Anaerobic downflow stationary fixed-film treatment (III). Anaerobic expanded/fluidized bed treatment (V) Comparison of high-rate anaerobic wastewater treatment reactors (VI).

Tomado de: Anaerobic wastewater treatment-attached growth and sludge blanket process.
Por: S. Vigneswaran; B.L.N. Balasuriya y T. Viraraghavan
Bangkok: ENSIC, 1986

General

Downflow stationary fixed film (DSFF) reactors are a relatively recent addition to the family of advanced high-rate anaerobic reactors, all of which are based on retention of the active biomass. The DSFF reactor distinguishes itself from other type of advanced reactors by the downflow mode of operation, the architecture of its packing (fixed biofilm support), and the absence or near-absence of suspended growth.

Principles

The method of operation of DSFF reactors is shown in Fig. 3.1. Waste is pumped in at the top, together with recycle effluent when desired, and effluent is withdrawn from the bottom. When started, inoculum from an active digester is recirculated and bacteria attach themselves to the channel walls of the support material to form a biofilm. The downflow mode allows any settleable material which might otherwise accumulate in the system to be removed with the effluent and reduce the risk of column plugging. Mixing in the reactor is produced entirely by the action of rising gas bubbles. Hence the high concentration of substrates are immediately dispersed and there is little need for an elaborate agitation system. With counter-current two-phase flow, the system exhibits good mixing characteristics and reduces potential for high localized concentrations of inhibitors or volatile acids (Hall & Melcer, 1984).

At present time, the anaerobic DSFF reactor is in an infant stage of development and its application for municipal waste is not being tested although studies with laboratory, pilot scale reactors and industrial scale reactors are documented.

System Design

The design of DSFF reactor consists of wastewater distribution system, a biofilm support structure, head space, effluent draw off and recycle facilities.

The downflow mode of operation disctated a stationary film support to maintain the film of micro-organisms in the reactor. Additionally, to prevent setting of suspended solids on parts of the film support surface, the stationary film support is arranged in more or less vertical channels and are made of potters clay, draintile clay, needle punched polyester or polyvinyl chloride. Reactor size and height have relatively little effect on the performance (when expressed in surface- to-volume ratio) but the reactor configuration and operation have marked effect on performance. Generally muti-channel reactors have not performed as well as reactors with only few channels (van den Berg et al. 1985).

Recirculation

As quoted by van den Berg et al.(1985), recirculation generally improves the performance of the DSFF reactor (Duff & Kennedy, 1982; Samson et al., 1984. With wastes containing large amounts of hard-to-digest suspended solids (e.g. piggery waste) recirculation helps to keep these in suspension and aid in their degradation (Kennedy & van den Berg, 1982b). Recirculation of effluent helps to maintain a uniform and relatively thin film (van den Berg et al., 1985).

Design Criteria

        Support Material

Support material affects the rate of start-up markedly. The effect of support material on start-up and steady-state performance is shown in Fig. 3.2

Reactors made from rigid foam polyvinyl chloride could not be started at all, while the glass reactors were slow to start up, presumably because bacteria had difficulty attaching themselves to the smooth inert surface. Solid polyvinyl chloride (PVC, used extensively in biological wastewater treatment) was substantially better than glass as a film support, buy not as good as the fired clay and needle punched polyester. The earlier studies also indicated that inert support media with a rough surface enhanced biofilm accumulation and reactor performance. Physical roughening of smooth plastic surfaces and addition of sawdust to clay support media before firing (unpublished results) have been shown to enhance start-up and overall reactor performance (van den Berg et.al., 1985).

             Surface-to-Volume Ratio

The amount of retained biomass in a DSFF reactor depends on the surface to volume ratio and is therefore limited by the support matrix area. The importance of surface-to-volume ratio on start-up and ultimate loading rates for reactors of the same height is shown in Fig. 3.3. Reactors with larger surface-to-volume ratio achieved higher space loading rates and higher rates of methane gas production (Kennedy & Droste, 1984).

       Organic Volumetric Loading Rate

Loading rates and organic removal efficiencies depended on the total amount of active biomass retained as well as on type of waste. For a wide variety of industrial wastes, loading rates of 5 to 15 kg COD/m³-d were readily obtained with 70-95% COD removal efficiency, depending on loading rate and type of waste (van den Berg et al., 1985).

Van den Berg et.al. (1980) with their experience with bean blanching waste, found that the small fixed film reactors could be loades substantially higher than larger ones and this difference is not explained. The size of the reactor did not affect the COD removal efficiency.

Maximum COD loading rates, using bean blanching waste (9,500-10,500 COD mg/L, mainly soluble) were as high as 20 kg COD/m3-d (depending on surface to volume ratio and size) to achieve 86% removal efficiency (Table 3.1)

          Table 3.1   Performance data for anaerobic fixed film reactors of two different sizes, fed
                           with bean blanching waste (van den Berg et.al., 1980)

     a  surface to volume ratio, 140 m²/m^3.
     b  surface to volume ratio, 120 m²/m³ - reactor may not have reached maximum loading
         rate in test run
     c  Independent of waste strength (0.5-2.0% COD)

For simulated sewage sludge (55,000 COD mg/L, over 45,000 which is suspended) the maximum COD loading rate of 12 kg COD/m³-d with an efficiency of 70% removal is reported

          Temperature

Kennedy & van den Berg (1981) reported the effect of temperature on the performance of DSFF reactors treating bean blanching waste and chemical industry wastes. The maximum loading rates decreased linearly with the temperature. The fixed film reactor treating bean blanching waste, mainly containing soluble starch and protein (Total COD = 10 g/L) was reported to be capable of achieving high loading rates without a substantial change in COD removal efficiency at a temperature range of 10°C to 35°C.
They observed that a decrease in temperature from 35°C to 25°C, decreased the maximum steady-state loading rate by 37%, while at 10°C the loading rate was reduced by 75% of   the maximum loading rate at 35° C (i.e.18.4 kg COD/m³d). The COD removal efficiency was independent of temperature and remained at 88 + 3%.

