- Email: [email protected]
Multiscale hydrodynamic investigation to intensify the biogas production in upflow anaerobic reactors
Accepted Manuscript Multiscale hydrodynamic investigation to intensify the biogas production in upflow anaerobic reactors Jiankai Jiang, Jing Wu, Jinbai Zhang, Souhila Poncin, Huai Z. Li PII: DOI: Reference:
S0960-8524(13)01919-6 http://dx.doi.org/10.1016/j.biortech.2013.12.079 BITE 12802
To appear in:
Bioresource Technology
Received Date: Revised Date: Accepted Date:
28 October 2013 14 December 2013 19 December 2013
Please cite this article as: Jiang, J., Wu, J., Zhang, J., Poncin, S., Li, H.Z., Multiscale hydrodynamic investigation to intensify the biogas production in upflow anaerobic reactors, Bioresource Technology (2013), doi: http:// dx.doi.org/10.1016/j.biortech.2013.12.079
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Multiscale hydrodynamic investigation to intensify the biogas production in upflow anaerobic reactors Jiankai Jianga, Jing Wu b, Jinbai Zhanga, Souhila Poncina and Huai Z. Lia* a
Laboratory of Reactions and Process Engineering, Université de Lorraine, CNRS, 1, rue
Grandville, BP 20451, 54001 Nancy cedex, France b
State Key Joint Laboratory of Environment Simulation and Pollution Control, School of
Environment, Tsinghua University, Beijing 100084, P.R. China *Corresponding author: [email protected]
Abstract Hydrodynamics plays a main role for the performance of an anaerobic reactor involving three phases: wastewater, sludge granules and biogas bubbles. The present work was focused on an original approach to investigate the hydrodynamics at different scales and then to intensify the performance of such complex reactors. The experiments were carried out respectively in a 3D reactor at macroscale, a 2D reactor at mesoscale and a 1D anaerobic reactor at microscale. A Particle Image Velocimetry (PIV), a micro-PIV and a high-speed camera were employed to quantify the liquid flow fields and the relative motion between sludge granules and bubbles. Shear rates exerted on sludge granules were quantified from liquid flow fields. The optimal biogas production is obtained at mean shear rate varying from 28-48 s-1, which is controlled by two antagonistic mechanisms. The multiscale approach demonstrates pertinent mechanisms proper to each scale and allows a better understanding of such reactors.
1
Keywords: upflow anaerobic reactor, biogas production, hydrodynamics, multiscale approach, PIV and micro-PIV
INTRODUCTION Due to the fast consumption of fossil fuels coupled with an increasing demand of energy, the energy shortage appears to be one of great worldwide challenges. Plenty of attentions have been paid to develop alternate energy such as solar energy, wind energy, biofuel, etc. The anaerobic digestion is an efficient and convenient approach to generate biogas, which can be converted into heat, electricity or vehicle fuel after purification and compression. In recent years, significant progress has been obtained in anaerobic digestion through the evolution of anaerobic reactors based on granular sludge. Upflow anaerobic reactors including upflow anaerobic sludge blanket (UASB) reactor, internal circulation anaerobic (ICA) reactor and expanded granular sludge blanket (EGSB) reactor etc. as the most advanced upflow anaerobic reactors are widely used and constructed in the world. The existence of UASB units are more than 1000 operating all over the world, while the total number of ICA and EGSB installations are over 300,and increase rapidly in recent 20 years (Kassam et al., 2003; Musee & Lorenzen, 2013). The organic loading of ICA reactor is 3-5 times higher than UASB reactor because of the enhanced mass transfer due to faster superficial liquid velocity and gas velocity. The superficial liquid velocity of the former reactor is 10-20 m/h while it is less than 1 m/h in the latter. Thus the hydrodynamics plays a crucial role in the upflow anaerobic reactors.
The understanding of the hydrodynamics in the upflow anaerobic reactors is still limited since the study is focused mainly on the performance of the reactors treating various wastewater and energy recovery up to now (Seghezzo et al., 1998; Uemura & Harada, 2000; Latif et al., 2
2011; Tauseef et al., 2013). The impacts of some key operational parameters such as hydraulic and organic loadings, temperature, nutrients and pH have been discussed (Chan et al., 2009; Ghangrekar et al., 2005; Liu & Tay, 2002; Wu et al., 2009). Some studies were devoted to the sludge granulation process and the characteristics of mature granular sludge, like the formation, structure, size, rheology, metabolism of granular sludge in the anaerobic reactor (Baloch et al., 2008; Batstone & Keller, 2001; Bhunia & Ghangrekar, 2007; Hulshoff Pol et al., 2004; Liu et al., 2003; Mu et al., 2006; Pevere et al., 2006; Wu et al., 2006). From experimental data, there are only a few studies on the hydrodynamic characteristics of upflow anaerobic reactor by establishing a global model or adopting Computational Fluid Dynamics (CFD) simulation (Ren et al., 2009; Vesvikar & Al-Dahhan, 2005; Wu, 2010; Wu & Chen, 2008). The hydrodynamics in UASB could be described by the consecutive CSTR (continuous stirred tank reactor) model including either equal-sized CSTRs or increasing-size CSTRs (Costello et al., 1991a; Costello et al., 1991b; Ren et al., 2009). It seems that the hydrodynamic behavior is basically dispersion-controlled. Other kinds of reactors were also studied like EGSB, fermentative hydrogen bioreactor (Chu et al., 2011; Zheng et al., 2012). However, to our best knowledge, accurate measurements of the local hydrodynamics in upflow anaerobic reactor are scarce. Thus, it is essential to gain the hydrodynamic understanding for improving the reactor performance and the biogas production.
The upflow anaerobic reactors based on sludge granules are typical three-phase (gas-liquidsolid) reactors, with really complex hydrodynamics inside. For example, ICA reactor has two reaction stages (namely, the 1st stage and 2nd stage), settling stage, riser, downcomer and gas/liquid separator. Internal circulation is normally caused by the different densities in the riser and downcomer. Most biogas is produced in the 1 st stage, then, biogas is trapped in the first section of gas-hoods and rises through the riser section to a gas-liquid separator placed on 3
top of the reactor including liquid to be recycled. The denser liquid after being degassed goes back through the downcomer to the first stage. The biogas thus drives an internal circulation flow, which results in good mixing in the bottom section and enhances mass transfer between granules and wastewater. As a key conception parameter in ICA reactors, liquid circulation velocity and its major effects have to be studied (Hu, 2011). Biomass retention takes place in the 2nd stage.
The present study aims at studying the hydrodynamic characteristics in upflow reactors of ICA type by means of an original multiscale approach. Especially, particle image velocimetry (PIV), micro-PIV and high-speed digital camera were employed for the first time to quantify the hydrodynamics at three different scales: firstly, in a 3D macroscale reactor; secondly, the liquid flow and shear rate fields in the reaction zone and downcomer in a 2D mesoscale reactor; and finally, the detailed liquid flow and shear rate fields around single granules in a 1D microscale reactor. These advanced techniques allow quantifying various phenomena at different scales including the role of bubbles’ size on the shear rate and the optimal flow conditions for the biogas production in such upflow anaerobic reactors.
MATERIALS AND METHODS It is still impossible to study the fluid flows in a real anaerobic reactor due to total opacity induced by biofilm developing on the reactor walls as well as concentrated sludge granules. Thus tap water and nitrogen were used to simulate the wastewater (liquid) and generated biogas respectively in this study at both macroscale and mesoscale.
4
Experimental features of the 3D reactor at macroscale Experiments were conducted in a transparent 3D ICA reactor made of Plexiglas with diameter 50 mm and height 1400 mm (Fig. 1a). A transparent cylinder of diameter 15 mm was used as downcomer for liquid circulation outside. Nitrogen was introduced into the reactor through a perforated panel of 0.5 mm installed on the bottom. Demineralized water was pumped by a peristaltic pump into the reactor from the bottom. The flow rates of both water and nitrogen were monitored by flow meters.
