INFLUENCE OF SUBSTRATE LOADING INTENSITY ON FLOC SIZE IN ACTIVATED SLUDGE PROCESS

(1)

Pergamon

0043-1354(94)00326-2

I N F L U E N C E OF SUBSTRATE L O A D I N G INTENSITY ON FLOC SIZE IN ACTIVATED S L U D G E PROCESS

K R Z Y S Z T O F BARBUSIlqSKI* and H E L E N A K O S C I E L N I A K

Departmelat of Environmental Engineering, Silesian Technical University, 44-101 Gliwice, Poland (First received February 1994; accepted in revised form November 1994)

Abstract--The activated sludge floes size variability and their size distribution were analyzed for the floes sizes larger than 10 #m and for different intensities (rates) of bidirectional organic loading changes. Under these conditions the distributions of floc sizes expressed as average diameter and longest dimension fitted well the log-normal model. The logarithmic average floc size variability proved to be proportional to the loading if a time delay due to the inertia of biochemical and physical processes of activated sludge was taken into account. It has also been found that floes larger than 50 #m constituted the main source of the surface area and volume of the activated sludge. The tests performed have shown that organic loading strongly influences floc size distribution in the activated sludge process. The long-term loading changes caused larger disturbances to the floc size distribution than more rapid but shorter ones. Generally, after the substrate overload occurred, the floes were more prone to breakup (pin-point floc effect).

Key words--activated sludge, substrate overloads, floc size distribution, geometric properties of floe, surface area, volume, pin-point floc

INTRODUCTION

The effectiveness of the activated sludge process is related to the physical properties of the flocs. In the literature, a number of parameters have been ident- ified as adequate indicators of the structure and physical properties of these microbiological aggregates. The basic ones are: floes size and shape, their density and porosky, settling velocity, as well as coherence and specific surface area. Activated sludge floes are an important research subject, as their physical state and properties influence both the mass transport and the effectiveness of separation of treated wastewater from the activated sludge biomass.

The above physical properties of microbiological aggregates are directly related to their size. This was confirmed in a number of publications (Lauben- berger, 1970; Parker et al., 1971; Magara et al., 1976;

Tambo and Watanabe, 1979; Sezgin et al., 1978;

Hermanowicz and Ganczarczyk, 1983). Recent research by Ganczarczyk and co-researchers (Li and Ganczarczyk, 1987, 1988, 1992; Zahid and Ganczarczyk, 1990; Nfimer and Ganczarczyk, 1993) proved that the settling velocity of microbiological aggregates is a function of the geometric parameters of these particles, as well as of their density and porosity. As the particle size increased, the density decreased, and the porosity increased. It has also been shown that the ]porosity of activated sludge floes

*Author to whom all carrespondence should be addressed.

increases at a higher rate when floc sizes are smaller than 200 # m (longest dimension) as compared with that of the floes larger than 2 0 0 # m (Li and Ganczarczyk, 1987). The particle size and especially their size distribution are the result of dynamic equilibrium state between formation, transformation and breakage of microbiological aggregates. Size of activated sludge flocs may change in a broad range with reference to the conditions of process performance. Therefore, better recognition of the floc size variability under varying conditions of biological wastewater treatment may lead to more profound understanding and better control of the activated sludge process.

Also, presence of filamentous microorganisms play a very important role in forming the floc structure.

According to Sezgin et al, (1978), the filaments are believed to form a "backbone" of activated sludge floes on which floe-forming bacteria are fixed by means of extraceilular polymers. On the other hand, the excessive growth of filamentous microorganisms mechanically prevents floes from a better com- paction. As a result, rather loose, low density aggregates are formed (filamentous bulking). Nevertheless, the excessive appearance of filamentous microorganisms need not always cause filamentous bulking. It depends on the filament type, the shape and mor- phology of the floes (Pipes, 1979; Wanner, 1992;

Nov~k et al., 1993). These problems are very complex and can be evaluated from different viewpoints.

In practice, the control of activated sludge systems is carried out on the basis of the organic loading or 1703

(2)

1704 Krzysztof Barbusifiski and Helena Ko~cielniak ( F / M ratio). Previous research into the affect o f

organic loading o n the physical properties of activated sludge has c o n c e n t r a t e d o n the settleability a n d sludge volume i n d e x - - S V I (Bisogni a n d Lawrence, 1971; C h a o a n d K e i n a t h , 1979; P a l m et al., 1980;