Similar results were obtained for chemical industry waste (Total COD = 14 g/L).
The maximum steady-state loading rate and the volumeric methane production rate decreased by less than 25% between 35° C and 25°C. As with the bean blanching waste there was no appreciable change in the COD removal efficiency or in digester gas composition with temperature. The effect of temperature on the performance of stationary fixed film reactors digesting chemical industry waste at maximum loading rate is presented in Table 3.2. The effect of temperature on loading rate and rate of methane production is given in Fig. 3.4.

       Table 3.2. Effect of temperature on the performance of stationary fixed-film reactors
                       digesting chemical industry waste at maximum loading rates (Kennedy & van
                       den Berg, 1981)

Kennedy & van den Berg (1981) proposed relationships between temperature and loading rate and temperature and methane production rate expressed by: 

                GP = 0.167T - 0.692
where: GP = methane production rate (m³ (STP)/m³-d)
                 T   = operating temperature (°C)

and          LR = 0.557T - 1.546

where:     LR = COD loading rate (kg COD/m³-d)

The coefficients in these relationships presumably depend on the nature of the substrate and the type of fixed-film support material as well as the support material configuration and surface-to-volume ratio. 

van den Berg et al. (1985) reported the studies conducted by Kennedy & van den Berg in 1982 which stated that the DSFF reactors can be operated at thermophilic temperatures (55°C), but maximum loading rates and COD removals were similar to those at mesophilic temperatures. 

Hydraulic Retention Time

Hydraulic retention time (HRT) is the ratio between the void volume of the reactor and the volumetric flow rate. An intermediate HRT is desirable for some high strength waste due to poor conversion of wastes to methane at very short HRT's. With laboratory experience, it was found that the HRT varies from few hours to number of days depending on the wastes characteristics and the strength. 

Application Status

van den Berg et al. (1985) summarized the performance data of downflow stationary fixed film reactors (DSFF) at steady-state loading (Table 3.3) for a wide variety of wastes tested and for a wide range of reactor sizes. COD loading rates between 5 to 15 kg/m³-d were readily maintained and COD removal efficiencies of 60 to 95% are reported. The composition of wastes treated in DSFF reactors is given in Table 3.4. 

Samson et al., 1984 reported the experiences with an industrial scale 400 m³ reactor, using channelled ceramic blocks as support material which has been operating on cheese factory effluent for a year and is quoted by van den Berg et al. (1985). The effluent varied widely from day to day and from hour to hour. Total COD varied from about 800 to over 2,500 mg/L and the reactor was designed for a loading rate of about 7 kg COD/m³-d at an average HRT of 7 hours. A COD removal efficiency of 60-70% at 8 kg COD/m³-d loading rate was reported. 

A couple of 50 m³ DSFF reactors for piggery waste have been built in Canada and initial findings confirm the results presented above (Hall, 1983). A very large (13,000 m³) DSFF reactor treating rum stillage waste at a concentration of up to 100,000 ppm is in operation in Puerto Rico (Szendrey, 1983, 1984). The plant operates at a loading rate of about 8 kg/m³-d and a COD conversion efficiency of 75%. it has been found to be very stable in withstanding variations in loading rate and wastes composition and its ability to start-up after a shut down period (van den Berg et al., 1985).

Table 3.3. Performance data for downflow stationary fixed film reactors (van den Berg et al., 1985)

 Table 3.4.  Composition of wastes treated in downflow stationary fixed film reactors
                    (van den Berg et.al.,  1985)

    ^a Average or most common concentration used; range used in brackets
    ^b Sodium hydroxide added to increase alkalinity
    ^c NH₄HCO₃ added
    ^d Sodium and potassium phosphate added

Applicability

Stationary fixed film reactors could be changed over from one waste to another with relatively littie loss of capacity and could adapt readily to changes in temperature as low as 10°C. This is important for installations where the character of the wastewater changes rapidiy due to the season or production schedules (van den Berg, 1982). van den Berg et al. (1981) reported that the reactors could start-up very quickly after a period of starvation (one or two days to reach maximum capacity after 3 weeks of starvation). 

This reactor could be used to remove the treated waste water intermittently than continuously and intermittent addition of wastes appears to be feasible (van den Berg, 1982). lntermittent loading increases the rate of methane production and hence the rate of conversíon of COD, but decreases the COD removal. The latter may be caused by the short hydraulic retention time for part of the waste (van den Berg et al., 1981). 

In stationary fixed film reactors COD removal depends on types of wastes and hydraulic retention times. Waste with hard to digest solids showed lower removals, particularly at short hydraulic retention time. Reactors could handle both low and high nitrogen wastes (pear peeling waste, piggery waste). Further, this fixed film reactor with effluent removal from the bottom has been found to produce methane with high suspended solids contents (van den Berg, 1982). 