Experimental features of the 2D reactor at mesoscale To be able to observe interactions never observed so far between bubble and granular sludge and calculate shear rate exerted on granules, a transparent 2D reactor made of Plexiglas with a 30 × 10 mm cross section and 250 mm height was employed to mimic the reaction zone of ICA reactor (Fig. 1b). Both a sintered metal distributor and a sieve distributor were equipped successively in the bottom of the reactor to inject nitrogen with various diameters of bubbles. Liquid entered into the reactor from the bottom. The liquid flow rate was 75 mL/min during the experiments and the superficial liquid velocity was 4.17×10 -3 m/s (i.e. 15 m/h), the typical value for full-scale reactors. And nitrogen flow rate was varied to change the liquid circulation intensity.
Experimental features of the 1D reactor at microscale Experiments under anaerobic conditions were conducted in a 1D transparent microreactor made of Plexiglas: width × height: 2 × 2 mm (Fig. 1c). A single granule of various sizes (equivalent diameter from 1.05 to 1.57 mm) was set up by a small grid. The synthetic substrate of Chemical Oxygen Demand (COD) of 3000 mg/L was prepared in a reservoir of 15 L for each experiment, where the composition of the synthetic substrate was same as 5
described in previous work (Zhang et al., 2012). It was designed to supply sufficient substrate for maintaining a constant value of COD at the inlet, and pumped into the microreactor. The temperature was controlled at 30ºC by a thermobath. The sludge granules used in our experiments were selected under the batch conditions with the same feeding solution used in the experiments for their proven methanogenic efficiency. The experiments ran under various liquid superficial velocities (0 - 5.21× 10-3 m.s-1) in order to obtain different hydrodynamic conditions, zero liquid velocity corresponding to stagnant batch condition.
Besides the hydrodynamic characterization, the biogas production was monitored as well in this microreactor. The size of biogas microbubbles was determined using a quantitative image analysis program in Matlab and the shape of produced microbubbles was fairly spherical due to their small size (typically, diameter < 0.36 mm). To quantify the biogas production rate, during a certain time interval (typically 15 - 20 min), the recordings of all the biogas microbubbles successively formed were performed. The volume of all the microbubbles recorded during the interval was employed to evaluate the biogas production rate at this moment. Eight such samples were consecutively recorded for each granule during 3 days, e.g. for a given liquid velocity to check the repeatability of the measured biogas production rate.
Liquid flow and shear rate fields Instantaneous velocity fields in the 3D and 2D reactor was obtained with a PIV system (Dantec Dynamics, Denmark), which was composed of two pulsed Nd-YAG Lasers (New Wave Research, USA), two CCD cameras, a control unity (FlowMap 1500) and a computer. Water was inseminated with silvered glass microspheres of average diameter 15 µm and density 1400 kg/m3 as seeding particles to trace the liquid. The sedimentation velocity of the seeding particles was around 30 µm.s-1. Compared to the liquid velocity used in this study
6
(4200 µm/s), the errors in liquid velocities measurements did not exceed 1%. The laser beams crossed a cylindrical lens firstly to produce laser sheet with strong light intensity and low thickness (2.0 × 10-3 m maximum). They were focused and superposed on one measuring window crossing the vertical symmetry axis of the dispersed phase. The camera placed perpendicular to the laser sheet took successively images at the maximum intensity of the laser pulse. Then these raw successive image pairs were analyzed by the FlowMap software (Dantec Dynamics, Denmark) to produce flow fields. The error of PIV was less than 5% via the experimental calibration of the velocity in an empty cylinder.
For the 1D microreactor, a micro-PIV system (Dantec Dynamics, Denmark) was used to measure the instantaneous liquid velocity flow field by adding the carrier fluid with calibrated Latex seeding particles (Merck, France) of diameter 0.88 µm and density 1056 kg/m3 and visualizing the flow through a × 5 objective mounted on an inverted microscope with a numerical aperture of 0.12 (Dietrich et al., 2008). The concentration of the seeding particles in the solution was prepared to ensure approximately five particles in each interrogation area. The flow was illuminated by a laser. By shadowgraph of the seeding particles, images of the flow were taken by a double image digital camera through the microscope. The duration of the laser pulse was 0.01 µs. After exposure, the image field was transferred to the storage on the digital camera, so that a second image field could be recorded by the camera. After a specified time delay, ranging from 10 to 300 µs depending on the liquid velocities in our case, a second laser pulse was used to record another image. Both images were then recorded on a computer for further analyzing.
The shear rate fields are calculated as follows from the original database of flow fields acquired by PIV or micro-PIV devices:
7
γx = γ y =
u1x − u2 x dy u1 y − u2 y
(1)
dx
γ = γx2 + γ y2 where u 1 and u2 are two arbitrary adjoining vectors in flow fields separated by a distance d; ux and u y are x axis component and y axis component of liquid velocity u; and γx , γ y are the axis projections of the shear rate γ .
Collisions between bubbles and sludge granules A high-speed digital camera (Optronis GMBH, Germany) was employed to visualize collisions between bubbles and sludge granules in the 3D and 2D reactors. In this part of experiments, 200 frames per second (fps) and an exposure time of 1/500 s were used. An evident contrast demanded by the image analysis was insured by strong LED backlight with emitting surface of 100×100 mm (Stemmer Imaging, United Kingdom).
RESULTS AND DISCUSSIONS 3D reactor at macroscale In industrial ICA reactors, the internal circulation is a key parameter to obtain the optimal hydrodynamic conditions. On one hand, flow rate of the liquid circulation, up to several times higher than that of the influent, can dilute toxic substance’s concentration thereby to improve the capacity of resistance to shock loadings. On the other hand, the alkalinity in the circulation keeps pH stable without the addition of alkali. Thus, main factors affecting the internal
8
circulation were extensively investigated by the PIV technique in a lab-scale 3D macroscale reactor.
Three phases, gas bubbles, liquid and sludge granules, contribute to the internal circulation through complex interactions in an anaerobic reactor. Since the liquid circulation was mainly driven by gas bubbles, the circulation velocity increased with the superficial gas velocity as shown in Fig. 2a. It is worth noting that the velocity in the presence of sludge granules was significantly lower than that without granules. This indicates that effects of suspended sludge cannot be ignored. In fact, the presence of granules increases the viscosity of the mixture and induces then higher energy consumption. And the circulation velocity decreased slightly with the increase of influent liquid flow rate in spite of the presence or absence of granular sludge. Fig. 2b shows clearly that the circulation velocity was inversely correlated to the influent liquid flow rate and the presence of granules increased the slope. The relative importance of the three phases was ordered as gas > solid > liquid for the liquid circulation velocity under experimental conditions.
The liquid and gas velocities are key hydrodynamic parameters for reactors at macroscale. But they are not sufficient yet to describe the complex hydrodynamic conditions because they can neither reflect interactions among three phases nor local flow parameters such as the shear on the sludge granule. Even the global pattern like dead zones isn’t directly involved in the velocities of gas and liquid. The PIV technique provides then a powerful tool to investigate the local hydrodynamic conditions in the reactors. Within the reactor, it is inevitable to meet both general and local circulations because the average liquid ascension velocity is smaller than local liquid velocity generated by bubbles. It can be expected that the superficial gas 9
velocity is dominant for the shear rate with respect to the superficial liquid velocity. The mean shear rate in industrial ICA reactors was roughly estimated about 30 to 40 s-1, and was attributed to the high superficial gas velocity (Pereboom & Vereijken, 1994).
