BIOTECHNOLOGY AND BIOENGINEERING, VOL. XIX, PAGES 43-53 (1977)
Effects of Quantitative Shock Loadings on the Constant Recycle Sludge Concentration Activated-Sludge Process Y. K. WONG and P. Y. YANG,* Division of Environmental Engineering, Asian Institute of Technology, Bangkok, Thailand
Summary The reliability of the process of Ramanathan and Gaudy (Biotechnol Bioeng., 13, 125 (1971)) for the completely mixed activated-sludge process holding the recycle cell concentration, X R , as a system constant with respect to step changes in
hydraulic retention time was investigated. The experiments were run a t initial dilution rates of t , +, f, and 4 hr-’ treating a soft drink bottling wastewater. The influent substrate concentration was maintained a t 1000 mg/liter chemical oxygen demand and the hydraulic recycle ratio a t 0.3. The recycle sludge concentration was maintained at about 7000 mg/liter. It was found that the system could accommodate hydraulic shock loads up to 200% positive changes and down to 50% negative changes without disruption of the effluent quality. Shorter retention time of the range studied, from 2 t o 8 hr, has the advantage of shorter response time with respect to the response of the concentration of biological solids in the reactor.
INTRODUCTION I n 1961 Herbert2 developed equations for cell recycle systems. Figure 1 is the flow scheme for a completely mixed activated-sludge process employing “controlled” cell feedback. In applying Herbert’s continuous-flow steady-state model, the concentration factor, c, and recirculation ratio, a, should be maintained as constant as possible. Ramanathan and Gaudy3 noted that due to the heterogeneity of the populations, there were wide fluctuations in the concentration of the biological solids in the “steady state.” Therefore, they made a computational analysis of the steady-state activated-sludge process with controlled sludge feedback and found that the process using constant X R as a system parameter for design and operation was less *Present address: Department of Agricultural Engineering, University of Hawaii at Manoa, Honolulu, Hawaii 96822. 43
@ 1977 by John Wiley & Sons, Inc.
WONG AND YANG
44
INFLUENT F tsi txi
rn 4
S X
v
(lto9F
7
F’SIXe
EFFLUENT WASTE SLUOGE
sensitive to high-dilution rates than Herbert’s model. This could also exert a steadying influence on the “steady-state” concentration of the biological solids in the reactor. In addition, a sludge reaeration and consistency tank was installed to ensure that X R was constant. It may be expected that any substrate carried over in the recycle sludge should be further reduced prior to recycle. Gaudy and Sriniva~araghavan~ conducted experiments to test the new operational process and found the process approached the steady state with heterogeneous populations more closely than did Herbert’s model, and the high degree of treatment efficiency predicted by the model was demonstrated experimentally. In 1975 Srinivasaraghavan and Gaudy5 tested the process experimentally by using sludge developed from different origins. The results indicated that the performance with respect t o X , S , and waste sludge Xw was fairly reproducible. They also suggested that the maintenance coefficient would be recommended when the prediction of X W was considered. Bonotan-Dura and YangG applied the process to treat an organic industrial wastewater and concluded that a “steady state” with respect t o reactor biological solids and effluent chemical oxygen demand (COD) could be attained by varying the dilution rates. The process has proved that it is very useful in the design and operation of activated sludge process. However, its stability under shock loadings (quantitative or/and qualitative) is still unknown. Storer and Gaudy7 observed “growth rate hysteresis” of heterogeneous populations in response to a threefold increase in influent organic concentration (synthetic glucose waste) and concluded that the specific growth rate of the heterogeneous populations did not respond instantaneously to the step changes in the influent substrate concentration as was assumed by the Monod equation. George and
QUANTITATIVE SHOCK LOADINGS
45
Gaudy*made a study of systematic step changes in the dilution rates of a completely mixed heterogeneous once-through system and reported that hydraulic shocks (with Si constant) constituting a n increase in D were more deleterious than a decrease in D. For a detention time of 8 hr, an increase in D of 100% could be successfully accommodated. Eckhoff and Jenkins9 found that the lower the steady growth rate prior to the shock, the better was the response to the shock. Gaudy'o reported that the initial growth rate of a heterogeneous culture was of significant importance to the successfulness of response to step changes in pH and temperature. Lower specific growth rates usually had better recovery power and shorter recovery time. I n the present study, a laboratory-scale pilot plant was set up to investigate the effect of quantitative shock loading, i.e., step increase and decrease of the dilution rate with a fixed influent substrate concentration, on a system where a constant recycle sludge concentration, XR, was maintained as a system constant in a completely mixed activated-sludge process.