Due to the self mixing feature of the fixed film reactors, they could treat (this mixing is produced by the rising gas bubbles which causes every channel to act as gas lift pumps) dilute and concentrated wastes equally well. The rapid self mixing distribute wastes quickly through the reactor before local high concentration of volatile acids could develop. 

The DSFF reactors could handle severe hydraulic overloading and organic shock loads without serious problems and could be operated at temperatures lower than optimum and still be loaded at high rates without affecting digester gas composition or COD removal efficiency. 

Mesophilic DSFF reactors tolerated sudden organic shock loads at constant hydraulic loading (caused by sudden increase in waste strength) and recovered normal performance within a few days, if the alkalinity was sufficiently high to maintain the pH above 6.2 (Kennedy et al., 1984). 

Kennedy & van den Berg (1981) reported that for chemical industry waste (TCOD = 14 9/L) the DSFF reactors could be over loaded 8 times their normal rate for a 24-hour period and, recovery to be possible within 12-48 hours while still being loaded normally. COD removal decreased with increasing overloading rates and was temperature dependent. During overloading at a loading rate of 61 kg COD/m3-d (0.43 kg COD/m2-d), COD removal efficiencies at 25°C and 35°C were 44% and 61% respectively. Results of overloading tests including wastes other than chemical wastes are given in Table 3.5. 

lt is evident that the COD removal during overloading decreased with increased rate of overloading while methane production rates increased. Further, Kennedy & van den Berg (1981) stated that repeated overloading improve the reactors as they are more stable and could be loaded at higher steady state rates than before overloading. This may be due to activation of inactive film by the availability of substrate and nutrients (Kennedy & van den Berg, 1981).

lt is reported that sloughing of the biofilm occurred during organic and hydraulic shock loadings (van den Berg et al., 1985). 

Tests to determine the ability of DSFF reactors to handle toxic shock loads are underway. Initial results indicate that the DSFF reactor is able to withstand large toxic shock loads (van den Berg et al., 1985). 

Problems Associated with Downflow Stationary Fixed Film Reactor

Plugging of the Reactor

Non-uniformity of biofilm thickness in DSFF reactor occurs due to excessive growth of biofilm near the top of the reactor. Under certain conditions, this non-uniform growth can cause plugging at the top of the packing (Hall, 1983) and partial plugging of some channels (Samson et al., 1985).

 Factors like, width of the channels, smoothness of the packing (will determine the smoothness of occasional sloughing), recirculation rate and the composition of the waste determine whether or not plugging wiil occur (van den Berg et al., 1985). 

van den Berg et al. (1985) reported on several methods which have potencial for maintaining a reasonably thin biofilm in DSFF reactors. These are: 

- organic and hydraulic shock loads - the effect on film sloughing will be greatest near the top of
   the reactor where the load enters. This method will not be of much use for channels already
   locked. 

- Recirculation of effluent - as already discussed, for wastes with a high suspended solids
   content, intermittent pumping of liquid from the bottom to the top of ttie reactor helps to
   maintain a uniform, relatively thin film. 

- Recirculation of gas - large gas bubbles rising in channels should help the sloughing process
   ard may even open blocked channels. 

- Reactor configuration - a relatively thin layer of coarse packing on top of the ordinary packing
   may accumulate the excess biofilm and cope with it. 

- Improvement in the flow distribution system on top of the reactor to avoid too low liquid
   velocity in these channels. 

- Horizontal spacing of channels to improve rnixing and reduce dead space.

          Start-up of the Reactor

Rate of start-up depends on the type of inoculum, the type and strength of waste, level of volante acids maintained and the characteristics of the support material used. For example reactors were difficult to start-up with chemical industry waste (toxicity could not be demonstrated), while reactors started readily on food processing wastes or sugar waste. Sewage digester sludge generally required a longer time to adapt than inoculum from an active digester fed with food processing waste. The rate of start-up was faster with sugar waste at 5,000 mg COD/L than with a strength of 10,000 mg COD/L or higher. Also, reactors started up faster when volante acid levels were maintained at about 1,000 mg/L than with levels below 600 mg/L (Kennedy & Droste,

1984).: Several factors presumably play a role: concentration of critical types of bacteria, ecological relationships and how close the waste resembled the substrate to which the inoculum was accustomed.

Support material as well as the number of channels in a reactor affect the start-up and ultimate loading rates (van den Berg et al., 1985). This effect was caused by differences in rnixing patterns because it affects the amount of dead space and short circuiting. Recirculation rate also improves mixing patterns and rate of start-up (Samson et al., 1985). Horizontal spaces in banks of vertical channels also provide an improvement in the rate of start-up by reducing the amount of dead space (Samson et al. , 1985).

With DSFF reactors, the rate of increase in COD loading rate was less consistent. According to the experiments carried out by van den Berg et al. (1980), it was reported that when simulated sewage siudge is fed, doubling of the COD loading rate took generally 30 days or longer while with bean blanching waste the time required to reach a doubling of the COD loading rate was as low as 10-15 days for 0.8 liter reactors and as high as 40-50 days for 35 liter reactors.

In addition, the COD loading rate had got 'stuck' for several weeks at a loading rate well below the ultimate maximum loading rate (Fig. 3.5). This reflected the variable and often less optimum degree of mixing in fixed film reactor, during the start-up. it may be possible that the types of organisms occupying fixed film surfaces during the start-up are not present in a relationship which is quantitatively and qualitatively optimal for high rate performance and readjustment during the start-up may therefore be necessary from time to time, causing a delay in reaching the maximum COD loading rate (van den Berg et al., 1980).