The circulation liquid velocity in the downcomer was also measured by the PIV in the present work and resulting instantaneous liquid flow fields are shown in Fig. 3b. The flow regime is typically laminar as the velocity profile in the half-pipe section integrated from the detailed flow fields is perfectly parabolic under various conditions (Fig. 3a). Moreover, it is interesting to consider a liquid Reynolds number in the downcomer defined as follows:
Re =
ρ L ud µL
(2)
where ρ L is the liquid density, kg.m-3; u the liquid mean velocity, m.s-1; d, diameter of the downcomer, m; µ L , the liquid dynamical viscosity, Pa.s. The range of the corresponding Reynolds number gathered in Table 1 shows the maximum is 455, much less than the threshold of 2100 for the laminar flow regime in an empty tube.
Due to these detailed flow fields and velocity profiles obtained by the PIV, it is easy to compute local shear rates by deriving flow fields as well as to estimate a mean shear rate in the downcomer. Certainly, the shear rate is not homogeneous across the pipe section due to the velocity profile. The local shear rate in the center of the downcomer was zero due to the symmetry of the velocity profile while the maximum located close to the walls. As expected, a linear relationship between the mean shear rate and the mean liquid circulation velocity does exist: the shear is to some extent intensified by the circulation.
10
2D reactor at mesoscale
Similar experiments were carried out in the 2D reactor at mesoscale with the PIV technique. Fig. 4a shows that mean shear rate in the main section increases continuously with the superficial gas velocity. Implicitly, this augmentation could be attributed to the size of bubbles generated: higher liquid velocities as well as mean shear rate are induced by more and bigger bubbles.
In order to demonstrate the effect of bubbles’ size on the mean shear rate in the liquid, two different gas distributors were employed to generate different sizes of bubbles with the same superficial gas velocity: a sintered distributor and a sieve distributor. By means of the PIV measurements, Fig. 4b illustrates the computed results with bubbles of various sizes generated by both the gas distributors. Clearly, there is again a close relationship between the mean shear rate in liquid and mean bubble diameter whatever the kind of distributors and superficial gas velocity. Due to the low supercial liquid velocity in upflow reactors, these results confirm that the relative motion between gas bubbles and liquid is the dominant parameter affecting the shear rates exerted on sludge granules. In addition, smaller bubbles whose diameter was less than 1 mm display few effects on the velocity field or the shear rate. Four bubbles of diameter 0.82 mm, 0.85 mm, 0.86 mm, 0.88 mm respectively in mid of the reactor have in fact no significant impact on the shear rate which is almost below to 5 s-1 everywhere. In contrary, bigger bubbles having a diameter more than 2 mm affect substantially local liquid flow fields and the shear rates due to intense buoyancy. It is worth noting that the complex coupling between a local liquid flow field around a bubble and the wake left by preceding ones induces additional difficulties for detailed quantification. 11
The PIV technique can give detailed liquid flow fields to quantify the shear rate exerted by the liquid on sludge granules, even in the simultaneous presence of three phases in the 2D reactor at mesoscale. The derivation of the flow field leads to the shear rate field around sludge granules whose position is traced by cross symbols in Fig. 4c. The maximum shear rate around the sludge granules is between 10 and 20 s-1, higher than the mean shear rate ranging from 5 to 10 s-1. In a real reactor, a higher shear rate at the surface of a sludge granule is favorable for a more efficient mass transfer between sludge granules and liquid substrates. It is worth mentioning that within the 2D reactor, the liquid height is still too limited to allow consecutive coalescences between bubbles. In fact, the range of shear rates corresponds to bubbles whose size is much smaller than that encountered in industrial reactors with substantial height.
According to these observations, the following mechanisms could be proposed for what happens in industrial upflow reactors. Small biogas bubbles exist only in the bottom of reactors just after the nucleation, growth and detachment from sludge granules. The shear rate at the bottom of reactors is relatively low due to small sizes of biogas bubbles. During the rise of small biogas bubbles, the coalescence takes place to produce increasing sizes of bubbles so that shear rate fields are intensified by larger bubbles after consecutive coalescences. It is then interesting to gain some insight into the coalescence phenomena, in particular, to visualize and locate coalescing zones where main coalescences occur in the reactor. Unlike the 3D reactor at macroscale, the 2D reactor at mesoscale offers the possibility to visualize complex interactions between sludge granules and bubbles, especially coalescence between bubbles with the help of the high-speed digital camera. Some interesting scenarios were also observed
12
related to interactions between sludge granules and bubbles and coalescence between bubbles. The spatially distributed sludge granules formed obstacles to the free rise of bubbles. The modification of the rise trajectory of a bubble could produce a coalescence with another one in the passage. After the coalescence, the resulting bigger bubble displays higher buoyancy to induce a more intense liquid flow fields around the neighboring sludge granule. In an industrial installation, it can be expected that the presence of numerous granules and biogas bubbles can enhance these phenomena of interactions and coalescences.
1D reactor at microscale
Under the optimal biogas production, the liquid velocity field measured by the micro-PIV around a granule of equivalent diameter 1.36 mm is shown in Fig. 5a as an example. The shear rate field was then deduced from the liquid flow field. All shear rates around the granule were averaged to get a mean value. Besides the classical relationship between the liquid flow and the mass transfer, the sludge granule can undergo sharp deformation as a soft matter due to the shear stresses exerted by the liquid to inhibit the mass transfer. The mean shear rate exerted on the granule is then the most suitable parameter describing reliably the local hydrodynamic conditions under which the microbial population produces the biogas. This can then help to better understand the hydrodynamic effects on the biogas production on sludge granules rather than the liquid Reynolds number which is a global parameter. The results presented in Fig. 5b show that the optimal mean shear rates (illustrated by red circle) vary from 28 - 48 s-1 respectively for the granules with an equivalent diameter ranging from 1.05 1.57 mm in order to achieve a maximum biogas production rate.
13
These original results reveal that there are two antagonist mechanisms to control the biogas production rate, i.e. the enhanced mass transfer and the compaction of granules caused by the high shear and they result in the existence of the optimal shear rate. It is worth noting that the mean shear rates obtained in this study are quantitatively consistent with the optimal values estimated in industrial ICA reactors as mentioned before. The biogas production is straightforwardly related to the shear rate exerted on the granules of different sizes. These data could provide new insight into the design and optimization of hydrodynamic conditions in various anaerobic reactors. In the practice, the strategy of a stepped increase of shear rate could be envisaged to find out the optimal range for both the wastewater treatment and biogas production.
CONCLUSIONS
The multiphase hydrodynamics was investigated in the upflow anaerobic reactors at different scales. A relationship between liquid circulation in downcomer and superficial gas velocity in central zone was established in the 3D reactor at macroscale. In the 2D reactor at mesoscale, hydrodynamics resulting from complex interactions between three phases was investigated and the relationship shear rates/sludge granules was related to the size of gas bubbles. In the 1D microreactor, a straightforward relationship between the shear rates exerted on the granule and the optimal biogas production is found and the optimal mean shear rates vary from 28 48 s-1.
14
ACKNOWLEDGEMENTS
The financial assistance provided by China Scholarship Council and the Agence Nationale de la Recherche (ANR PROMET) to the French group, and by the Natural Science Foundation of China (51061130555 and 50978147) to the Chinese group is gratefully acknowledged.
REFERENCES
1. Baloch, M.I., Akunna, J.C., Kierans, M., Collier, P.J. 2008. Structural analysis of anaerobic granules in a phase separated reactor by electron microscopy. Bioresource Technology, 99(5), 922-929.
2. Batstone, D.J., Keller, J. 2001. Variation of bulk properties of anaerobic granules with wastewater type. Water Research, 35 (7), 1723-1729. 3. Bhunia, P., Ghangrekar, M.M. 2007. Required minimum granule size in UASB reactor and characteristics variation with size. Bioresource Technology, 98 (5), 994-999. 4. Chan, Y.J., Chong, M.F., Law, C.L., Hassell, D.G. 2009. A review on anaerobic–aerobic treatment of industrial and municipal wastewater. Chemical Engineering Journal, 155 (1-2), 1-18.
