bioeng. 49, 229–246.
Marchaim, U., Krause, C., 1993. Propionic to acetic acid ratios in overloaded
Stams, A.J., 1994. Metabolic interactions between anaerobic bacteria in methano-
anaerobic digestion. Bioresour. Technol. 43, 195–203.
genic environments. Antonie van Leeuwenhoek 66, 271–294.
Martínez, C., Celis, L., Cervantes, F., 2013. Immobilized humic substances as redox
Tartakovsky, B., Mu, S.J., Zeng, Y., Lou, S.J., Guiot, S.R., Wu, P., 2008. Anaerobic
mediator for the simultaneous removal of phenol and Reactive Red 2 in a UASB
digestion model no. 1-based distributed parameter model of an anaerobic
reactor. Appl. Microbiol. Biotechnol. 97, 9897–9905. reactor: II. Model validation. Bioresour. Technol. 99, 3676–3684.
Martin, A., 2000. Interpretation of residence time distribution data. Chem. Eng. Sci. Thamsiriroj, T., Nizami, A.S., Murphy, J.D., 2012. Use of modeling to aid design of a
55, 5907–5917. two-phase grass digestion system. Bioresour. Technol. 110, 379–389.
Mirzoyan, N., Gross, A., 2013. Use of UASB reactors for brackish aquaculture sludge Xiao, X., Sheng, G.-P., Mu, Y., Yu, H.-Q., 2013. A modeling approach to describe ZVI-
digestion under different conditions. Water Res. 47, 2843–2850. based anaerobic system. Water Res. 47, 6007–6013.
Mu, S.J., Zeng, Y., Tartakovsky, B., Wu, P., 2007. Simulation and control of an upflow Yu, L., Zhao, Q., Ma, J., Frear, C., Chen, S., 2012. Experimental and modeling study of
anaerobic sludge blanket (UASB) reactor using an ADM1-based distributed a two-stage pilot scale high solid anaerobic digester system. Bioresour. Technol.
parameter model. Ind. Eng. Chem. Res. 46, 1519–1526. 124, 8–17.
Mu, S.J., Zeng, Y., Wu, P., Lou, S.J., Tartakovsky, B., 2008. Anaerobic digestion model Zeng, Y., Mu, S.J., Lou, S.J., Tartakovsky, B., Guiot, S.R., Wu, P., 2005. Hydraulic modeling
no. 1-based distributed parameter model of an anaerobic reactor: I. Model and axial dispersion analysis of UASB reactor. Biochem. Eng. J. 25, 113–123.
development. Bioresour. Technol. 99, 3665–3675. Zhao, B.-H., Mu, Y., Dong, F., Ni, B.-J., Zhao, J.-B., Sheng, G.-P., Yu, H.-Q., Li, Y.-Y.,
Nordstedt, R., Thomas, M., 1985. Startup characteristics of anaerobic fixed bed Harada, H., 2010. Dynamic modeling the anaerobic reactor startup process. Ind.
reactors. Trans. ASAE 28, 1242–1247 1252. Eng. Chem. Res. 49, 7193–7200.