    Advantages and Disadvantages

          Advantages

          -   Elimination of mechanical mixing (mixing in the reactor is provides entirely by the
               action of rising bubbles)

          -   Recycling is not necessary

          -   Better stability at higher loading rates

          -   Simplicity in construction

        Disadvantages

         -  Lower quality in circumstances where the influent suspended solids concentration is
             high.

         -  Requires more care in starting-up of the reactor.

Conclusion

The DSFF reactors appear to be outstanding in reliability and ability to withstand adverse conditions, but generally do not achieve very high loading rates or COD removal efficiencies. The DSFF reactor is at an infant stage of development and further experiences at pilot plant and full scale level are needed.

Anaerobic Expanded/Fluidized Bed Treatment

   General

The anaerobic fixed bed process was found to be hampered  by clogging and inefficient contact of the micro-organisms and the wastewater due to influent suspended solids or excess biomass which creates inactive zones and channeling. Unlike fixed bed reactors, bed fluidization results in little or no short-circuiting and small pressure gradients. Jewell (1974) proposed the attached film expanded bed process as a means of optimizing aerobic systems. This was based on the assumption that large biomass concentrations could be achieved on the large surface area provided by the small sand size particles. The small particles, when fluidized, would minimize diffusionall limitations and eliminate clogging problems. Later Jewell & Switzenbaum (1980) demonstrated that it was possible to utilize this concept, using an anaerobic film.

Anaerobic Expanded / Fluidized Bed Reactor

Expanded/fluidized bed reactors have much larger surface area per unit reactor volume, which increases the reactor micro-organism concentration. The larger specific surface area allows shorter hydraulic retention times or lower operating temperatures for the same degree of treatment in a given volume. A specific surface area of 3000 m2 / m3 has been reported for fluidized bed reactor and the concentration of the microorganism of 30 g/L have been measured in anaerobic processes (Boening & Larson, 1982).

For anaerobic treatment of wastewater, only a few studies using expanded / fluidized beds have been reported. Switzenbaum (1978) used anaerobic attached film expanded process (AAFEB) and found it to be effective for the treatment of low strength soluble organic wastes at reduced temperatures, at short retention timers and at high organic loading rates. Walker (1981) tested pilot plants containing large scale fluidized bed bioreactors as advanced treatment processes for denitrification of industrial wastewaters and he demonstrated long-term stable operation of the unit, as well as the ability to meet stringent discharge limits. Hickey and Owens (1981) used a similar process and have shown this to be effective for the simultaneous generation of methane gas and stabilization of high strength wastewaters including dairy, chemical, food processing, soft drink bottling and heart transfer liquids.

Packed bed reactors such as anaerobic filters often experience problems or increased pressure drop due to the accumulation of biomass. To prevent plugging, relatively large voidage must be maintained which limits specific surface area and biomass concentration. Expanded/fluidized bed reactors overcome these problems allowing the use of low voidage, high surface area of the media.

Principles and Theory of Fluidization

In an expanded bed, the particies remain in stationary contact whilst in a fluidized bed the particles are in free motion. When the liquid is passed upward through an unrestrained bed of particles, the bed will initially expand slightly to take up a loose packed arrangement. If the flow is increased, the pressure drop across the bed increases as shown by the line OA in Fig. 4.1. Eventually the pressure drop equals the force of gravity (corrected for the buoyancy in the liquid) on the particles and the grains begin to move. This is the point 'A' in the figure. During this period, the prosity increases and the pressure drop rises more slowly than before due to the net effect of increased porosity and velocity. When point 'B' is reached, the bed is in extremely loose condition with the grains still in contact. Between points 'A 'and 'B ' the bed is unstable, the particles begin to loose contact and then adjust their position to present as little resistance to the flow as possible.

As the velocity is further increased, the grains separate and true fluidization begins. This is the point 'F' on the figure. By the time this point is reached, all particles are in motion and beyond that point, the bed continues to expand and the particles move in a more rapid and independent motion. The bed continues to expand as the velocity is increased and maintains a uniform character. As the fluid velocity is increased further, the porosity increases, the bed of solids expands and eventually at point 'P' on the figure, all of the particles have been entrained in the fluid. The porosity approaches one and the bed ceases to exist. From this point on, there exists the simultaneous flow of two phases. At point 'P' the superficial velocity is approximately equal to the terminal settling velocity of the particles.

The systems, as applied to wastewater treatment, consist of inert sand-sized particles in a column which expand with the upward flow of waste through the column. The inert particles act as a support surface for the growth of attached organisms. Fluidization allows the entire surface area of each particle to become available for biological attachment and subsequent reaction.

In anaerobic systems, both the expanded and fluidized beds operate at less than full fluidization (Switzenbaum, 1983).

The biomass hold up is achieved by allowing natural mechanism of flocculation and adhesion to take place in a low shear environment and then to remove excess growth by particle/particle or particle/wall contact.

Design Elements

Like the anaerobic filter, the design of an expanded/fluidized bed consists of a wastewater distributor, a medium support structure, medium, head space, effluent draw off and recycle facilities. For high strength wastes, a device for separating the excess biomass from the support medium and subsequent wasting of this excess growth is generally incorporated into de design.

Recycling

Since the medium should be kept in the fluidized state, upflow velocities must be high enough to keep the particles in suspension, and thus effluent recycle is practised.