MATERIALS AND METHODS The laboratory-scale pilot unit was exactly the same as the unit used by Bonotan-Dura and Yang.6 The flow diagram consisted of a feed reservoir, an aerator, a clarifier, and a homogenizing reaeration tank (see Fig. 2). The rectangular aeration tank had a n effective volume of 5.0 liters. Compressed air was introduced through six stone diffusers to provide complete mixing. Feed solution was pumped to the reactor from the feed reservoir (50 liters capacity) by a chemical feed pump (model C-630P, Engineered Products Mfg. Co., U.S.A.). To preserve the characteristics of the feed waste, a cooling coil was continuously dipped into the feed reservoir. The feed lines were sterilized periodically by a 1% chlorox solution and then rinsed by tap water for several minutes. The reactor effluent flowed by gravity into a 60" hopper-type clarifier with a capacity of 4.3 liters. Settled sludge was drawn a t 4 to 6-hr intervals and stored in the reaeration tank. Sludge diluted to 7000 mg/liter was added to the homogenizing tank a t 12-hr intervals for recycle. At times, the sludge withdrawn from the clarifier was below the desirable level for recycle; it was then concentrated by centrifugation and diluted to the desired level. The homogenizing tank (11.0 liters) was equipped with four stone diffusers. Sludge feedback was performed continuously by a variable speed peristatic action pump (Type
WONG AND YANG
46
^i" EFFLUENT
OF,XR
1 p 1 Rf-AmaTIoN
mw K)MwENUMG T W
! I I
Fig. 2. Flow diagram for a compfetely mixed reactor employing constant concentration of recycle cells, X R .
M H R E 7, Watson Marlow Ltd., England) through a silicone rubber tubing. The soft drink bottling combined wastewater (provided by nearby Pepsi Cola Company) was used in this work as the carbon source (growth-limiting substrate). Analysis showed that the waste had a COD of over 1000 mg/liter and was deficient in phosphorus and nitrogen. The low suspended solids content of the wastewater had the advantage that the organic content was mostly soluble. The raw waste was diluted to the desirable value (in this case, 1000 mg/ liter COD) and of the necessary amount of nitrogen and phosphorous added. Activated sludge was developed from the seeds obtained from the other activated sludge pilot unit treating pineapple canning waste. Microscopic observations revealed that the seeds were free of filamentous organisms. To start the continuous-flow operation, 1 liter of the sludge was added to the reactor and then diluted to 5 liters using tap water. The feed solution (about 1000 mg/liter COD, containing the appropriate amounts of ammonium sulfate and phosphate buffer solution, P H about 7.0) was pumped into the reactor a t the proposed rate. Sludge in the homogenizing tank was pumped into the reactor a t 0.3 times the value of the influent waste flow rate.