5. Chu, C.Y., Wu, S.Y., Wu, Y.C., Lin, C.Y. 2011. Hydrodynamic behaviors in fermentative hydrogen bioreactors by pressure fluctuation analysis. Bioresource Technology, 102 (18), 8669-8675.
6. Costello, D.J., Greenfield, P.F., Lee, P.L. 1991a. Dynamic modeling of a single-stage highrate anaerobic reactor .1. Model derivation. Water Research, 25 (7), 847-858. 7. Costello, D.J., Greenfield, P.F., Lee, P.L. 1991b. Dynamic modeling of a single-stage highrate anaerobic reactor .2. Model verification. Water Research, 25(7), 859-871. 15
8. Dietrich, N., Poncin, S., Midoux, N., Li, H.Z. 2008. Bubble Formation Dynamics in Various Flow-Focusing Microdevices. Langmuir, 24(24), 13904-13911. 9. Ghangrekar, M.M., Asolekar, S.R., Joshi, S.G. 2005. Characteristics of sludge developed under different loading conditions during UASB reactor start-up and granulation. Water Research, 39(6), 1123-1133.
10. Hu, J.C. 2011. A method of design and calculation for internal circulation (IC) anaerobic reactor (in Chinese). China Biogas, 29, 3-6. 11. Hulshoff Pol, L.W., de Castro Lopes, S.I., Lettinga, G., Lens, P.N.L. 2004. Anaerobic sludge granulation. Water Research, 38 (6), 1376-1389. 12. Kassam, Z.A., Yerushalmi, L., Guiot, S.R. 2003. A market study on the anaerobic wastewater treatment systems. Water Air and Soil Pollution, 143(1-4), 179-192. 13. Latif, M.A., Ghufran, R., Wahid, Z.A., Ahmad, A. 2011. Integrated application of upflow anaerobic sludge blanket reactor for the treatment of wastewaters. Water Research, 45(16), 4683-4699. 14. Liu, Y., Tay, J.-H. 2002. The essential role of hydrodynamic shear force in the formation of biofilm and granular sludge. Water Research, 36 (7), 1653-1665. 15. Liu, Y., Xu, H.-L., Yang, S.-F., Tay, J.-H. 2003. Mechanisms and models for anaerobic granulation in upflow anaerobic sludge blanket reactor. Water Research, 37 (3), 661673. 16. Mu, Y., Yu, H.-Q., Wang, G. 2006. Permeabilities of anaerobic CH4-producing granules. Water Research, 40(9), 1811-1815.
17. Musee, N., Lorenzen, L. 2013. Market dynamics as a driver towards the evolution of research needs: the case of up-flow anaerobic sludge blanket seeding granules. Water SA, 39(1),131-142.
16
18. Pereboom, J.H.F., Vereijken, T. 1994. Methanogenic granule development in full-scale internal circulation reactors. Water Science and Technology, 30(8), 9-21. 19. Pevere, A., Guibaud, G., van Hullebusch, E., Lens, P., Baudu, M. 2006. Viscosity evolution of anaerobic granular sludge. Biochemical Engineering Journal, 27(3), 315322. 20. Ren, T.-T., Mu, Y., Ni, B.-J., Yu, H.-Q. 2009. Hydrodynamics of upflow anaerobic sludge blanket reactors. AIChE Journal, 55(2), 516-528. 21. Seghezzo, L., Zeeman, G., van Lier, J.B., Hamelers, H.V.M., Lettinga, G. 1998. A review: The anaerobic treatment of sewage in UASB and EGSB reactors. Bioresource Technology, 65 (3), 175-190.
22. Tauseef, S.M., Abbasi, T., Abbasi, S.A. 2013. Energy recovery from wastewaters with high-rate anaerobic digesters. Renewable and Sustainable Energy Reviews, 19, 704741.23. Uemura, S., Harada, H. 2000. Treatment of sewage by a UASB reactor under moderate to low temperature conditions. Bioresource Technology, 72(3), 275-282. 24. Vesvikar, M.S., Al-Dahhan, M. 2005. Flow pattern visualization in a mimic anaerobic digester using CFD. Biotechnology and Bioengineering, 89(6), 719-732. 25. Wu, B. 2010. CFD simulation of gas and non-Newtonian fluid two-phase flow in anaerobic digesters. Water Research, 44(13), 3861-3874. 26. Wu, B., Chen, S. 2008. CFD simulation of non-Newtonian fluid flow in anaerobic digesters. Biotechnology and Bioengineering, 99(3), 700-711. 27. Wu, J., Lu, Z.Y., Hu, J.C., Feng, L., Huang, J.D., Gu, X.S. 2006. Disruption of granules by hydrodynamic force in internal circulation anaerobic reactor. Water Science and Technology, 54 (9), 9-16.
17
28. Wu, J., Zhou, H.M., Li, H.Z., Zhang, P.C., Jiang, J. 2009. Impacts of hydrodynamic shear force on nucleation of flocculent sludge in anaerobic reactor. Water Research, 43(12), 3029-3036. 29. Zhang, J.B., Wu, J., Poncin, S., Hamelin, M., Li, H.Z. 2012. Microscale Investigation of Anaerobic
Biogas
Production
under
Various
Hydrodynamic
Conditions.
Environmental Science & Technology, 46(16), 8698-8704.
30. Zheng, M.X., Wang, K.J., Zuo, J.E., Yan, Z., Fang, H., Yu, J.W. 2012. Flow pattern analysis of a full-scale expanded granular sludge bed-type reactor under different organic loading rates. Bioresource Technology, 107, 33-40.
Caption of figures
Figure 1. Schematic presentation of (a) 3D macroreactor; (b) 2D mesoreactor and (c) 1D microreactor.
Figure 2. Circulation velocity in the 3D macroreactor. (a) in function of superficial gas velocity with different influent liquid flow rates in the presence of granular sludge (10 g SS/L), (b) in function of influent liquid flow rate with different superficial gas velocities without or with granular sludge (10 g SS/L).
Figure 3. (a) Liquid velocity profile of in a half-pipe section and (b) instantaneous liquid flow fields in the downcomer of the 3D macroreactor.
Figure 4. Variation of the mean shear rate in the main section of the 2D reactor at mesoscale. (a) in function of the superficial gas velocity, (b) in function of the mean bubble diameter 18
under a gas velocity of 9.61×10-4 m/s with a sintered distributor and a sieve distributor respectively. (c) Instantaneous shear rate field with cross symbols representing sludge granules’ position at superficial gas velocity of 2.99×10-4 m·s-1 in the 2D reactor.
Figure 5. (a) Flow fields around a granule of 1.36 mm producing biogas obtained by the micro-PIV. (b) Relationship between the rate of biogas production and the mean shear rate exerted by the liquid on the granules of different diameter.
19
(a)
(c)
(b)
Figure 1
20
Figure 2
21
(b)
(a)
Figure 3
22
Figure 4
23
(a)
(b)
Figure 5
24
Table 1. Experimental values of Reynolds number under different superficial gas and liquid velocities
influent liquid
influent liquid
influent liquid
superficial gas
velocity 0 m/s with
velocity 2.21×10-3
velocity 4.24×10-3
velocity (m/s)
granular sludge 10 g
m/s with granular
m/s with granular
SS/L
sludge 10 g SS/L
sludge 10 g SS/L
3.80×10-5
204
195
140
6.69×10-5
286
193
139
1.56×10-4
301
295
265
1.94×10-4
395
312
262
2.31×10-4
455
342
350
25
Mouvement and size of bubble Liquid flow field Shear rate field Circulation velocity
ro s
Hydrodynamics in upflow anaerobic reactors
26
PIV Comprehension and improvement
al e
M ac
c os icr
Mesoscale
M
ca le
High-speed digital camera
Micro-PIV Microscope
Highlights • Hydrodynamics plays a major role in improving the performance of upflow reactors. • Gas > solid > liquid can be ordered as for the circulation velocity. • Motion between bubbles and liquid is the key parameter affecting the shear rates. • Shear rate exerted on a sludge granule displays an optimum biogas production rate.