In addition to this, a certain amount of recycle is useful as it can:

-  help neutralize the pH of the incoming wastewater;
-  reduce the amount of alkalinity required;
-  reduce the effect of toxic biodegradable compounds;
-  minimize the effect of shock loadings; and
-  compensate for variability of influent flow rate.

Increasing recycle in effect allows the process to tend towards the results and operational characteristics of a completely mixed system.

Separation Equipment

In fluidized beds, some sort of separation equipment or other biomass retention measure is incorporated in the design as process failure could result in total loss of biomass within 15 minutes. Expanded bed reactors thus have some inherent risk.

Design Criteria

      General

The expanded/fuidized bed reactor can be designed using either the organic volumetric loading rate (OVLR) or solid retention time (SRT) approach. The kinetics of substrate removal in the fluidized bed reactor will determine the actual SRI required for a given degree of treatment efficiency. Once the SRT is established together with values for the kinetic parameters, the kinetic equations can be used together with the information from fluidization mechanics to establish values for other design parameters (hydraulic retention time, fluidization velicity, recycle ratio, etc.).

A less rigorous and strictly empirical approach will be to use the organic volumetric loading rate (OVLR) to achieve a given degree of treatment.

   Solid Retention Time

The solid retention time (SRT) is the average retention time of organisms in the system. In the expanded/fluidized bed process the SRT is normally defined as:

SRT  =       volatile suspended solids (VSS) in the reactor         
             volatile solids lost in the effluent or intentionally wasted/day

In a biological reactor the organism specific growth rate is equal to the reciprocal of the solid retention times of the system.

The organism specific growth rate is expressed according to the kinetic equation given:  

   u  =  1    dx      =   YK - b

where;

  u  =  organisms specific growth rate, m³/m³-h
  x  =  organisms concentration, kg. VSS/m³
  y  =  organisms specific yeild coefficient
  k  =  specific substrate utilization rate, kg. COD/kg. VSS-h

  b  =  organism decay coefficient, h ^-1

thus

    1      = YK - b
  SRT

      Organic Volumetric Loading Rate (OVLR)

Rudd et.al. (1985) found that the biomass concentration was affected by variation in the organic loading, influent substrate concentration, and hydraulic retention time. Schrra and Jewell (1984) observed that the COD loading appears to be the major determinant of biomass concentration.

Numerous authors (Stephenson & Murply, 1980; Sutton et.al., 1981; Jewell, 1981) have used this parameter to illustrate the efficiency of fluidized bed reactors in comparison to other systems in treating wastewaters. The OVLR is used when it is difficult to determine the reactor biomass concentration.

The organic volumetric losding rate (OVLR) to the system is defined as:

OVLR  =  QS_o  =  S_o
                                  
                  V           T

Where

     Q =  influent flowrate, m3/d

     S_o  = influent substrate concentration, kg/m3

     V   = reactor volume, m3

     T   = reactor hydraulic retention time, d

Numerous pilot scale tests have shown high COD removal of more than or equal to 80% at COD loading of 10-20 kg/m³-d for variety of industrial wastes.

Hydraulic Retention Time  (HRT)

Hydraulic retention time (HRT) is calculated on the basis of expanded/fluidized bed volume. The HRT is the ratio between the expanded/fluidized bed volume and the influent flow rate of the wastewater.

In the case of wastewater treatment, due to the relative insensitivity of the process performance to HRT, the system should be designed at a low HRT (of the order of several hours). The actual design HRT depend on wastewater organic strength (Switzenbaum et al., 1984). 

Laboratory and Pilot-scale experiments have been conducted with HRT vaiues varying from a low of five minutes to several days for various industrial wastes. COD loadings have ranged from 0.65 to 60 kg/M3-d (Switzenbaum, 1983). 

Rudd et al. (1985) and Schraa & Jewel (1984) reported that the organic removal efficiency d'ecreased with decreasing HRT at a constant loading rate. Optimum HRT for mesophiiic (350C) anaerobic fluidized bed reactors lie within the range of 6-13 hours (Rudd et al., 1985). 

Recycle Ratio

Expanded and fluidized beds operate with very high recycle rates. For concentrated wastes, a very high degree of recycle is needed in order lo keep the bed particles in suspension and at the same time to dilute the organic materials present in high coricentration. For dilute wastes, like municipal wastewater, the recycle ratio is reduced to reasonably low values. 

The recycle ratio (µ ) is given by the ratio between the influent flow rate (Q) and the recycle flow rate (Q_r).

Thus, recycle ratio ( µ )         =     Q 
                                                              Q_r

Based on the design values of substrate removal rate, vx = 0.01 kg.COD/kg.VSS.h; biomass concentration, X = 20 kg. VSS/m³; hydraulic loading rate, Q = 10 m³/m²-h; COD removal efficiency, E = 80%; and total particle density, p = 1.02 x 103 kg/m³; the following figure (Fig 4.2) is developed to show the effect of influent wastewater concentration, Co upon the degree of recycle required in expanded and fluidized bed reactors.

However, the assumed upflow superficial velocity of 10 m³/m²-h is an estimation based on an overall bed particle specific density identical for pure biological flocs. For very thin biofilms on high density media, the particle size will be of great importance in considerations regarding the needed superficial upflow velocity (Jewel, 1982).

Filter Media

The medium used for biofilm attachment is comprised of small diameter inert particles, such as sand, anthracite, or granular activated carborn, which are maintained in the fluidized state.