QUANTITATIVE SHOCK LOADINGS
47
The reactor was checked for complete mixing by measuring the biological solids and filtrate COD of samples taken from the middle and effluent of the reactor. When there was close agreement between corresponding parameters from both sampling points, the reactor was considered to be completely mixed. The COD was measured according to the Standard Method.” The concentration of biological solids was determined as dry weight by the membrane filter technique using glass fiber filter paper with 0.8-p pore size. The conditions of “steady state’’ were approached by measuring the filtrate COD in the reactor and the effluent at hourly intervals until both the level of solids and filtrate COD in the reactor reached some constant value. After a sampling period of a certain dilution rate, the influent flow rate was slowly altered (in one or two days) to the next predetermined initial flow condition of the successive run. The system was again allowed to go t o another steady state. Shock loadings were introduced by increasing the inflow rate 100 or ZOO’% of the initial value, or by decreasing the initial influent flow rate by 50y0. The schedule is summarized in Table I. TABLE I Summary of Parameters and Variables Initial Final Influent Hydraulic influent influent % recycle dilution rate dilution rate Change Run XR COD D (hr-l) D (hr-l) in D no. S, (mg/liter) (mg/liter) ratio, a 1 2 3 4 5 6
1/3
1/ 2 1 /4 1/2.65 1/1.5 114 1/ 8 1/2 1/1.33
lo00
7 8 9
+loo -50 +200 +3m 100 -50
+
++200 100
RESULTS AND DISCUSSION Operation Under Step-Hydraulic Shocks
Before the step-hydraulic shocks were applied, the initial steady state a t different dilution rates was developed. Nine experimental
WONG A N D YANG
48
e Y
'".I ILO
-
-_ m
RVI No6 lD=l/a-l/4hr-', 100%) No4 (D=l/O-l/ZVZ.65b-'.XX)XI
0 RVI
Irn -
2
8 Y L
noM-
=: . a
I , , , , , , , , , , , , , , , , , , , ,
runs were made in which the initial steady state was disrupted by applying a step change in influent flow rate only. The recycle sludge flow rate was altered correspondingly to keep the recycle flow ratio a at a constant of 0.3. I n Figures 3, 4, and 5, the disruption due to different degrees of shocks at each of the dilution rates, D = &,&,and hr-l, respectively, is compared. I n Figure 3, the initial value of D is hr-' and 100 and 2007& increases in D are shown. It can be seen that for the 100% step increase in D there is a rapid increase in cells and the substrate decreases. Then there is a lag of 2 hr and an increase in the concentration of biological solids occurs until the final steady level. The filtrate COD varied between 24 and 54 mg/liter during the transient period. On the other hand, there is a lag of 4 hr for the 200% step increase in D for the biological solid concentration in the reactor before it rises to a peak a t the ninth hour and then drops to the final steady-state level. During the transient period, the filtrate COD fluctuates from 40 to 80 mg/liter. The new steady-state value of filtrate COD is almost the same as the one before the shock was applied (about 40 mg/liter COD). The transient period takes about 14 hr in both cases. The difference in the modes of response can be explained by recalling that the microorganisms employed are heterogeneous in nature. The mass culture consists of microorganisms growing a t different specific growth rates. When D is changed from
49
QUANTITATIVE SHOCK LOADINGS
Fig. 4.
Comparison of loo%, 200%, and 300% changes in D for system at initial D of hr-1.
+
+
to $ hr-I (100% change), and to 1/2.65 hr-I (200% change), those microorganisms which can adapt to the new growth condition will be developed. Therefore, the lag period is needed before these species become the dominant populations.
a
I
0
RU(Y)S(D=c/4-c/t33k-',',%)
0
RUW NO 6 (D.1/4-!/2
b-', 1 0 0 % )
Rvw WO.7(o.ln-l/.
b-'.-w)%)
m
D
I4
m
T W .k
Fig. 5. Comparison of
loo%,
200%, and -50% changes in D for system at initial D of t hr-l.
50
WONG AND YANG
I n Figure 4,similar patterns of response are shown for initial D a t a t different levels of shock loading. It can be seen that the lag period for the biological solids concentration in the reactor takes 3 hr in the case of 100 and 200% increases of D while it takes 2 hr for a 3oOy0 increase of D. For the 3000/, increase of D, instead of increasing the biological solids concentration in the transient period, the concentration has been decreased to another steady state. For the filtrate COD concentration, there is not much change (between 35 and 70 mg/liter) following the 100 and 200y0 increases of D in the transient period, but there is a major change (between 40 and 140 mg/liter) following the 300y0 increase of D. The transient period for the case of 100 and 200% increases of D takes about 7 hr, and 15 hr for a 300% increase of D. I n Figure 5 shocks have been made a t initial D of hr-' a t a different level of hydraulic shocks. Unlike the previous two cases, no time lag for biological solids and COD concentration in the reactor is observed even when a 50% negative change in D is included. In general, the transient period of these three cases takes about 10 hr. I n Figures 6, 7, and 8 the same degree of hydraulic shocks and different initial dilution rates are compared. It has been found that using the constant XRprocess, a faster growing culture (higher initial
+ hr-l
a
Fig. 6. Comparison of hydraulic shocks (100% increase) a t different initial dilution rates.