27
S0960-8524(13)01919-6 http://dx.doi.org/10.1016/j.biortech.2013.12.079 BITE 12802
To appear in:
Bioresource Technology
Received Date: Revised Date: Accepted Date:
28 October 2013 14 December 2013 19 December 2013
Please cite this article as: Jiang, J., Wu, J., Zhang, J., Poncin, S., Li, H.Z., Multiscale hydrodynamic investigation to intensify the biogas production in upflow anaerobic reactors, Bioresource Technology (2013), doi: http:// dx.doi.org/10.1016/j.biortech.2013.12.079
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Multiscale hydrodynamic investigation to intensify the biogas production in upflow anaerobic reactors Jiankai Jianga, Jing Wu b, Jinbai Zhanga, Souhila Poncina and Huai Z. Lia* a
Laboratory of Reactions and Process Engineering, Université de Lorraine, CNRS, 1, rue
Grandville, BP 20451, 54001 Nancy cedex, France b
State Key Joint Laboratory of Environment Simulation and Pollution Control, School of
Environment, Tsinghua University, Beijing 100084, P.R. China *Corresponding author: [email protected]
Abstract Hydrodynamics plays a main role for the performance of an anaerobic reactor involving three phases: wastewater, sludge granules and biogas bubbles. The present work was focused on an original approach to investigate the hydrodynamics at different scales and then to intensify the performance of such complex reactors. The experiments were carried out respectively in a 3D reactor at macroscale, a 2D reactor at mesoscale and a 1D anaerobic reactor at microscale. A Particle Image Velocimetry (PIV), a micro-PIV and a high-speed camera were employed to quantify the liquid flow fields and the relative motion between sludge granules and bubbles. Shear rates exerted on sludge granules were quantified from liquid flow fields. The optimal biogas production is obtained at mean shear rate varying from 28-48 s-1, which is controlled by two antagonistic mechanisms. The multiscale approach demonstrates pertinent mechanisms proper to each scale and allows a better understanding of such reactors.
1
Keywords: upflow anaerobic reactor, biogas production, hydrodynamics, multiscale approach, PIV and micro-PIV
INTRODUCTION Due to the fast consumption of fossil fuels coupled with an increasing demand of energy, the energy shortage appears to be one of great worldwide challenges. Plenty of attentions have been paid to develop alternate energy such as solar energy, wind energy, biofuel, etc. The anaerobic digestion is an efficient and convenient approach to generate biogas, which can be converted into heat, electricity or vehicle fuel after purification and compression. In recent years, significant progress has been obtained in anaerobic digestion through the evolution of anaerobic reactors based on granular sludge. Upflow anaerobic reactors including upflow anaerobic sludge blanket (UASB) reactor, internal circulation anaerobic (ICA) reactor and expanded granular sludge blanket (EGSB) reactor etc. as the most advanced upflow anaerobic reactors are widely used and constructed in the world. The existence of UASB units are more than 1000 operating all over the world, while the total number of ICA and EGSB installations are over 300,and increase rapidly in recent 20 years (Kassam et al., 2003; Musee & Lorenzen, 2013). The organic loading of ICA reactor is 3-5 times higher than UASB reactor because of the enhanced mass transfer due to faster superficial liquid velocity and gas velocity. The superficial liquid velocity of the former reactor is 10-20 m/h while it is less than 1 m/h in the latter. Thus the hydrodynamics plays a crucial role in the upflow anaerobic reactors.
The understanding of the hydrodynamics in the upflow anaerobic reactors is still limited since the study is focused mainly on the performance of the reactors treating various wastewater and energy recovery up to now (Seghezzo et al., 1998; Uemura & Harada, 2000; Latif et al., 2
2011; Tauseef et al., 2013). The impacts of some key operational parameters such as hydraulic and organic loadings, temperature, nutrients and pH have been discussed (Chan et al., 2009; Ghangrekar et al., 2005; Liu & Tay, 2002; Wu et al., 2009). Some studies were devoted to the sludge granulation process and the characteristics of mature granular sludge, like the formation, structure, size, rheology, metabolism of granular sludge in the anaerobic reactor (Baloch et al., 2008; Batstone & Keller, 2001; Bhunia & Ghangrekar, 2007; Hulshoff Pol et al., 2004; Liu et al., 2003; Mu et al., 2006; Pevere et al., 2006; Wu et al., 2006). From experimental data, there are only a few studies on the hydrodynamic characteristics of upflow anaerobic reactor by establishing a global model or adopting Computational Fluid Dynamics (CFD) simulation (Ren et al., 2009; Vesvikar & Al-Dahhan, 2005; Wu, 2010; Wu & Chen, 2008). The hydrodynamics in UASB could be described by the consecutive CSTR (continuous stirred tank reactor) model including either equal-sized CSTRs or increasing-size CSTRs (Costello et al., 1991a; Costello et al., 1991b; Ren et al., 2009). It seems that the hydrodynamic behavior is basically dispersion-controlled. Other kinds of reactors were also studied like EGSB, fermentative hydrogen bioreactor (Chu et al., 2011; Zheng et al., 2012). However, to our best knowledge, accurate measurements of the local hydrodynamics in upflow anaerobic reactor are scarce. Thus, it is essential to gain the hydrodynamic understanding for improving the reactor performance and the biogas production.
The upflow anaerobic reactors based on sludge granules are typical three-phase (gas-liquidsolid) reactors, with really complex hydrodynamics inside. For example, ICA reactor has two reaction stages (namely, the 1st stage and 2nd stage), settling stage, riser, downcomer and gas/liquid separator. Internal circulation is normally caused by the different densities in the riser and downcomer. Most biogas is produced in the 1 st stage, then, biogas is trapped in the first section of gas-hoods and rises through the riser section to a gas-liquid separator placed on 3
top of the reactor including liquid to be recycled. The denser liquid after being degassed goes back through the downcomer to the first stage. The biogas thus drives an internal circulation flow, which results in good mixing in the bottom section and enhances mass transfer between granules and wastewater. As a key conception parameter in ICA reactors, liquid circulation velocity and its major effects have to be studied (Hu, 2011). Biomass retention takes place in the 2nd stage.
The present study aims at studying the hydrodynamic characteristics in upflow reactors of ICA type by means of an original multiscale approach. Especially, particle image velocimetry (PIV), micro-PIV and high-speed digital camera were employed for the first time to quantify the hydrodynamics at three different scales: firstly, in a 3D macroscale reactor; secondly, the liquid flow and shear rate fields in the reaction zone and downcomer in a 2D mesoscale reactor; and finally, the detailed liquid flow and shear rate fields around single granules in a 1D microscale reactor. These advanced techniques allow quantifying various phenomena at different scales including the role of bubbles’ size on the shear rate and the optimal flow conditions for the biogas production in such upflow anaerobic reactors.
MATERIALS AND METHODS It is still impossible to study the fluid flows in a real anaerobic reactor due to total opacity induced by biofilm developing on the reactor walls as well as concentrated sludge granules. Thus tap water and nitrogen were used to simulate the wastewater (liquid) and generated biogas respectively in this study at both macroscale and mesoscale.
4
Experimental features of the 3D reactor at macroscale Experiments were conducted in a transparent 3D ICA reactor made of Plexiglas with diameter 50 mm and height 1400 mm (Fig. 1a). A transparent cylinder of diameter 15 mm was used as downcomer for liquid circulation outside. Nitrogen was introduced into the reactor through a perforated panel of 0.5 mm installed on the bottom. Demineralized water was pumped by a peristaltic pump into the reactor from the bottom. The flow rates of both water and nitrogen were monitored by flow meters.
