design of a lamella settler for biomass recycling in continuous ethanol fermentation process

Design of a Lamella Settler for Biomass Recycling in Continuous Ethanol Fermentation Process

J. Tabera* and M.A. lznaola Instituto de Fermentaciones Industriales (C.S.I. C.) Juan de la Cierva, 3, 28006 Madrid, Spain

Accepted for publication July 18, 1988

The design and application of a settler to a continuous fermentation process with yeast recycle were studied. The compact lamella-type settler was chosen to avoid large volumes associated with conventional settling tanks. A rationale of the design method is covered. The sedimentation area was determined by classical batch settling rate tests and sedimentation capacity calculation. Limitations on the residence time of the microorganisms in the settler, rather than sludge thickening considerations, was the approach employed for volume calculation. Fermentation rate tests with yeast after different sedimentation periods were carried out to define a suitable residence time. Continuous cell recycle fermentation runs, performed with the old and new sedimentation devices, show that lamella settler improves biomass recycling efficiency, being the process able to operate a t higher sugar concentrations and faster dilution rates.

INTRODUCTION

Biomass recycle is a technique often used to overcome restrictions imposed to continuous alcoholic fermentation by inhibitory effect of ethanol. By cell recycle, higher concentrations inside the fermentor, productivities, and conver- sion efficiencies of ethanol can be obtained when compared with the simple continuous fermentation technique. ’

Various cell-liquid separation procedures have been reported: centrifugation: f i l trati~n,~ and gravity ~et t l ing.~ Among these, gravity settling is advantageous from an en- ergetic and maintenance point of view. The major draw- back of this method is the large size of sedimentation tanks needed to harvest efficiently the microorganisms from the fermentation broth.’ To circumvent this problem, biomass recycle by sedimentation is usually performed with floccu- lating microorganisms. Microbial flocs, with greater settling rates than individual cells, can be adequately harvested in smaller sedimentation devices. On the other hand, keeping in mind that the basic efficiency-determining factor in sedimentation is area, a well-designed, high surface- volume ratio settler may enhance the whole fermentation process performance.

A continuous process of alcoholic fermentation from beet molasses, with yeast recycle by sedimentation, has been

* To whom all correspondence should be addressed.

developed in the Instituto de Fermentaciones Industriales (IFI), CSIC.6 The aim of this article is to describe the design of a lamella settler to be used in our fermentation process and the profits derived from its utilization.

THEORY

When a particulate suspension is allowed to rest, the gravity force leads the particles to settle in several ways depending on the suspension characteristics. If solids concentration is very low, each particle moves separately to each other. Their movement is exclusively due to gravita- tional force and to the buoyancy applied. Constant falling terminal velocity can be described according to theoretical equations (Stokes’ law, Newton’s law, etc.) for this “free sedimentation.” In concentrated suspensions interaction between neighboring particles or aggregates does exist; then, a block, plug-flow “hindered settling” takes place. A clarified liquid-suspension interface becomes sharply defined, and other progressively less distinguishable concentration zones can also be seen. The settling rate of this kind of suspension is a function of the solids concentration, and their sedimentation profiles must be described from laboratory-scale tests. Once the particles are settled at the bottom of the tank, a process of thickening (also named compression) starts that can be described by the following first-order kinetics’:

A certain sludge height (a certain sludge concentration) will be achieved after a compression time t . This time can be calculated from experimentally determined h,, h,, and k.

Design of Continuous Settler

A continuous settler is a basin fed by a suspension, from which a clarified liquid and a thickened sludge must be obtained. Therefore, the settler must accomplish the following functions: (1) efficient retention of the solids from the suspension feeding and (2) suitable thickening of the sludge.

Clarification efficiency depends on the existence of a zone with very low solids contents in the upper part of the

Biotechnology and Bioengineering, Vol. 33, Pp. 1296-1305 (1989) 0 1989 John Wiley & Sons, Inc. CCC 0006-3592/89101001296-010$04.00

settling tank. Particles will settle in this free sedimentation zone if their terminal sedimentation rate is greater than the liquid feed ascending velocity. This condition will be ac- complished if the tank has enough horizontal area, the volume having no influence. The thickening of the settled sludge depends on its residence time in the tank (the volume being the controlling factor) as well as on the sludge bed thickness.