B a r a h o n a a n d Eckenfelder, 1984). The works listed a b o v e c o n t a i n n o a d d i t i o n a l m e a s u r e m e n t s o f average floc size or its distribution. The investigations were limited to microscopic o b s e r v a t i o n s o f the sludge flocs, a n d aimed at d e t e r m i n i n g trends in their size distribution. In some cases the dispersion o f the flocs was m e a s u r e d in a simplified way, e.g. as a c o n c e n t r a t i o n o f effluent s u s p e n d e d solids ( C h a o a n d K e i n a t h , 1979). Li a n d G a n c z a r c z y k (1993) per- f o r m e d a n extensive study to assess the influence o f t r e a t m e n t p l a n t o p e r a t i n g c o n d i t i o n s o n the size d i s t r i b u t i o n a n d dispersion o f the activated sludge flocs. They c o n c l u d e d t h a t the organic loading a n d the availability o f dissolved oxygen per unit o f organic loading D O / ( F / M ) were the two m o s t significant factors influencing the size d i s t r i b u t i o n o f the activated sludge flocs.

C u r r e n t research is aimed at o b s e r v i n g the variability in average floc size a b o v e 10 p m , a n d floc size d i s t r i b u t i o n , in response to the degree (or intensity) o f organic loading. These experiments were per- f o r m e d at different rates o f bidirectional changes in organic loading. In each experiment, total p e r i o d o f loading changes included g r a d u a l increase o f organic loading followed by its decrease to the initial level. As the d u r a t i o n o f total p e r i o d was s h o r t e n e d , m o r e drastic c o n d i t i o n s in p a r t i c u l a r experiments were created. The applicability o f the floc size measure- m e n t for a n on-line c o n t r o l o f the activated sludge process is also being evaluated.

MATERIALS AND METHODS Operational conditions of the laboratory unit

The five parallel continuous-fow laboratory units, each consisting of a bioreactor (aeration basin) coupled with a clarifier was used in this study. Aeration and mixing was achieved using compressed air, introduced through a porous stone diffusers in the bottom of the reactors. The air flow intensity was adjusted with the aid of precalibrated rotame- ters to maintain the content of dissolved oxygen (DO) in each reactor 1.5 mg 1 - i. The inoculum was activated sludge provided by the municipal wastewater treatment plant in

Pyskowice (Poland). This is a conventional activated sludge plant of 2 MGD (7400 m 3 per day) capacity, receiving both domestic (85%) and industrial (15%) wastes. Compressed air is used to aerate the content of aeration tanks, in which the mixed liquor hydraulic retention time is 5.5 h. During period of the study, concentration of the raw sewage was approximately 286 mg COD 1-~.

The adaptation of microorganisms to synthetic wastewater, consisting of peptone (as the carbon source) and mineral salts was carried out at an organic loading close to 0.2 kg COD/kg MLSS/day. In this phase, the units were operated for a period of 4 times the sludge age to establish steady-state conditions. After completing this phase, ad- ditional testing was carried out for a few days to confirm factual stabilization of biochemical and physical properties of the sludge (cf. Barbusifiski, 1991). To attain this the control comprised such parameters as: substrate removal rate, sludge growth, oxygen uptake rate, dehydrogenase activity (TTC test), as well as sludge volume index and floc size. The adaptation period was regarded to be finished when the range of changes of particular parameters within 7 days' time did not exceed 10-15%. Moreover, periodic microscopic observations were conducted to determine the stabilization of activated sludge biocenosis composition.

After the acclimatization and stabilization period, five parallel experimental series were performed at different rates of bidirectional changes in organic loading. Changes in loading were done by altering the feed concentration (COD). The particular reactors were operated at a constant hydraulic retention time of 6 h (in Series I-IV) and 4 h (in Series V). In each experiment, total period of loading changes comprised gradual increase of organic loading and following gradual decrease to the initial level (Table 1).

Total duration (period) of loading changes in succeeding experimental series was shortened from 42 to 1 day. This created more drastic conditions of process performance.

Frequency of loading changes in individual experiments varied as well (Table 1). In each experiment, after total period of loading changes, a final stage run was carried out.

Here steady-state conditions were provided, the organic loading being about 0.2 kg COD/kg MLSS/day. This stage was introduced to allow the activated sludge to stabilize again. The MLSS concentration was maintained at approxi- mate level of 3000 mg 1 - i, by wasting the excess sludge from the reactors once or twice per day (depending on sludge growth rate). The pH within the reactors was in the range of 6.8-7.2. The operating parameters of each experimental series are presented in Table 1.