Experimental features of the 2D reactor at mesoscale To be able to observe interactions never observed so far between bubble and granular sludge and calculate shear rate exerted on granules, a transparent 2D reactor made of Plexiglas with a 30 × 10 mm cross section and 250 mm height was employed to mimic the reaction zone of ICA reactor (Fig. 1b). Both a sintered metal distributor and a sieve distributor were equipped successively in the bottom of the reactor to inject nitrogen with various diameters of bubbles. Liquid entered into the reactor from the bottom. The liquid flow rate was 75 mL/min during the experiments and the superficial liquid velocity was 4.17×10 -3 m/s (i.e. 15 m/h), the typical value for full-scale reactors. And nitrogen flow rate was varied to change the liquid circulation intensity.
Experimental features of the 1D reactor at microscale Experiments under anaerobic conditions were conducted in a 1D transparent microreactor made of Plexiglas: width × height: 2 × 2 mm (Fig. 1c). A single granule of various sizes (equivalent diameter from 1.05 to 1.57 mm) was set up by a small grid. The synthetic substrate of Chemical Oxygen Demand (COD) of 3000 mg/L was prepared in a reservoir of 15 L for each experiment, where the composition of the synthetic substrate was same as 5
described in previous work (Zhang et al., 2012). It was designed to supply sufficient substrate for maintaining a constant value of COD at the inlet, and pumped into the microreactor. The temperature was controlled at 30ºC by a thermobath. The sludge granules used in our experiments were selected under the batch conditions with the same feeding solution used in the experiments for their proven methanogenic efficiency. The experiments ran under various liquid superficial velocities (0 - 5.21× 10-3 m.s-1) in order to obtain different hydrodynamic conditions, zero liquid velocity corresponding to stagnant batch condition.
Besides the hydrodynamic characterization, the biogas production was monitored as well in this microreactor. The size of biogas microbubbles was determined using a quantitative image analysis program in Matlab and the shape of produced microbubbles was fairly spherical due to their small size (typically, diameter < 0.36 mm). To quantify the biogas production rate, during a certain time interval (typically 15 - 20 min), the recordings of all the biogas microbubbles successively formed were performed. The volume of all the microbubbles recorded during the interval was employed to evaluate the biogas production rate at this moment. Eight such samples were consecutively recorded for each granule during 3 days, e.g. for a given liquid velocity to check the repeatability of the measured biogas production rate.
Liquid flow and shear rate fields Instantaneous velocity fields in the 3D and 2D reactor was obtained with a PIV system (Dantec Dynamics, Denmark), which was composed of two pulsed Nd-YAG Lasers (New Wave Research, USA), two CCD cameras, a control unity (FlowMap 1500) and a computer. Water was inseminated with silvered glass microspheres of average diameter 15 µm and density 1400 kg/m3 as seeding particles to trace the liquid. The sedimentation velocity of the seeding particles was around 30 µm.s-1. Compared to the liquid velocity used in this study
6
(4200 µm/s), the errors in liquid velocities measurements did not exceed 1%. The laser beams crossed a cylindrical lens firstly to produce laser sheet with strong light intensity and low thickness (2.0 × 10-3 m maximum). They were focused and superposed on one measuring window crossing the vertical symmetry axis of the dispersed phase. The camera placed perpendicular to the laser sheet took successively images at the maximum intensity of the laser pulse. Then these raw successive image pairs were analyzed by the FlowMap software (Dantec Dynamics, Denmark) to produce flow fields. The error of PIV was less than 5% via the experimental calibration of the velocity in an empty cylinder.
For the 1D microreactor, a micro-PIV system (Dantec Dynamics, Denmark) was used to measure the instantaneous liquid velocity flow field by adding the carrier fluid with calibrated Latex seeding particles (Merck, France) of diameter 0.88 µm and density 1056 kg/m3 and visualizing the flow through a × 5 objective mounted on an inverted microscope with a numerical aperture of 0.12 (Dietrich et al., 2008). The concentration of the seeding particles in the solution was prepared to ensure approximately five particles in each interrogation area. The flow was illuminated by a laser. By shadowgraph of the seeding particles, images of the flow were taken by a double image digital camera through the microscope. The duration of the laser pulse was 0.01 µs. After exposure, the image field was transferred to the storage on the digital camera, so that a second image field could be recorded by the camera. After a specified time delay, ranging from 10 to 300 µs depending on the liquid velocities in our case, a second laser pulse was used to record another image. Both images were then recorded on a computer for further analyzing.
The shear rate fields are calculated as follows from the original database of flow fields acquired by PIV or micro-PIV devices:
7
γx = γ y =
u1x − u2 x dy u1 y − u2 y
(1)
dx
γ = γx2 + γ y2 where u 1 and u2 are two arbitrary adjoining vectors in flow fields separated by a distance d; ux and u y are x axis component and y axis component of liquid velocity u; and γx , γ y are the axis projections of the shear rate γ .
Collisions between bubbles and sludge granules A high-speed digital camera (Optronis GMBH, Germany) was employed to visualize collisions between bubbles and sludge granules in the 3D and 2D reactors. In this part of experiments, 200 frames per second (fps) and an exposure time of 1/500 s were used. An evident contrast demanded by the image analysis was insured by strong LED backlight with emitting surface of 100×100 mm (Stemmer Imaging, United Kingdom).
RESULTS AND DISCUSSIONS 3D reactor at macroscale In industrial ICA reactors, the internal circulation is a key parameter to obtain the optimal hydrodynamic conditions. On one hand, flow rate of the liquid circulation, up to several times higher than that of the influent, can dilute toxic substance’s concentration thereby to improve the capacity of resistance to shock loadings. On the other hand, the alkalinity in the circulation keeps pH stable without the addition of alkali. Thus, main factors affecting the internal
8
circulation were extensively investigated by the PIV technique in a lab-scale 3D macroscale reactor.
Three phases, gas bubbles, liquid and sludge granules, contribute to the internal circulation through complex interactions in an anaerobic reactor. Since the liquid circulation was mainly driven by gas bubbles, the circulation velocity increased with the superficial gas velocity as shown in Fig. 2a. It is worth noting that the velocity in the presence of sludge granules was significantly lower than that without granules. This indicates that effects of suspended sludge cannot be ignored. In fact, the presence of granules increases the viscosity of the mixture and induces then higher energy consumption. And the circulation velocity decreased slightly with the increase of influent liquid flow rate in spite of the presence or absence of granular sludge. Fig. 2b shows clearly that the circulation velocity was inversely correlated to the influent liquid flow rate and the presence of granules increased the slope. The relative importance of the three phases was ordered as gas > solid > liquid for the liquid circulation velocity under experimental conditions.
The liquid and gas velocities are key hydrodynamic parameters for reactors at macroscale. But they are not sufficient yet to describe the complex hydrodynamic conditions because they can neither reflect interactions among three phases nor local flow parameters such as the shear on the sludge granule. Even the global pattern like dead zones isn’t directly involved in the velocities of gas and liquid. The PIV technique provides then a powerful tool to investigate the local hydrodynamic conditions in the reactors. Within the reactor, it is inevitable to meet both general and local circulations because the average liquid ascension velocity is smaller than local liquid velocity generated by bubbles. It can be expected that the superficial gas 9
velocity is dominant for the shear rate with respect to the superficial liquid velocity. The mean shear rate in industrial ICA reactors was roughly estimated about 30 to 40 s-1, and was attributed to the high superficial gas velocity (Pereboom & Vereijken, 1994).
The circulation liquid velocity in the downcomer was also measured by the PIV in the present work and resulting instantaneous liquid flow fields are shown in Fig. 3b. The flow regime is typically laminar as the velocity profile in the half-pipe section integrated from the detailed flow fields is perfectly parabolic under various conditions (Fig. 3a). Moreover, it is interesting to consider a liquid Reynolds number in the downcomer defined as follows:
Re =
ρ L ud µL
(2)
where ρ L is the liquid density, kg.m-3; u the liquid mean velocity, m.s-1; d, diameter of the downcomer, m; µ L , the liquid dynamical viscosity, Pa.s. The range of the corresponding Reynolds number gathered in Table 1 shows the maximum is 455, much less than the threshold of 2100 for the laminar flow regime in an empty tube.