The evaluation of the horizontal area and the volume of a continuous settler can be supported on batch sedimentation tests because batch and continuous sedimentation pro- cesses have two points in common: (1) the relationship between the falling velocity of the solids and their concentration, v = v(C), and (2) the relationship between solids concentration and time, C = C(t).’

Relationships v = 4 C ) and C = C(t ) Establishment

According to Kynch,* if we allow a concentrated suspension to settle in a flask, an ideal, horizontal layer with fixed values of solids concentration C and sedimentation rate v starts to appear at the bottom of the flask and will ascend to maintain C and v constant. When this layer meets the descending clarified liquid-suspension interface, all the solids in the flask have crossed throughout it. The ascensional velocity u of the ideal layer can be demon- strated to be constant.

If we assume A to be the cross-sectional area of the flask and Co and ho the initial solids concentration and initial height of the interface, respectively, the total mass of solids in the flask, m,, will be

m, = AhoCo ( 2 )

On the other hand, if the considered ascending layer reaches the interface level at height h and time t , we obtain

(3)

u = h/t (4)

AhoCo = A(v + u)rC

where

Solving for h from the preceding equations,

This expression describes the variation of the interface height h with time t in a batch sedimentation run and can be obtained experimentally. For each experimental curve point a tangent line can be drawn; from its intercept the solids concentration of the interface at time t can be derived, and its slope accounts for their settling rate at drfined concentration C. Thus, the relationships C = C(t) and v = v(C) can be established.

Determining Area of Settler

Figure 1 shows the flow and concentration pattern in a continuous settler. At steady state, a solids mass balance

1 QR C~

Figure 1. mow diagram in continuous settler.

leads to

Qo Co = Qc = QR CR (6)

provided there are no solids in the clarified effluent. To avoid overflow of solids, each horizontal layer of the

settler must be able to accept the descending solids flow; that is, the falling velocity of the solids must at least be equal to the ascending velocity of the clarified liquid,

(7)

The maximum amount of solids entering a level by unit area and unit time is called the sedimentation capacity G of that level. This magnitude depends on the concentration and velocity reached by the solids at the considered level. The smaller the sedimentation capacity of a level, the larger the area needed to meet the condition in equation (7).

The area of the settler must obviously be calculated from the minimum sedimentation capacity level, that is, from the maximum needed area level. By combining equations (6) and (7),

and according to the sedimentation capacity definition,

(9)

Since the relationship v = v(C) is known, by defining the desired sludge concentration CR , the sedimentation capacity can be calculated for each C value. The required sedimentation area A must be determined from the minimal G value.

Determining Volume of Settler

As previously outlined, the residence time of the solids in the settling tank controls the sludge thickening, only guaranteeing equation (9) no solids overflowing when sludge concentration C, is attained. Due to the exponential nature of equation (l), high thickening requires very high residence times. When handling biological suspensions, the residence time of the microorganisms apart from their cultivation conditions becomes the more important factor.

TABERA AND IZNAOLA: DESIGN OF LAMELLA SETTLER FOR BIOMASS RECYCLING 1297

So, the maximum residence time must be established based on biological considerations, and from this value, the volume of the settler must be calculated.

A well-designed continuous settler is divided into two sections by the feeding level: the volume occupied by the clarified liquid, with no solids, and the volume occupied by the suspension, where all the solids in the settler are confined. This second volume can be expressed as

v, = m , / c , (10)

assuming a homogeneous suspension concentration and constant total solids mass at the steady state. The residence time of the solids in the settler, 7, can be considered as the necessary time to perform the total solids contents turn- over. Taking into account the mass balance of equation (6),

(11)

On the other hand, although the suspension homogene- ity condition is not actually met, its concentration would be the same as the one obtained if the sediment, always confined in the volume V,, was mixed. Because the same relationship C = C(t ) is valid for the flask as for the continuous settler, the problem can be solved by determining the mean concentration of the sediment at time T:

(12)

In this way, the volume occupied by the suspension can be determined. The total volume of the settler can be obtained by addition of the clarification zone volume, with no influence on the free sedimentation. Thus, providing that a correct surface calculation to avoid solids overflow has been made, this clarification volume can be freely adjusted having in mind design security criteria.

m, = TQR CR = 7Q0 C,

C, = C, = C&,,fh,

Basis of Lamella Settler

When free sedimentation conditions are achieved, one can take advantage of the fact that area is the main design parameter. Let us consider an ideal, horizontally fed, continuous sedimentation tank (Fig. 2).9 If we dispose in it n - 1 parallel plates, we obtain n sedimentation compart- ments (lamellae) with the same horizontal surface; thus, the same tank is able to handle an n-fold greater incoming

Effective area - A

flow. To facilitate the drainage of the sludge, the tank must be inclined, the horizontal projection of the total area of the plates being the resulting effective sedimentation area:

A = A,n cos (Y (13)

As can be seen from Figure 2, compactness is the main advantage of the lamella settler. Furthermore, the plates act as deflecting devices, avoiding turbulences and “short- circuits”, which are important considerations in the design of conventional continuous settlers.