QUANTITATIVE SHOCK LOADINGS
51
Fig. 7. Comparison of hydraulic shocks (200% increase) at different initial dilution rates.
dilution rate) responded to an increase in D better than a slower growing culture (lower initial dilution rate). When imposing transient conditions on a pure culture of s. cerevisiae fed ethanol by altering the change in hydraulic retention time, Mor and Fiechter12 observed that the time delay for the system to respond to a shock was dependent on the microbial net specific growth rate, p. A short time delay was recorded when the specific growth rate was high, and a long delay was recorded when the value of p was low. I n this study it is found that under constant X R operation, the system can accommodate a step increase in D up to 200% and a step decrease in D by 50% without much disruption of the effluent quality based on the COD measurement. In general, the response time for the new steady state to bc approached depends on the initial dilution rates. The higher t h r value of initial D , the shorter will be the response time. However, the initial D value should not be too high because the growth rate will bc too close to the maximum specific growth rate, p m . Any perturbations due to an increase in D will lead to cell washout and substrate leakage for the cultures which can no longer adapt to the new condition of change. I n the present study, it is found that an initial dilution rate of t hr-l is optimal to givv thc shortmt response time and it can prevent cell washout and substrat(>lcakagv \iit hiri t h(x present range of hydraulic shock loading.
52
WONG AND YANG
Fig. 8. Comparison of hydraulic shocks (50% decrewe) a t different initial dilution rates.
CONCLUSIONS On the basis of the findings in this study, the following conclusions are drawn. 1) With the initial dilution rates employed (D = $, $, hr-l), the system can accommodate a step increase in D up to 200% and a step decrease by 50% without disrupting the effluent quality. 2) The higher the value of initial dilution rates (e.g., 2 hr-l) the shorter the response time.
Nomenclature a c
D F S, S SR V X
X,
xi
xR
recirculation ratio; ratio of rate of underflow from the clarifier to the rate of inflowing waste concentration factor; the ratio of the biological solids in the clarifier underflow to that in the reactor F / V , dilution rate of the system (hr-1) rate of flow of incoming substrate (liter/hr) influent substrate concentration (mg/liter COD) steady-state concentration of substrate in the reactor (mg/liter COD) filtrate COD in the recycle sludge (mg/liter) effective volume of reactor (liter) steady-state biological solid concentration in the reactor (mg/liter) biological solid concentration in the effluent of the system (mgiliter) biological solid level in the influent waste b g / h t e r ) concentration of cells feedback (mdhter)
QUANTITATIVE SHOCK LOADINGS
53
XW sludge wastage (kg/day) p CL,
specific growth rate of microorganism (hr-l) maximum specific growth rate of microorganisms (hr-1)
This work was supported by a scholarship donated by the Government of Canada to the Asian Institute of Technology.
References 1. M. Ramanathan and A. F. Gaudy, Jr., Biotechnol. Bioeng., 13, 125 (1971). 2. D. Herbert, Chem. Znd. Monogr., 12, 21 (1961). 3. M. Ramanathan and A. F. Gaudy, Jr., Biotechnol. Bioeng., 11, 207 (1969). 4. A. F. Gaudy, Jr. and R. Srinivasaraghavan, Biotechnol. Bioeng., 16, 723 (1974). 5. R. Srinivasaraghavan and A. F. Gaudy, Jr., J. Water Pollut. Contr. Fed., 47, 1946 (1975). 6. F. M. Bonotan-Dura and P. Y. Yang, Biotechnol. Bioeng., 18, 145 (1976). 7. F. F. Storer and A. F. Gaudy, Jr., Environ. Sn’. Technol., 3, 143 (1969). 8. T. K. George and A. F. Gaudy, Jr., J. Environ. Eng. Div., Proc. ASCE, 99, 593 (1973). 9. D. W. Eckhoff and D. Jenkins, “Transient loading effects in the activated sludge process,” presented a t 3rd Int. Conf. Water Pollut. Res., Section 11, Paper 14, 1966. 10. A. F. Gaudy, Jr., BioteChnoZ. Bioeng., 17, 1051 (1975). 11. Standard Methods for the Examination of Water and Wastewater, 13th ed., American Public Health Association, New York, 1971. 12. J. R. Mor and A. Fiechter, Biotechnol. Bioeng., 10, 787 (1968).