Various support media have been tested, including sand, PVC particles, granular activated carbon, and diatomaceous earth. A range of particle sized and densities have been examined. There are trade-offs between size and density of particles and stability of operation with these systems. Smaller particles provide greater specific surface area-to-volume ratio and thus provide greater surfaces for attached biofilms. In addition, lighter particles can be fluidized at lower upflow velocities which reduce the recycle rate necessary to achieve a given HRT.

                    Fig. 4.2  Recycle rates for expanded/fluidized beds

In the design of an anaerobic expanded/fluidized bed, small and light particles which are easy to flidize and provide a large specific surface are being used. There would exist however, a minimal particle size and/or density in order to prevent carry-over.

If a high shear stage is to be used to disengage the biomass and the medium then it is likely that a tough medium, such as sand, would be chosen in preference to a more fragile one like activated carbon (Cooper & Wheeldon, 1980).

The technical data based on available literature are given in Table 4.1

                       Table 4.1. Technical data (Henz & Harremoes, 1982)

   Temperature

Temperature was found to be an important variable affecting process efficiency, but the process was found to compensate well for changes in temperature. However, the optium temperature for the treatment is around 35 ° C.

Schraa & Jewell (1984), reported high rate conversions of soluble organics with anaerobic fluidized bed reactors at termophilic (55° C) temperatures, after a 5 month accumulation of biomass. A mature microbial attached film was developed successfully at 55°C over a short time period and high solids concentrations and film depths (60 g/L and 170 cm respectively) were achieved with these thermophilic films Medium strength (1.5 to 3 g/L COD) and high strength (5-16 g/L COD) soluble wastes were treated with a 70% removal efficiency at a volumetric loading rate of 30 g/L.d COD.

Results obtained by Rudd et.al (1985) indicate that thermophilic anaerobic fluidized bed reactors are inferior to mesophilic reactors in several ways. The most important was the inferior organic removal efficiencies ahieved by thermophilic anaerobic fluidized bed reactors under a number of operating conditions.

Process Design

Since anaerobic fluidized bed treatment technology is relatively a new approach, process design criteria is not available at present time.

Pilot plant studies were completed by Dorr-Oliver T.M. of U.S.A., for industrial wastewater treatment as reported by Sutton and Li (1982) in order to derive information for process design of single phase and two phase anaerobic fluidized bed systems. Sutton & Li (1982) presented the range of operating conditions and the range of process design parameter values for single-and two-phased fluidized bed systems (Tables 4.2., 4.3 and 4.4.)

The design values presented by Sutton and Li (1982) for single phase and two phase fluidized bed reactors are summarized in Table 4.5.

                  Table 4.2  Operating conditions for fluidized bed reactor of
                                    single phase system (Sutton & Li, 1982)

     * Percent expansion of settled sand bed.

               Table 4.3. Operating conditions for fluidized bed reactors
                                 of two phase system (Sutton & Li, 1982)

  * Percent expansion of settled sand bed.

Tabla 4.4. Design parameter values for single phase and two-phase fluidized bed
                  reactors operated at 30°C to 35°C (Sutton & Li, 1982)

Different Design Configurations

Dorr-Oliver's Anitron System

Fig. 4.3 shows a pilot plant that was used by Dorr-Oliver and reported by Sutton & Li (1982). The units were skid-mounted, self contained units. The fluidized bed reactor consisted of a clear PVC column having a diameter of 16.2 cm. The height of the reactor to the effluent out-let port was approximately 3m. The fluidized bed height was controlled below 2.44m. Additional pilot plant components included a refrigerated feed tank, feed and recycle pumps, a means for foam control, gas liquid separator, a wet-tect meter for gas flow measurement, and various other instrumentation. A refrigerated holding tank was installed between the two pilot plants to help balance the feed flow rate during two phaseed operation. Temperature control of each reactor was achieved by adjusting the temperature of the recycle stream using an electrically controlled heat exchnger. The pH of each reactor was controlled by the addition of either sodium bicarbonate or sodium hydroxide. Ammonium dibasic phosphates were added to ensure a proper nutrient balance for biological growth.

          Table 4.5     Single - and two-phase fluidized bed design values
                                                       
       

In tapered configuration (Fig 4.4), the cross sectional area gradually increases from bottom to the top of the reactor. This configuration provides the flow patterns a minimal backmixing, especially at the feed entry point and prevents plugging by maintaining a high inlet velocity. Relatively stable flow through the reactor can be achieved when the entry cross section is sufficiently small and the expansion is gradual (an angle of few degree). Since the fluid velocity decreases with the reactor, the height of the column allows a wide range of flow rates without the loss of bed material, Hence, the tapered fluidized bed can effectively operate over a wide range of feed flow rates. This system can maintain a stable operation over a wide range of volumetric flow rates than the conventional fluidized bed.

Application Status

Municipal Applications

The initial research on the anaerobic expanded/fluidized bed was conducted on primary effluent from the Ithaca (New York) Sewage Treatment Plant, using a process called anaerobic attached film expanded bed process. The work reported by Jewell et.al (1981) found the process to be capable of achieving COD removals of 60-80% at HRT of 0.5-8 h at 20°C. Effluent quality was 40-5- mg/L COD and 5-15 mg/L SS at an HRT of 1 to 5 h.

A later laboratory study conducted by Froster (1984), for settled sewage obtained daily from a sewerline running along the perimeter of the University of Birmingnam campus, U.K., showed an unsatisfactory performance of this process under unsteady conditions. The efficiencies were significantly lower than those quoted previously.