There are no special criteria to choose the number and shape of the plates, but three essential conditions must be met by the lamella width (separation between adjacent plates) relating to the countercurrent flow pattern shown in Figure 3:

1. Laminarjlow condition: Values of the Reynolds number ranging from 500 to 2000 have been described as suitable for lamella settling. However, values as low as Re = 350 have been reported as needed for certain suspensions”:

Re = v f D h / y 6 350 ( 14)

where D, is the hydraulic diameter for a rectangular cross section defined” as 2ab/(a + b). According to Figure 3,

Figure 3. Flow diagram in lamella.

Effective area = n A Effective area = n A C O S -

Figure 2. Operational principle of lamella settler.

1298 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 33, APRIL 1989

(15) 2s sin a

1 + (S sin a ) / B - - 2BS sin a

B + S sin a D, =

The ascending fluid velocity vf is

vr = q / B S sin a (16)

where q = Qo/n is the flow rate in a lamella. Equations (14)-( 16) lead to

s 350 (17) 29 y(B + S sin a)

Re =

2 . Flow stability condition: The Froude number” must be greater than lo-’ to assure stable flow”:

3 10-~ (18) v3 gD, 2gB2S3 sin’a

q2(B + s sin a) F r = - =

3. Critical fluid velocity condition: The ascending fluid velocity VJ must be less than a certain critical value from which settled particles could be resuspended. This value (5 X m/s) has been experimentally determined for our sludge (see experimental for settler design):

( 19)

Equations (17) and (18) establish maximal values for the lamella width, while equation (19) establishes a minimal

vr = q/BS sin a s 5 x

81

i B

/

value for it. Within this range, the separation of the plates can be chosen, as well as their number and dimensions, to obtain a suitable size and shape of the sedimentation tank.

MATERIALS AND METHODS

IF1 Fermentation Continuous Process

Microorganisms and Culture Media

Sacchuromyces ellipsoideus 159 strain of the IF1 collec- tion, selected for its fermentation rate and flocculation ability, l 3 was employed.

The fermentation medium was composed of beet molasses diluted to 120 g/L sucrose with 0.5 g/L of added (NH,),HPO, and pH adjusted to 4.5 with H2S04.

Fermentation Cell Recycle System

Figure 4a shows the primitive, two-branch, continuously fed fermentor. The B1 branch, of 18 L working volume, is a column fermentor with hydrodynamic mixing by an ex- ternal pump, while the B2 branch acts as a sedimentation device. The coupling between both branches must assure the B2 branch to be a quiet zone. Hence, this coupling has

Figure 4. IFI fermentation equipment: (a) two-branch fermentor; (b) fermentor with lamella settler.

1299 TABERA AND IZNAOLA: DESIGN OF LAMELLA SETTLER FOR BIOMASS RECYCLING

been made in a stirring restricted area. Yeast flocs sedi- mented at the bottom of B2 were reentered in B 1 by means of a dosifying pump; so, cell recycle was performed.

Within the course of the experiment, the need for improved yeast recycle efficiency was pointed out. There- fore, the replacement of the sedimentation branch by a better settling unit (Fig. 4b) was considered. The new lamella settler was coupled with the fermentation branch at the same place as B2 was done; thus, the advantages of restricted stirring area were maintained. Due to the enhanced clarifying efficiency of the lamella settler and in order to prevent culture aging, the harvested yeast was partially pumped out. l4

Experimental for Settler Design

According to principles covered in the theory section, batch sedimentation assays were carried out to determine C = C ( t ) and v = v ( C ) relationships and, therefore, the needed sedimentation area. Critical ascending fluid velocity was also determined with our yeast. To establish the residence time of the solids in the settler, necessary for volume calculation, we considered that fermentation rate was the best indicator of the influence of that time on the physiological state of the microorganisms. Thus, fermentation rate measurements of yeast maintained in sedimentation for different time intervals were performed.