Accepted for Publication August 20, 1976
Effects of Quantitative Shock Loadings on the Constant Recycle Sludge Concentration Activated-Sludge Process Y. K. WONG and P. Y. YANG,* Division of Environmental Engineering, Asian Institute of Technology, Bangkok, Thailand
Summary The reliability of the process of Ramanathan and Gaudy (Biotechnol Bioeng., 13, 125 (1971)) for the completely mixed activated-sludge process holding the recycle cell concentration, X R , as a system constant with respect to step changes in
hydraulic retention time was investigated. The experiments were run a t initial dilution rates of t , +, f, and 4 hr-’ treating a soft drink bottling wastewater. The influent substrate concentration was maintained a t 1000 mg/liter chemical oxygen demand and the hydraulic recycle ratio a t 0.3. The recycle sludge concentration was maintained at about 7000 mg/liter. It was found that the system could accommodate hydraulic shock loads up to 200% positive changes and down to 50% negative changes without disruption of the effluent quality. Shorter retention time of the range studied, from 2 t o 8 hr, has the advantage of shorter response time with respect to the response of the concentration of biological solids in the reactor.
INTRODUCTION I n 1961 Herbert2 developed equations for cell recycle systems. Figure 1 is the flow scheme for a completely mixed activated-sludge process employing “controlled” cell feedback. In applying Herbert’s continuous-flow steady-state model, the concentration factor, c, and recirculation ratio, a, should be maintained as constant as possible. Ramanathan and Gaudy3 noted that due to the heterogeneity of the populations, there were wide fluctuations in the concentration of the biological solids in the “steady state.” Therefore, they made a computational analysis of the steady-state activated-sludge process with controlled sludge feedback and found that the process using constant X R as a system parameter for design and operation was less *Present address: Department of Agricultural Engineering, University of Hawaii at Manoa, Honolulu, Hawaii 96822. 43
@ 1977 by John Wiley & Sons, Inc.
WONG AND YANG
44
INFLUENT F tsi txi
rn 4
S X
v
(lto9F
7
F’SIXe
EFFLUENT WASTE SLUOGE
sensitive to high-dilution rates than Herbert’s model. This could also exert a steadying influence on the “steady-state” concentration of the biological solids in the reactor. In addition, a sludge reaeration and consistency tank was installed to ensure that X R was constant. It may be expected that any substrate carried over in the recycle sludge should be further reduced prior to recycle. Gaudy and Sriniva~araghavan~ conducted experiments to test the new operational process and found the process approached the steady state with heterogeneous populations more closely than did Herbert’s model, and the high degree of treatment efficiency predicted by the model was demonstrated experimentally. In 1975 Srinivasaraghavan and Gaudy5 tested the process experimentally by using sludge developed from different origins. The results indicated that the performance with respect t o X , S , and waste sludge Xw was fairly reproducible. They also suggested that the maintenance coefficient would be recommended when the prediction of X W was considered. Bonotan-Dura and YangG applied the process to treat an organic industrial wastewater and concluded that a “steady state” with respect t o reactor biological solids and effluent chemical oxygen demand (COD) could be attained by varying the dilution rates. The process has proved that it is very useful in the design and operation of activated sludge process. However, its stability under shock loadings (quantitative or/and qualitative) is still unknown. Storer and Gaudy7 observed “growth rate hysteresis” of heterogeneous populations in response to a threefold increase in influent organic concentration (synthetic glucose waste) and concluded that the specific growth rate of the heterogeneous populations did not respond instantaneously to the step changes in the influent substrate concentration as was assumed by the Monod equation. George and
QUANTITATIVE SHOCK LOADINGS
45
Gaudy*made a study of systematic step changes in the dilution rates of a completely mixed heterogeneous once-through system and reported that hydraulic shocks (with Si constant) constituting a n increase in D were more deleterious than a decrease in D. For a detention time of 8 hr, an increase in D of 100% could be successfully accommodated. Eckhoff and Jenkins9 found that the lower the steady growth rate prior to the shock, the better was the response to the shock. Gaudy'o reported that the initial growth rate of a heterogeneous culture was of significant importance to the successfulness of response to step changes in pH and temperature. Lower specific growth rates usually had better recovery power and shorter recovery time. I n the present study, a laboratory-scale pilot plant was set up to investigate the effect of quantitative shock loading, i.e., step increase and decrease of the dilution rate with a fixed influent substrate concentration, on a system where a constant recycle sludge concentration, XR, was maintained as a system constant in a completely mixed activated-sludge process.