A study on anaerobic fluidized beds was conducted at U.S. EPA's Test and Evaluation Facility in Cincinnati, Ohio, investigating the treatment of a municipal wastewater (containing industrial wastes) with an average sultate concentration of 250 mg/L (Bowker, 1983). COD removal efficiencies ranged from 40 to 50% for hydraulic retention times of 6 to 24 hours. Methane production was found to be low due to the effect of the thermodynamically favored reduction of the influent sultate that was occurring in the reactors.

A pilot-scale testing of the anaerobic fluidized bed process was conducted by Switzengbaum et.al. (1984) for the primary effluent from the Amherst wastewater treatment plant at the University of Massachusetts/Amherst, U.S.A. Sand was used as the support material. This

study was conducted over a five-month intensive testing period and over a wide range of organic volumetric loading rates, (0.18-0.41 g BOD/L-d) and at a hidraulic retention time varying from 1.67 to 6.37 hrs; a mean effluent 5 day BOD of 47.2 mg/L (standard deviation = 15.5 mg/L) and a mean suspended solids (SS) concentration of 30.5 mg/L (standard deviation = 16.6 mg/L) were achieved. The influent BOD₅ and SS were 74.2 mg/L and 35.5 mg/L respectively. Relatively low percentages of removal were obtained. A comparison of pilot-scale data (Switzenbaum et.al, 1984) with the laboratory scale data (Jewell et al., 1981) is presented in Table 4.6.

               Table 4.6   Comparison of laboratory and pilot scale units

*   Data of Switzenbaum et.al.(1984).
** Data of Jewell et.al. (1981).

Culp/Wesner/Culp consultants (1980) concluded taht the fluidized bed SMAR appeared to have its greatest potential as a secondary treatment unit process due to its ability to treat low-strength soluble wastes and produce high quality effluents. They found the fluidized bed SMAR comparable to conventional aerobic systems in the treatment of dilute organic wastes in terms of organic remocal efficiencies, detention times and organic loading rates. They also developed process flow sheets and cost details for treatment schmes incorporating fluidized bed SMAR for wastewater flows of 1 mgd and 25 mgd, and found that these costs compared favorably with costs for conventional aerobic treatment processes.

So far no full scale application of the anaerobic expanded/fluidized bed process is being reported on sanitary wastewater.

Industrial Applications

For the treatment of industrial wastewater, Ecolotrol Inc., of U.S.A. has operated a number of pilot scale anaerobic fluidized bed reactors in order to establish the desingn parameters (Jeris, 1982)

Results of a pilot study conducted by Ecolotrol Inc. U.S.A. (Jeris 1982) for a food processing waste is given in Table 4.7. Temperatures were maintained at 35-37°C. For the four organic loading rates used, the BOD5 and COD removals varied between 86 to 93 and 75 to 86 percent respectively showing a relatively low spread of removal efficiency over the organic load range studied. Further evidence of the stability of this operation are the volatile acids which rarely exceeded 200 mg/L as acetic acid. Also, the suspended solids removal averaged close to 50% with an influent concentration of 1,140 mg/L and an effluent of 550 mg/L.

  Table 4.7  Anaerobic fluidized bed reactor treatment - food processing waste (Jeris, 1982)

The percentage COD reduction for five waste categories at different organic loading are summarized in Fig 4.5 (Jeris, 1982). It is evident that the removal efficiency in slightly affected upto 16 kg COD/m3-day for most wastes and that the efficiency remained high for some of these wastes to loadings in excess of 36 kg COD/m3-day.

Numerous pilot-scale tests have shown high removal percentages (80%) at loading rates ranging from 10 to 20 kg'm3-day for a variety of industrial wastes. The results are summarized in Table 4.8.

The COD removal efficiency of whey waste at different loading rates and at two temperatures were studied (24°C and 35°C). About 10% reduction in removal efficiency was caused by the temperature difference (Fig.4.6).

Greater efficiencies were reported for two stage application of fluidized bed reactors in series and the results for different types of waste are given in Table 4.8. In single stage application of influent COD in excess of 50,000 mg/L and for 80% removal, an effluent COD of 10,000 mg/L was reported, where as in the two stage application an influent COD of 52,200 mg/L was reduced to 3,200 mg/L to give an overall COD removal of 94%. A five day retention period was required.

Only two full scale applications of the anaerobic expanded/fluidized bed process for treating industrial wastes have been reported (Switzenbaum, 1983) (Table 4.9).

Extensive literature on denifrification in anaerobic fluidized and expanded beds are available but as yet there is little full scale operating experience. All the work that has been reported on anaerobic denitrifying fluidized bed systems are summarized by Cooper & Wheeldon (1980) and is presented in Table 4.10.

Applicability

Shock loadings (in terms of temperature and loading strength) had relatively little influence on the process (Jewell et.al., 1981). Expanded bed reactor was found to be efficient for the treatment of particulate wastes (Morris & Jewell, 1982)

           Table 4.9. Full-scale anaerobic fluidized bed installation (Switzenbaum, 1983)

The anerobic expanded bed process has been demonstrated to be effective for treating low strength wastes (COD less than 600 mg/L) at short retention times (several hours) and at high organic loading rates (up to 8 kg COD/m3-day) even at low temperatures (10°C. 20°C) (Switzenbaum & Jewell, 1980).

Due to the relative insensitivity of the process performance to hydraulic retention time, the system could be designed at a low HRT (on the order of several hours). The actual HRT will depend on wastewater organic strength. Little gas is produced when treating low strength wastes and as a result anaerobic treatment of sanitary wastewater cannot be regarded as a large energy producer.