Sedimentation Tests

A cylindrical graduated glass tube 0.08 m diameter and 1.00 m length was filled with yeast suspension up to a 0.80-m-height level. After agitation, the tube was set in a vertical position in a quiet place and suspension was allowed to settle. At this moment the time count was started. Every time the clarified liquid-suspension interface falls to 0.02 m, the time was counted.

The yeast suspension samples were taken from the two- branch fermentor at the place where the settler would be attached (see Fig. 4a, b) because in this zone of restricted stirring a little sedimentation was detected and the settling behavior of the suspension could be different from that of the stirred culture. Six different suspension samples were tested. The yeast concentration was 45 kg/m3 (d.w.).

Critical Fluid Velocity Determination

The yeast suspension was allowed to settle for 45 min in the glass tube used for sedimentation tests and inclined 55". Clean water was pumped up slowly until the clarified liquid was colorless. The maximum throughput was determined by increasing the feed flow rate until the overflow became cloudy, indicating resuspension of particles.

The very small lamella width [see eq. (26)] obtained from this critical velocity measurement is in the range of practical limitations due to solids clogging. Therefore, no more precise determinations of critical ascending fluid velocity were considered.

Fermentation Rate Tests

Yeast cuIture from the fermentor was allowed to settle in a flask. After 0, 1, 2, and 24 h, samples were taken from the bottom of the flask. To adjust the sizes of the inocula, dry mass determinations were carried out in the biomass.

Glass vessels of lo00 mL capacity were filled with 500 mL of the medium used in the fermentating plant, steam steril- ized, inoculated with 1 g/L (d.w.) yeast to be tested, closed with Mueller valves, weighed, and maintained at 30°C with- out stimng. Measurements of fermentation development by weighing the glass vessels were made at fixed periods of time. The difference between the weight obtained every time and the weight of the initial inoculated glass vessel was the lost CO,, which is proportional to the alcohol concentration reached in the medium.

RESULTS AND DISCUSSION

Settling Area Determination

Mean values of the clarified liquid-suspension interface height, obtained from the six sedimentation tests, has been plotted vs. sedimentation time in Figure 5 . As previously illustrated [eq. (5)J from slopes and intercepts of tangent lines to this curve, we can obtain the values of C and v for each sedimentation time. The results of such a procedure are listed in columns 2, 3 and 4 in Table I.

To determine the sedimentation area, it is necessary to define the values Q,, and C, to be included in the sedimentation capacity expression [eq. (9>]. As previously stated for our process,6 to obtain 90% recycled biomass and a double yeast concentration in the sludge than in the broth, Q, = 4.264 x m3/s and C, = 120 kg/m-3 must be fixed. It must be remarked that these values have been derived from extreme conditions, that is, maximal dilution rate and maximal yeast concentration. Of course, these conditions have not simultaneously happened; so, settling area determination has been based on a conserva- tive approach.

Values of sedimentation capacity calculated from equation (9), being C, C,, and v known, are included in the last column of Table I. A minimal value, G = 2.384 x

kg/m3 s, corresponding to a solids concentration C = 75.789 kg/m3 and a settling rate v = 1.159 X m/s, accounts for the maximal solids entrance at a level of 0.32 m height. This level controls the effective area of the settler, also calculated from equation (9):

A = Q,C,/G = 8.05 X m2 (20)

which corresponds to an actual plate surface of 1.4035 x lo-' m2 if a 55" inclination angle is utilized.'

Lamella Pack Design

To select the separation between plates, equations (17)- (19) must be applied. Calculations are greatly facilitate if some approximations are assumed. Keeping in mind that the dimensions of the plates are usually much greater than


0 5-

0 4-

0 3- 0

/ 0 0 7 I I I I I 1 I 1 1 1

0 2 4 6 8 10 12 14 16 18 20

& 0 0 , I I I I I 1 I 1 1 1

0 2 4 6 8 10 12 14 16 18 20

Sedimentation time x ( 8 )

Figure 5. Experimental interface falling curve for 45 kg/m3 Succharomyces ellipsoideus IF1 159 suspension.

Table I. Sedimentation capacity calculation from experimental h vs. t plot (Fig. 5 ) .