MATERIALS AND METHODS The laboratory-scale pilot unit was exactly the same as the unit used by Bonotan-Dura and Yang.6 The flow diagram consisted of a feed reservoir, an aerator, a clarifier, and a homogenizing reaeration tank (see Fig. 2). The rectangular aeration tank had a n effective volume of 5.0 liters. Compressed air was introduced through six stone diffusers to provide complete mixing. Feed solution was pumped to the reactor from the feed reservoir (50 liters capacity) by a chemical feed pump (model C-630P, Engineered Products Mfg. Co., U.S.A.). To preserve the characteristics of the feed waste, a cooling coil was continuously dipped into the feed reservoir. The feed lines were sterilized periodically by a 1% chlorox solution and then rinsed by tap water for several minutes. The reactor effluent flowed by gravity into a 60" hopper-type clarifier with a capacity of 4.3 liters. Settled sludge was drawn a t 4 to 6-hr intervals and stored in the reaeration tank. Sludge diluted to 7000 mg/liter was added to the homogenizing tank a t 12-hr intervals for recycle. At times, the sludge withdrawn from the clarifier was below the desirable level for recycle; it was then concentrated by centrifugation and diluted to the desired level. The homogenizing tank (11.0 liters) was equipped with four stone diffusers. Sludge feedback was performed continuously by a variable speed peristatic action pump (Type
WONG AND YANG
46
^i" EFFLUENT
OF,XR
1 p 1 Rf-AmaTIoN
mw K)MwENUMG T W
! I I
Fig. 2. Flow diagram for a compfetely mixed reactor employing constant concentration of recycle cells, X R .
M H R E 7, Watson Marlow Ltd., England) through a silicone rubber tubing. The soft drink bottling combined wastewater (provided by nearby Pepsi Cola Company) was used in this work as the carbon source (growth-limiting substrate). Analysis showed that the waste had a COD of over 1000 mg/liter and was deficient in phosphorus and nitrogen. The low suspended solids content of the wastewater had the advantage that the organic content was mostly soluble. The raw waste was diluted to the desirable value (in this case, 1000 mg/ liter COD) and of the necessary amount of nitrogen and phosphorous added. Activated sludge was developed from the seeds obtained from the other activated sludge pilot unit treating pineapple canning waste. Microscopic observations revealed that the seeds were free of filamentous organisms. To start the continuous-flow operation, 1 liter of the sludge was added to the reactor and then diluted to 5 liters using tap water. The feed solution (about 1000 mg/liter COD, containing the appropriate amounts of ammonium sulfate and phosphate buffer solution, P H about 7.0) was pumped into the reactor a t the proposed rate. Sludge in the homogenizing tank was pumped into the reactor a t 0.3 times the value of the influent waste flow rate.