Due to these detailed flow fields and velocity profiles obtained by the PIV, it is easy to compute local shear rates by deriving flow fields as well as to estimate a mean shear rate in the downcomer. Certainly, the shear rate is not homogeneous across the pipe section due to the velocity profile. The local shear rate in the center of the downcomer was zero due to the symmetry of the velocity profile while the maximum located close to the walls. As expected, a linear relationship between the mean shear rate and the mean liquid circulation velocity does exist: the shear is to some extent intensified by the circulation.
10
2D reactor at mesoscale
Similar experiments were carried out in the 2D reactor at mesoscale with the PIV technique. Fig. 4a shows that mean shear rate in the main section increases continuously with the superficial gas velocity. Implicitly, this augmentation could be attributed to the size of bubbles generated: higher liquid velocities as well as mean shear rate are induced by more and bigger bubbles.
In order to demonstrate the effect of bubbles’ size on the mean shear rate in the liquid, two different gas distributors were employed to generate different sizes of bubbles with the same superficial gas velocity: a sintered distributor and a sieve distributor. By means of the PIV measurements, Fig. 4b illustrates the computed results with bubbles of various sizes generated by both the gas distributors. Clearly, there is again a close relationship between the mean shear rate in liquid and mean bubble diameter whatever the kind of distributors and superficial gas velocity. Due to the low supercial liquid velocity in upflow reactors, these results confirm that the relative motion between gas bubbles and liquid is the dominant parameter affecting the shear rates exerted on sludge granules. In addition, smaller bubbles whose diameter was less than 1 mm display few effects on the velocity field or the shear rate. Four bubbles of diameter 0.82 mm, 0.85 mm, 0.86 mm, 0.88 mm respectively in mid of the reactor have in fact no significant impact on the shear rate which is almost below to 5 s-1 everywhere. In contrary, bigger bubbles having a diameter more than 2 mm affect substantially local liquid flow fields and the shear rates due to intense buoyancy. It is worth noting that the complex coupling between a local liquid flow field around a bubble and the wake left by preceding ones induces additional difficulties for detailed quantification. 11
The PIV technique can give detailed liquid flow fields to quantify the shear rate exerted by the liquid on sludge granules, even in the simultaneous presence of three phases in the 2D reactor at mesoscale. The derivation of the flow field leads to the shear rate field around sludge granules whose position is traced by cross symbols in Fig. 4c. The maximum shear rate around the sludge granules is between 10 and 20 s-1, higher than the mean shear rate ranging from 5 to 10 s-1. In a real reactor, a higher shear rate at the surface of a sludge granule is favorable for a more efficient mass transfer between sludge granules and liquid substrates. It is worth mentioning that within the 2D reactor, the liquid height is still too limited to allow consecutive coalescences between bubbles. In fact, the range of shear rates corresponds to bubbles whose size is much smaller than that encountered in industrial reactors with substantial height.
According to these observations, the following mechanisms could be proposed for what happens in industrial upflow reactors. Small biogas bubbles exist only in the bottom of reactors just after the nucleation, growth and detachment from sludge granules. The shear rate at the bottom of reactors is relatively low due to small sizes of biogas bubbles. During the rise of small biogas bubbles, the coalescence takes place to produce increasing sizes of bubbles so that shear rate fields are intensified by larger bubbles after consecutive coalescences. It is then interesting to gain some insight into the coalescence phenomena, in particular, to visualize and locate coalescing zones where main coalescences occur in the reactor. Unlike the 3D reactor at macroscale, the 2D reactor at mesoscale offers the possibility to visualize complex interactions between sludge granules and bubbles, especially coalescence between bubbles with the help of the high-speed digital camera. Some interesting scenarios were also observed
12
related to interactions between sludge granules and bubbles and coalescence between bubbles. The spatially distributed sludge granules formed obstacles to the free rise of bubbles. The modification of the rise trajectory of a bubble could produce a coalescence with another one in the passage. After the coalescence, the resulting bigger bubble displays higher buoyancy to induce a more intense liquid flow fields around the neighboring sludge granule. In an industrial installation, it can be expected that the presence of numerous granules and biogas bubbles can enhance these phenomena of interactions and coalescences.
1D reactor at microscale
Under the optimal biogas production, the liquid velocity field measured by the micro-PIV around a granule of equivalent diameter 1.36 mm is shown in Fig. 5a as an example. The shear rate field was then deduced from the liquid flow field. All shear rates around the granule were averaged to get a mean value. Besides the classical relationship between the liquid flow and the mass transfer, the sludge granule can undergo sharp deformation as a soft matter due to the shear stresses exerted by the liquid to inhibit the mass transfer. The mean shear rate exerted on the granule is then the most suitable parameter describing reliably the local hydrodynamic conditions under which the microbial population produces the biogas. This can then help to better understand the hydrodynamic effects on the biogas production on sludge granules rather than the liquid Reynolds number which is a global parameter. The results presented in Fig. 5b show that the optimal mean shear rates (illustrated by red circle) vary from 28 - 48 s-1 respectively for the granules with an equivalent diameter ranging from 1.05 1.57 mm in order to achieve a maximum biogas production rate.
13
These original results reveal that there are two antagonist mechanisms to control the biogas production rate, i.e. the enhanced mass transfer and the compaction of granules caused by the high shear and they result in the existence of the optimal shear rate. It is worth noting that the mean shear rates obtained in this study are quantitatively consistent with the optimal values estimated in industrial ICA reactors as mentioned before. The biogas production is straightforwardly related to the shear rate exerted on the granules of different sizes. These data could provide new insight into the design and optimization of hydrodynamic conditions in various anaerobic reactors. In the practice, the strategy of a stepped increase of shear rate could be envisaged to find out the optimal range for both the wastewater treatment and biogas production.
CONCLUSIONS
The multiphase hydrodynamics was investigated in the upflow anaerobic reactors at different scales. A relationship between liquid circulation in downcomer and superficial gas velocity in central zone was established in the 3D reactor at macroscale. In the 2D reactor at mesoscale, hydrodynamics resulting from complex interactions between three phases was investigated and the relationship shear rates/sludge granules was related to the size of gas bubbles. In the 1D microreactor, a straightforward relationship between the shear rates exerted on the granule and the optimal biogas production is found and the optimal mean shear rates vary from 28 48 s-1.
14
ACKNOWLEDGEMENTS
The financial assistance provided by China Scholarship Council and the Agence Nationale de la Recherche (ANR PROMET) to the French group, and by the Natural Science Foundation of China (51061130555 and 50978147) to the Chinese group is gratefully acknowledged.
REFERENCES
1. Baloch, M.I., Akunna, J.C., Kierans, M., Collier, P.J. 2008. Structural analysis of anaerobic granules in a phase separated reactor by electron microscopy. Bioresource Technology, 99(5), 922-929.
2. Batstone, D.J., Keller, J. 2001. Variation of bulk properties of anaerobic granules with wastewater type. Water Research, 35 (7), 1723-1729. 3. Bhunia, P., Ghangrekar, M.M. 2007. Required minimum granule size in UASB reactor and characteristics variation with size. Bioresource Technology, 98 (5), 994-999. 4. Chan, Y.J., Chong, M.F., Law, C.L., Hassell, D.G. 2009. A review on anaerobic–aerobic treatment of industrial and municipal wastewater. Chemical Engineering Journal, 155 (1-2), 1-18.
5. Chu, C.Y., Wu, S.Y., Wu, Y.C., Lin, C.Y. 2011. Hydrodynamic behaviors in fermentative hydrogen bioreactors by pressure fluctuation analysis. Bioresource Technology, 102 (18), 8669-8675.