Interface Settling Solids Sedimentation height, Intercept, rate, concentration, capacity, h (m) Coho/C (m) v ( x 10’) (m/s) C (kg/m’) G ( x 10’) (kg/m’ s)

~

0.800 0.580 0.480 0.460 0.440 0.420 0.400 0.380 0.360 0.340 0.320 0.310 0.300 0.290 0.280

0.800 0.770 0.750 0.710 0.690 0.655 0.605 0.590 0.565 0.520 0.475 0.435 0.405 0.360 0.325

5.800 5.233 4.750 4.283 3.900 3.417 2.600 2.400 2.233 1.583 1.159 0.900 0.733 0.450 0.267

45.000 46.753 48.000 50.704 52.174 54.962 59.504 61.017 63.717 69.231 75.789 82.759 88.889

100.000 110.769

4.176 4.008 3.800 3.761 3.600 3.456 3.069 2.979 3.034 2.590 2.384 2.400 2.513 2.700 3.845

the separation between them, not much error is involved in neglecting S sin a /B against 1 in the lamella hydraulic diameter expression:

2s sin a 1 + (S sin a) /B D, = = 2s sin a

Equations (17) and (18) then reduce to

Re = 2q/yB S 350

Fr = q2/2gB2S3 sin3a 2 lo-’

(22)

(23)

On the other hand, if we assume the kinematic viscosity of the fluid to be that of the water (y = W6 m’/s), from equation (22) we obtain

q/B < 1.75 x (24)

which, substituted in equation (23) with a = 55”, leads to

S s 6.57 X m (25)

and, by substituting equation (24) in (19),

S 3 0.43 X lo-’ m

Therefore, the separation between the plates can be adjusted to any value within the range 0.43 X 10-2-6.57 X

m. Note that if no assumptions were made, an enlarged range for separation would be obtained. Hence, the simplified calculation leads to a more restrictive design.


60

50

- - \ 0 I

0 z 40

0 0

c 0

0 0

.-

n L

c

0

30

20

10

0 0 10 20 30 40 so 60 70

Fermentation time ( h ) Figure 6. ent sedimentation periods.

Fermentation curves of Saccharomyces ellipsoideus IFI 159 after differ-

The obtained 1.4035 X lo-’ m2 plate area must be divided into n plates, each of A, = BL surface, having no precise criteria to choose neither their number nor their dimensions. Nevertheless, most of the commercial lamella settlers use plates with a 1.5-2.5 length-width ratio. Also, lamella widths of 2-5 cm, values included in our range, are very often employed in small settlers.

pack of 18.00 X lo-* m length, 10.81 x lo-’ m width, 17.72 X lo-’ m height, and 3.448 X m3 volume can be made. Note that a conventional settling tank with the same effective area and the same height would have a volume of 14.264 X m3, more than 4 times that of the lamella pack.

By establishing a length-width plate ratio LIB = 2 and a lamella width S = 3 x lo-’ m, in addition to the set-

Suspension Determination

tling area imposed by the sedimentation capacity [eq. (20)], we have a set of constraints to be satisfied by the number and size of the plates.

To obtain a lamella pack of suitable shape from the several pairs of values A,, n satisfying the above constraints, the following have been selected: n = 6 plates (the lower wall of the container included), L = 21.63 X lo-* m and B = 10.81 x lo-’ m, from which a lamella

Fermentation curves of S. ellipsoideus IFI 159 strain after 0, 1, 2, and 24 h sedimentation are shown in Figure 6. Because we want to evaluate the delay in fermentation activity due to spent time outside the fermentor, the initial slope of these curves is of major concern. As can be seen, this slope decreases as the sedimentation period increases. The maximal relative difference is attained between 0 and 1 h sedimentation, which can be interpreted as the fastest

1302 BIOTECHNOLOGY A N D BIOENGINEERING, VOL. 33, APRIL 1989

loss of fermentation activity is produced in the first hour. Therefore, in an attempt to find a compromise between fermentation activity and thickening, we have chosen 45 min as a suitable settling residence time. Entering 7 = 2700 s in the interface falling ? m e (Fig. 5), h, = 63.5 X

lo-’ m is obtained, and from equations (lo)-( 12), the volume to be occupied by the suspension is 9.137 X m3. The total volume of the settler will be obtained by addition of the lamella pack volume.