QUANTITATIVE SHOCK LOADINGS
47
The reactor was checked for complete mixing by measuring the biological solids and filtrate COD of samples taken from the middle and effluent of the reactor. When there was close agreement between corresponding parameters from both sampling points, the reactor was considered to be completely mixed. The COD was measured according to the Standard Method.” The concentration of biological solids was determined as dry weight by the membrane filter technique using glass fiber filter paper with 0.8-p pore size. The conditions of “steady state’’ were approached by measuring the filtrate COD in the reactor and the effluent at hourly intervals until both the level of solids and filtrate COD in the reactor reached some constant value. After a sampling period of a certain dilution rate, the influent flow rate was slowly altered (in one or two days) to the next predetermined initial flow condition of the successive run. The system was again allowed to go t o another steady state. Shock loadings were introduced by increasing the inflow rate 100 or ZOO’% of the initial value, or by decreasing the initial influent flow rate by 50y0. The schedule is summarized in Table I. TABLE I Summary of Parameters and Variables Initial Final Influent Hydraulic influent influent % recycle dilution rate dilution rate Change Run XR COD D (hr-l) D (hr-l) in D no. S, (mg/liter) (mg/liter) ratio, a 1 2 3 4 5 6
1/3
1/ 2 1 /4 1/2.65 1/1.5 114 1/ 8 1/2 1/1.33
lo00
7 8 9
+loo -50 +200 +3m 100 -50
+
++200 100
RESULTS AND DISCUSSION Operation Under Step-Hydraulic Shocks
Before the step-hydraulic shocks were applied, the initial steady state a t different dilution rates was developed. Nine experimental
WONG A N D YANG
48
e Y
'".I ILO
-
-_ m
RVI No6 lD=l/a-l/4hr-', 100%) No4 (D=l/O-l/ZVZ.65b-'.XX)XI
0 RVI
Irn -
2
8 Y L
noM-
=: . a
I , , , , , , , , , , , , , , , , , , , ,
runs were made in which the initial steady state was disrupted by applying a step change in influent flow rate only. The recycle sludge flow rate was altered correspondingly to keep the recycle flow ratio a at a constant of 0.3. I n Figures 3, 4, and 5, the disruption due to different degrees of shocks at each of the dilution rates, D = &,&,and hr-l, respectively, is compared. I n Figure 3, the initial value of D is hr-' and 100 and 2007& increases in D are shown. It can be seen that for the 100% step increase in D there is a rapid increase in cells and the substrate decreases. Then there is a lag of 2 hr and an increase in the concentration of biological solids occurs until the final steady level. The filtrate COD varied between 24 and 54 mg/liter during the transient period. On the other hand, there is a lag of 4 hr for the 200% step increase in D for the biological solid concentration in the reactor before it rises to a peak a t the ninth hour and then drops to the final steady-state level. During the transient period, the filtrate COD fluctuates from 40 to 80 mg/liter. The new steady-state value of filtrate COD is almost the same as the one before the shock was applied (about 40 mg/liter COD). The transient period takes about 14 hr in both cases. The difference in the modes of response can be explained by recalling that the microorganisms employed are heterogeneous in nature. The mass culture consists of microorganisms growing a t different specific growth rates. When D is changed from
49
QUANTITATIVE SHOCK LOADINGS
Fig. 4.
Comparison of loo%, 200%, and 300% changes in D for system at initial D of hr-1.
+
+
to $ hr-I (100% change), and to 1/2.65 hr-I (200% change), those microorganisms which can adapt to the new growth condition will be developed. Therefore, the lag period is needed before these species become the dominant populations.
a
I
0
RU(Y)S(D=c/4-c/t33k-',',%)
0
RUW NO 6 (D.1/4-!/2
b-', 1 0 0 % )
Rvw WO.7(o.ln-l/.
b-'.-w)%)
m
D
I4
m
T W .k
Fig. 5. Comparison of
loo%,
200%, and -50% changes in D for system at initial D of t hr-l.
50
WONG AND YANG
I n Figure 4,similar patterns of response are shown for initial D a t a t different levels of shock loading. It can be seen that the lag period for the biological solids concentration in the reactor takes 3 hr in the case of 100 and 200% increases of D while it takes 2 hr for a 3oOy0 increase of D. For the 3000/, increase of D, instead of increasing the biological solids concentration in the transient period, the concentration has been decreased to another steady state. For the filtrate COD concentration, there is not much change (between 35 and 70 mg/liter) following the 100 and 200y0 increases of D in the transient period, but there is a major change (between 40 and 140 mg/liter) following the 300y0 increase of D. The transient period for the case of 100 and 200% increases of D takes about 7 hr, and 15 hr for a 300% increase of D. I n Figure 5 shocks have been made a t initial D of hr-' a t a different level of hydraulic shocks. Unlike the previous two cases, no time lag for biological solids and COD concentration in the reactor is observed even when a 50% negative change in D is included. In general, the transient period of these three cases takes about 10 hr. I n Figures 6, 7, and 8 the same degree of hydraulic shocks and different initial dilution rates are compared. It has been found that using the constant XRprocess, a faster growing culture (higher initial
+ hr-l
a
Fig. 6. Comparison of hydraulic shocks (100% increase) a t different initial dilution rates.