6. Costello, D.J., Greenfield, P.F., Lee, P.L. 1991a. Dynamic modeling of a single-stage highrate anaerobic reactor .1. Model derivation. Water Research, 25 (7), 847-858. 7. Costello, D.J., Greenfield, P.F., Lee, P.L. 1991b. Dynamic modeling of a single-stage highrate anaerobic reactor .2. Model verification. Water Research, 25(7), 859-871. 15
8. Dietrich, N., Poncin, S., Midoux, N., Li, H.Z. 2008. Bubble Formation Dynamics in Various Flow-Focusing Microdevices. Langmuir, 24(24), 13904-13911. 9. Ghangrekar, M.M., Asolekar, S.R., Joshi, S.G. 2005. Characteristics of sludge developed under different loading conditions during UASB reactor start-up and granulation. Water Research, 39(6), 1123-1133.
10. Hu, J.C. 2011. A method of design and calculation for internal circulation (IC) anaerobic reactor (in Chinese). China Biogas, 29, 3-6. 11. Hulshoff Pol, L.W., de Castro Lopes, S.I., Lettinga, G., Lens, P.N.L. 2004. Anaerobic sludge granulation. Water Research, 38 (6), 1376-1389. 12. Kassam, Z.A., Yerushalmi, L., Guiot, S.R. 2003. A market study on the anaerobic wastewater treatment systems. Water Air and Soil Pollution, 143(1-4), 179-192. 13. Latif, M.A., Ghufran, R., Wahid, Z.A., Ahmad, A. 2011. Integrated application of upflow anaerobic sludge blanket reactor for the treatment of wastewaters. Water Research, 45(16), 4683-4699. 14. Liu, Y., Tay, J.-H. 2002. The essential role of hydrodynamic shear force in the formation of biofilm and granular sludge. Water Research, 36 (7), 1653-1665. 15. Liu, Y., Xu, H.-L., Yang, S.-F., Tay, J.-H. 2003. Mechanisms and models for anaerobic granulation in upflow anaerobic sludge blanket reactor. Water Research, 37 (3), 661673. 16. Mu, Y., Yu, H.-Q., Wang, G. 2006. Permeabilities of anaerobic CH4-producing granules. Water Research, 40(9), 1811-1815.
17. Musee, N., Lorenzen, L. 2013. Market dynamics as a driver towards the evolution of research needs: the case of up-flow anaerobic sludge blanket seeding granules. Water SA, 39(1),131-142.
16
18. Pereboom, J.H.F., Vereijken, T. 1994. Methanogenic granule development in full-scale internal circulation reactors. Water Science and Technology, 30(8), 9-21. 19. Pevere, A., Guibaud, G., van Hullebusch, E., Lens, P., Baudu, M. 2006. Viscosity evolution of anaerobic granular sludge. Biochemical Engineering Journal, 27(3), 315322. 20. Ren, T.-T., Mu, Y., Ni, B.-J., Yu, H.-Q. 2009. Hydrodynamics of upflow anaerobic sludge blanket reactors. AIChE Journal, 55(2), 516-528. 21. Seghezzo, L., Zeeman, G., van Lier, J.B., Hamelers, H.V.M., Lettinga, G. 1998. A review: The anaerobic treatment of sewage in UASB and EGSB reactors. Bioresource Technology, 65 (3), 175-190.
22. Tauseef, S.M., Abbasi, T., Abbasi, S.A. 2013. Energy recovery from wastewaters with high-rate anaerobic digesters. Renewable and Sustainable Energy Reviews, 19, 704741.23. Uemura, S., Harada, H. 2000. Treatment of sewage by a UASB reactor under moderate to low temperature conditions. Bioresource Technology, 72(3), 275-282. 24. Vesvikar, M.S., Al-Dahhan, M. 2005. Flow pattern visualization in a mimic anaerobic digester using CFD. Biotechnology and Bioengineering, 89(6), 719-732. 25. Wu, B. 2010. CFD simulation of gas and non-Newtonian fluid two-phase flow in anaerobic digesters. Water Research, 44(13), 3861-3874. 26. Wu, B., Chen, S. 2008. CFD simulation of non-Newtonian fluid flow in anaerobic digesters. Biotechnology and Bioengineering, 99(3), 700-711. 27. Wu, J., Lu, Z.Y., Hu, J.C., Feng, L., Huang, J.D., Gu, X.S. 2006. Disruption of granules by hydrodynamic force in internal circulation anaerobic reactor. Water Science and Technology, 54 (9), 9-16.
17
28. Wu, J., Zhou, H.M., Li, H.Z., Zhang, P.C., Jiang, J. 2009. Impacts of hydrodynamic shear force on nucleation of flocculent sludge in anaerobic reactor. Water Research, 43(12), 3029-3036. 29. Zhang, J.B., Wu, J., Poncin, S., Hamelin, M., Li, H.Z. 2012. Microscale Investigation of Anaerobic
Biogas
Production
under
Various
Hydrodynamic
Conditions.
Environmental Science & Technology, 46(16), 8698-8704.
30. Zheng, M.X., Wang, K.J., Zuo, J.E., Yan, Z., Fang, H., Yu, J.W. 2012. Flow pattern analysis of a full-scale expanded granular sludge bed-type reactor under different organic loading rates. Bioresource Technology, 107, 33-40.
Caption of figures
Figure 1. Schematic presentation of (a) 3D macroreactor; (b) 2D mesoreactor and (c) 1D microreactor.
Figure 2. Circulation velocity in the 3D macroreactor. (a) in function of superficial gas velocity with different influent liquid flow rates in the presence of granular sludge (10 g SS/L), (b) in function of influent liquid flow rate with different superficial gas velocities without or with granular sludge (10 g SS/L).
Figure 3. (a) Liquid velocity profile of in a half-pipe section and (b) instantaneous liquid flow fields in the downcomer of the 3D macroreactor.
Figure 4. Variation of the mean shear rate in the main section of the 2D reactor at mesoscale. (a) in function of the superficial gas velocity, (b) in function of the mean bubble diameter 18
under a gas velocity of 9.61×10-4 m/s with a sintered distributor and a sieve distributor respectively. (c) Instantaneous shear rate field with cross symbols representing sludge granules’ position at superficial gas velocity of 2.99×10-4 m·s-1 in the 2D reactor.
Figure 5. (a) Flow fields around a granule of 1.36 mm producing biogas obtained by the micro-PIV. (b) Relationship between the rate of biogas production and the mean shear rate exerted by the liquid on the granules of different diameter.
19
(a)
(c)
(b)
Figure 1
20
Figure 2
21
(b)
(a)
Figure 3
22
Figure 4
23
(a)
(b)
Figure 5
24
Table 1. Experimental values of Reynolds number under different superficial gas and liquid velocities
influent liquid
influent liquid
influent liquid
superficial gas
velocity 0 m/s with
velocity 2.21×10-3
velocity 4.24×10-3
velocity (m/s)
granular sludge 10 g
m/s with granular
m/s with granular
SS/L
sludge 10 g SS/L
sludge 10 g SS/L
3.80×10-5
204
195
140
6.69×10-5
286
193
139
1.56×10-4
301
295
265
1.94×10-4
395
312
262
2.31×10-4
455
342
350
25
Mouvement and size of bubble Liquid flow field Shear rate field Circulation velocity
ro s
Hydrodynamics in upflow anaerobic reactors
26
PIV Comprehension and improvement
al e
M ac
c os icr
Mesoscale
M
ca le
High-speed digital camera
Micro-PIV Microscope
Highlights • Hydrodynamics plays a major role in improving the performance of upflow reactors. • Gas > solid > liquid can be ordered as for the circulation velocity. • Motion between bubbles and liquid is the key parameter affecting the shear rates. • Shear rate exerted on a sludge granule displays an optimum biogas production rate.
27