The approach utilized in deriving equations (lo)-( 12) for the suspension volume determination is rigorous only in a conventional continuous settler. If the settler is well designed, the absence of solids above the feeding level is a good assumption. This is not the case for the lamella settler. Because of the countercurrent flow pattern, the lamellar space (above the feeding level) is partially occupied by the suspension undergoing thickening within it. Analysis of the fluid mechanics of the lamella sedimenta-

makes it evident that the suspension will occupy a significant fraction of the lamella volume. These facts point to a possible oversizing of the thickening chamber and an unduly enlarged residence time of the yeast in the settler.

These considerations are of minor concern in our case. The calculated suspension volume for 7 = 3600 s is 10.36 x

m3, with an increase over the value for 7 = 2700 s of 1.22 X m3, equal to 35% of the lamella pack volume. Due to the flocculation ability of our yeast, promot- ing particle agglomeration, the portion occupied by the suspension was observed to be much less than 35% of the lamella pack volume. So, the added volume to that calculated is small and the residence time of the microorganisms in the settler was maintained near 2700 s and far from the 3600 s considered as the limiting time from the fermentation rate tests.

tion15,16

Constructive Details and Total Volume Determination

The lamella pack must be mounted in a container where suitable flow conditions can be achieved (Fig. 7). Uniform feed distribution between lamellae and low velocity of the liquid leaving the unit, to avoid resuspension of settled solids, are principal factors to be considered.” Influent comes into the settler by two opposite sides, contribut- ing to uniform feeding. The two supporting walls of the lamella pack force the suspension to come in laterally near the bottom of the plates and to upflow through the lamellae. Attached to the supporting walls two channels per- forms a slow flow recovery of the effluent.

These constructive details are volume consuming and have been absorbed in a 1.25 increasing volume security factor,’* leading to a settler volume of about 15 x m . The length of the sedimentation plates has been increased by the same factor, with a corresponding security margin in the sedimentation area. The volume to be occupied by the suspension has been maintained at its primitive value in order not to increase the residence time of the microorganisms in the settler. The bottom of the settler has a pyra- midal shape both to facilitate the sludge drainage and to achieve enough solids bed thickness.

3

Improvements in Cell Recycle and Fermentation

Results of steady-state continuous fermentation runs, obtained with the old and new fermentating systems oper- ating at the same conditions, are compared in Table 11. Values for ethanol concentration in beer, P, yeast concentration inside the fermentor, X, ethanol-substrate yield, Y, and ethanol productivity, PD, are included.

1303

Table 11. fermentating plants.“

Continuous ethanol fermentation with cell recycle runs performed by the two IF1

Dilution rate D (h-’) 0.19, 0.30 0.43 Substrate concentration SO (g/L) 140 120 120

F2Bb FLS‘ F2B FLS F2B. FLS

Ethanol concentration f, g/L 56.6 64.6 55.7 56.2 - 53.0

Ethanol-substrate yieldd Y, % 74.7 86.9 84.9 89.0 - 84.1 Productivity‘ PD, g/L h 10.8 12.3 16.7 16.9 - 22.8

Yeast concentration X, g/L 37.0 35.8 53.0 60.0 - 57.4

a T = 30“C, pH 4.5. Fermentor with two branches. Fermentor plus lamella settler. Percentage of stoichiometnc 0.538 ethanol-sucrose yield. Referred to stirred volume.

At the dilution rate 0.30 h-’ and initial substrate concentration 120 g/L (the usual working conditions of the IF1 process), higher yeast concentration inside the fermentor and higher ethanol-substrate yield are achieved by the fermentor plus lamella settler system.

At D = 0.19 h-’ and So = 140 g/L, decreases in yeast concentration and ethanol substrate yield from the fermentor with two branches are registered; it is due to the decrease in substrate feeding. In the new fermentating plant the decrease in yeast concentration is due to limited feeding as well as to enhanced ethanol inhibition, the last point not so evident in the older design. This may explain why there is similar biomass concentration in both designs. Nevertheless, the better efficiency of the lamella settler improves the biomass utilization, with less amount of substrate spent to compensate the losses of cells in the effluent. This is the reason for achieving a higher ethanol-substrate yield than in the fermentor with the two branches. In both cases, for smaller values of ethanol productivity the low