QUANTITATIVE SHOCK LOADINGS
51
Fig. 7. Comparison of hydraulic shocks (200% increase) at different initial dilution rates.
dilution rate) responded to an increase in D better than a slower growing culture (lower initial dilution rate). When imposing transient conditions on a pure culture of s. cerevisiae fed ethanol by altering the change in hydraulic retention time, Mor and Fiechter12 observed that the time delay for the system to respond to a shock was dependent on the microbial net specific growth rate, p. A short time delay was recorded when the specific growth rate was high, and a long delay was recorded when the value of p was low. I n this study it is found that under constant X R operation, the system can accommodate a step increase in D up to 200% and a step decrease in D by 50% without much disruption of the effluent quality based on the COD measurement. In general, the response time for the new steady state to bc approached depends on the initial dilution rates. The higher t h r value of initial D , the shorter will be the response time. However, the initial D value should not be too high because the growth rate will bc too close to the maximum specific growth rate, p m . Any perturbations due to an increase in D will lead to cell washout and substrate leakage for the cultures which can no longer adapt to the new condition of change. I n the present study, it is found that an initial dilution rate of t hr-l is optimal to givv thc shortmt response time and it can prevent cell washout and substrat(>lcakagv \iit hiri t h(x present range of hydraulic shock loading.
52
WONG AND YANG
Fig. 8. Comparison of hydraulic shocks (50% decrewe) a t different initial dilution rates.
CONCLUSIONS On the basis of the findings in this study, the following conclusions are drawn. 1) With the initial dilution rates employed (D = $, $, hr-l), the system can accommodate a step increase in D up to 200% and a step decrease by 50% without disrupting the effluent quality. 2) The higher the value of initial dilution rates (e.g., 2 hr-l) the shorter the response time.
Nomenclature a c
D F S, S SR V X
X,
xi
xR
recirculation ratio; ratio of rate of underflow from the clarifier to the rate of inflowing waste concentration factor; the ratio of the biological solids in the clarifier underflow to that in the reactor F / V , dilution rate of the system (hr-1) rate of flow of incoming substrate (liter/hr) influent substrate concentration (mg/liter COD) steady-state concentration of substrate in the reactor (mg/liter COD) filtrate COD in the recycle sludge (mg/liter) effective volume of reactor (liter) steady-state biological solid concentration in the reactor (mg/liter) biological solid concentration in the effluent of the system (mgiliter) biological solid level in the influent waste b g / h t e r ) concentration of cells feedback (mdhter)
QUANTITATIVE SHOCK LOADINGS
53
XW sludge wastage (kg/day) p CL,
specific growth rate of microorganism (hr-l) maximum specific growth rate of microorganisms (hr-1)
This work was supported by a scholarship donated by the Government of Canada to the Asian Institute of Technology.
References 1. M. Ramanathan and A. F. Gaudy, Jr., Biotechnol. Bioeng., 13, 125 (1971). 2. D. Herbert, Chem. Znd. Monogr., 12, 21 (1961). 3. M. Ramanathan and A. F. Gaudy, Jr., Biotechnol. Bioeng., 11, 207 (1969). 4. A. F. Gaudy, Jr. and R. Srinivasaraghavan, Biotechnol. Bioeng., 16, 723 (1974). 5. R. Srinivasaraghavan and A. F. Gaudy, Jr., J. Water Pollut. Contr. Fed., 47, 1946 (1975). 6. F. M. Bonotan-Dura and P. Y. Yang, Biotechnol. Bioeng., 18, 145 (1976). 7. F. F. Storer and A. F. Gaudy, Jr., Environ. Sn’. Technol., 3, 143 (1969). 8. T. K. George and A. F. Gaudy, Jr., J. Environ. Eng. Div., Proc. ASCE, 99, 593 (1973). 9. D. W. Eckhoff and D. Jenkins, “Transient loading effects in the activated sludge process,” presented a t 3rd Int. Conf. Water Pollut. Res., Section 11, Paper 14, 1966. 10. A. F. Gaudy, Jr., BioteChnoZ. Bioeng., 17, 1051 (1975). 11. Standard Methods for the Examination of Water and Wastewater, 13th ed., American Public Health Association, New York, 1971. 12. J. R. Mor and A. Fiechter, Biotechnol. Bioeng., 10, 787 (1968).
Accepted for Publication August 20, 1976