However, for high strength industrial wastes (COD range 5,000-50,000 mg/L) BOD and COD removal varies between 60-95 and 65-85 percent respectively in 0.3 to 4.9 days HRT over a wide range of organic loading (4-25 kg COD/m³-day) with a considerable amount of methane production.

The experiments carried out on anaerobic fluidized beds by Jewell et.al.(1981) to study the effect of shock loads, showed that the process was unaffected by large instantaneous fluctuations in temperature, flow rate, organic concentrations, and organic loading rate.

However, the optimum temperature for the treatment is around 35° C and for lower temperatures, reduced removal efficiencies are observed. The energy produced could be used to heat the influent wastewater depending on tis temperature. Typically 0.4 lieters CH₄ were produced per gram of COD removed at 35° C. Methane content averages approximately 70% of the biogas with a range between 65- and 75% (Table 4.11)

               Table 4.11.     Summary of methane produced  (Jeris, 1982)

     *  Based on ultimate BOD, 35 days

Expanded/fluidized beds are able to withstand severe hydraulic overloading. They are less suited for organic overloading compared to fixed film reactors (Henz & Harremoes, 1982).

Laboratory scale experiments carried out by Norrman (1982) for the treatment of black liquor evaporator condensate from a kraft mill, showed that the process in the expanded bed was upset after exposure to the toxic wastes while the fluidized bed showed no signs of toxicity influence even at high volumetric loads. The fixed bed reactor survived the toxicity effects. Generally, expanded/fluidized beds are believed to have a relatively poor ability to withstand the effect of toxic compounds and other environmental factors compared to fixed bed reactors. ( Henz & Harremoes. 1982).

Results of a few pilot-plant scale experiments conducted indicate the feasibility of using expanded/fluidized bed reactor for denitrification.

Problem Associated with Expanded/Fluidized Bed Reactors

  Non- attached Biomass

In fluidized bed reactors, the high vertical velocity will take the free swimming organisms with the flow and deteriorate the effluent quality (Henz & Harremoes, 1982).

  Suspended Organics

Expanded and fluidized beds have their biomass structure damaged by significant amounts of suspended organics in the influent (Henz & Harremoes, 1982)

  Foaming

Foaming problems were reported in expanded/fluidized beds (Henz & Harremoes, 1982)/

  Gas bubbles

Problems with gas bubbles are frequently met with in expanded/fluidized beds as in fixed beds and sludge blanket reactors. Gas bubbles may adhere to flocs/bed particles and cause these to rise in the reactor, and may result in wash-out of biomass or deterioration of the effluent quality (Henz & Harremoes, 1982).

  Start-up

Although all anaerobic reactors have start-up problems, expanded and fluidized bed reactors are believed to be the most troublesome in this aspect. (Henz & Harremoes, 1982).

Advantages and Disadvantages

   Advantages

All the advantages claimed for the anaerobic expanded/fluidized bed reactors are derived directly or indirectly from the high concentration of biomass. Generally 10-4- kg/m³ of volatile solids loading can be achieved (Switzenbaum, 1983).

Expanded and fluidized beds have several important advantages over anaerobic filters. These include the following (Switzenbaum 1982)::

• no danger of clogging
• small headloss
• easier removal or addition of active material
• better hydraulic circulation (avoidance of short circuiting)
• ability to operate al lower required detention times and/or higher organic loading rates,
• greater surface area available per unit of reactor volume
• greater efficiency; and
• better mitigation against toxic shocks and dilution of high strength wastes.

Further, Cooper and Wheeldon (1980) have claimed the following advantages for biological fluidized bed systems:

• the high biomass concentration achievable leads to a small plant design which could reduce the land area requirement by up to 80 per cent; and
• a large reduction of capital cost is achieved due to the gratly reduced reactor volumes.

Disadvantages

One disadvantage of this system is that recycling of effluent may be necessary to achieve bed expansion, and the system is more complex.

Conclusion

An anaerobic fluidized  bed may be a good pretreatment process for sanitary wastewater, followed by some post-treatment process like aerobic treatment.

Numerous pilot tests on anaerobic expanded/fluidized beds have shown high removal efficiencies ate greater loading reates and low retention times for a variety of industrial wastes. For high stregth wastes, a considerable amount of methane is produced.

Comparison of high-rate anaerobic wastewater treatment reactors

The four reactors reviewed in this book are similar in that each is a means of maintaining the active biomass independently of the hydraulic retention time in the fermentation process. Thus high biomass concentrations, for systems stability and high efficiency, are achieved with low HRT, which in necessary for system economy. Each represents an advancement of conventional anaerobic digestion technology.

As reported by Lettinga et.al. (1984), so far no comparison study have been made at relevant scales between the various high-rate processes, although such a study is in progress at the Waste Water Technology Centre, Burlington, Canada (Hall, 1981) for a UASB process, an anaerobic stationary fixed film reactor, an anaerobic filter and a small-scale fluidized bed reactor.

The main factors determining the hold-up of viable biomass of various high-rate systems, as compiled by Lettinga et al.(1984) is given in Table 6.1.

Other important features of the various high-rate anaerobic wastewater treatment systems, such as the rate of start up, the capacity of the process to remove suspended solids, the risk of clogging, the need for effluent recycle, the installation of a sophisticated feed inlet distribution system, a gas-solids separator (GSS) and the use of packing materials are listed in Table 6.2.