c, D

* Dh Fr g G h ho h, h, k L ms n P PD 9 Q QL

PO QR Re S

mean solids concentration of sediment at time T, kg/m3 dilution rate, h-’ hydraulic diameter of lamella, m Froude number, dimensionless acceleration due to gravity, m/s’ sedimentation capacity, kg/m3 s interface height, m interface height at zero time, m interface height at infinite time, m interface height at time 7, m first-order kinetics constant, s - ’ length of plate, m total solids mass in settler, kg number of plates ethanol concentration in effluent beer, g/L ethanol productivity, g/L h flow rate in lamella, m3/s flow rate at arbitrary level, m3/s ascending flow rate of clarified liquid, m3/s incoming flow rate, m3/s sludge flow rate, m3/s Reynolds number, dimensionless lamella width, m

dilution rate is the major factor. SO initial substrate concentration, g/L f

v

v, v, settler volume occupied by x Y ethanol-substrate yield, %

The fermentor with lamella settler working at 0.43 h-’ dilution rate yields better results than the old fermentor at D = 0.30 h-’ in spite of the higher kinetic exigencies imposed by the former dilution rate value. When these results are compared with others of the new fermentor, good yeast concentration and yield values are revealed as well as a

sedimentation or compression time, s ascending velocity Of KYnch’s ideal layer, m/S

falling velocity of solids at interface, m/s ascending fluid velocity in lamellae, m/s

m3

yeast concentration inside fermentor, g/L

remarkably high ethanol productivity. a inclination angle of plates, degrees

All the advantages attained with the new fermenting design have been interpreted by us to be a consequence of

y kinematic viscosity of fluid, m’/s residence time of microorganisms in settler,

the enhanced separation efficiency of the lamella settler, which provides improved cell recycle. References

NOMENCLATURE

length of rectangular cross section, m height of rectangular cross section, m horizontal settling area, rn’ area of plate, m2 plate width, m solids concentration at interface, kg/m’ solids concentration of influent suspension, kg/m3‘ solids concentration of sludge, kg/m3 homogeneous solids concentration of settler suspension zone, kg/m3

1. G. E. Guidoboni, Enz. Microb. Technol., 6 , 194 (1984). 2. B. Maiorella, C R . Wilke, and H. W. Blanch, Adv. Biochem. Eng. ,

3. A. Margaritis and C . W. Wilke, Biotechnol. Bioeng., 20, 709 (1978). 4. C. B. Netto, A. Destruhaut, and G . Goma, Biotechnol. Letr, 7 , 355

5. T. J . Walsh and H. R. Bungay, Biotechnol. Bioeng.. 21, 1081 (1979). 6. J . Tabera, Ph.D. Thesis, Facultad de Ciencias Quimicas, Universidad

7. J . M. Coulson and J. F. Richardson, Chemical Engineering, 3rd ed.,

8. G . J. Kynch, Trans. Faraduy SOC., 48, 166 (1952). 9 . G . R. Mace and R. Laks, Chem. Eng. Prog., 74, 77 (1978).

20, 43 (1981).

(1985).

Aut6noma de Madrid, 1986.

Vol. 2 (Pergamon, Oxford, 1978), p. 177.


10. Granges Engineering, Sedimentacidn a Lamelas Sistema GEWE

11. R. H. Perry and C. H. Chilton, Chemical Engineer’s Handbook,

12. J. M. Coulson and J. F. Richardson, Chemical Engineering 3rd ed.

13. J . Tabera, M. A. Iznaola, I. Schnabel, and J. Ganido, Biorechnol.

(Granges Engineering, Stockholm, 1973).

5th ed. (McGraw-Hill, New York, 1973), pp. 5-24.

Vol. 1 (Pergamon, Oxford, 1977), pp. 66-67.

Lett., 7, 437 (1985).

14. S. J. Pirt and W. M. Kurowski, J. Gen. Microbiol., 63, 357 (1970). 15. A. Acrivos and E. Herbolzheimer, J . Fluid Mech., 92, 435 (1979). 16. W. F. h u n g and R. F. Probstein, Ind. Eng. Chem. Proc. Des. Dev.,

17. B. Forsell and B. Hedstrom, J . Water Pollut. Control Fed., 47, 834

18. R. H. Perry and C. H. Chilton, Chemical Engineer’s Handbook,

22, 58 (1983).

(1975).

5th ed. (McGraw-Hill, New York, 1973), pp. 19-53.


design of a lamella settler for biomass recycling in continuous ethanol fermentation process

Documents