Chemical oxygen demand (COD) of influents and effluents, dissolved oxygen content (DO), mixed liquor suspended solids in the bioreactors (MLSS), sludge volume index (SVI) and pH were measured in accordance with Standard Methods (APHA, 1980). COD, MLSS and SVI were analyzed just before and after each change of organic loading. DO and pH were measured continuously by an oximeter (OXI-196) and pH-meter (pH-196) respectively

Table 1. Operating data of the laboratory scale activated sludge units

Series

Total

duration Frequency of

Experiment of loading loading COD Changes in

time ~ changes b changes HRT range organic loading

(day) (day) (h) (h) (mgl-~) (kg COD/kg MLSS/day)

1 56 42 24 6 150- 1000 0.17--) 1.68--0.26

I1 25 16 24 6 150-700 0.18-~0.81---)0.22

III 16 4 12 6 150-700 0.20--*0.80--~0.17

IV 12 1 6 6 150-700 0.20-*0.99--.)0.2 I

V 12 1 4 4 100-800 0.22---* 1.80--.*0.20

~Including total duration of loading changes (unsteady-state) as well as initial and final stabilization stages (steady-state).

blncluding gradual increase and then gradual decrease of loading.

HRT--hydraulic retention time.

COD--chemical oxygen demand.

(3)

Floc size in activated sludge process (both made by WTW, Germany). SVI determinations were

performed in unstirred 1 litre graduated cylinder.

Size measurements

In previous research into the geometrical characteristics of biological floes (average diameter, longest dimension, surface area, etc.) a wide range of methods have been applied. Direct micreseopic observations of floes with an eyepiece micrometer were performed (Ganczarczyk and Kosarewicz, 1961; Finstein, 1967; Finstein and Heukelekian, 1967; Mueller et al., 1967; Parker et al., 1971;

Hermanowicz and Ganczarczyk, 1983; Sadalgekar et aL, 1988) or photographs of individual floes were taken (Laubenberger, 1970; Magara et al., 1976; Tambo and Watanabe, 1979; Li and Ganczarczyk, 1987). Recently, image analysis systems are being used more frequently for this purpose (Roth and Pinnow, t981; Glasgow et aL, 1983;

Li and Ganczarczyk, 1986, 1988, 1991, 1992). In the course of this research, direel, microscopic floc size measurements were performed using an eyepiece micrometer. Although this method does not allow all the geometric parameters of the floes, to be measured, e.g. their cross-sectional area and perimeter, its main advantage is simplicity. This enables practical application of this method both in a laboratory and on a technical scale.

The activated sludge was sampled using a wide mouthed micropipette, to minimise damage during transfer (Mueller et al., 1967; Sezgin et al., 1978; Li and Ganczarczyk, 1986, 1987). The samples were diluted in order to stop the floes forming larger agglomerates. Depending on floc size and degree of dispersion, a dilution ratio of between 1 : 5 and I : 10 was used. A single microscopic slide was then prepared from each sample, and the parameters of the floes were measured. Sampling, dilution and slide preparation was repeated for the appropriate number of measurements.

Each time 100 activated sludge floes were analyzed. The average diameter of individual floes was expressed as a halved sum of the longe.st and the shortest dimension of each floe. After analyzing 100 floes, mean value of their average diameters (average size) was calculated. A total of 6500 floes were measured altogether. Additionally, the surface area and the volume of individual floes were calculated as the surface area and the w)lume of a sphere with the diameter equal to the mean diameter of the floe. Because of irregular shapes of the activated sludge floes it is not possible to calculate their real surface area and volume on the basis of direct geometrical measurements. Therefore in practice, above quantities are fr,squently estimated with assumption that the floes are spi~erical (e.g. Mueller et al., 1967;

Laubenberger, 1970; Li and Ganczarczyk, 1991). This enables determination of general relations between external surface area, volume and size of floes.

RESULTS AND DISCUSSION F l o c s i z e d i s t r i b u t i o n

The floc size range considered here was 10-900/~m.

The lower limit was set for two reasons. Firstly, the errors associated with applying a direct observation technique to f l o c s < 10 # m are large. To obtain more accurate results, a Coulter Counter should be used.

Secondly, as proved by Li and Ganczarczyk (1991, 1993), flocs larger th~.n 10/~m are the main source of surface area, volume and mass in activated sludge.

A total of 65 floc size distributions in five measurement series were analyzed. In almost all the cases the distribution was log-normal, described by the formula

1705

xO'|n.~x/2II L 2alnx

where x is the characteristic size of a floc, and ~,x and cq,.~ are logarithmic average and logarithmic standard deviation of a floc size, respectively.

Only about 8% of the samples did not fit the model. The model parameters varied as follows: ~ from 3.63 to 5.35 and c%~ from 0.36 to 0.89. The log-normal distribution of biological aggregate sizes has only been reported a few times (Mueller et al.,

1967; Li and Ganczarczyk, 1991, 1993; Nfimer and Ganczarczyk, 1993). Although the experiments performed by the above authors concerned different levels of activated sludge loading, they have only been carried out under steady or almost steady state conditions, while the present research focused on rapid loading changes over a wide range from 0.17 to

1.8 kg COD/kg MLSS/day.

In Series I, the long-term changes in the activated sludge loading caused most visible perturbations in the floc size distribution in comparison with remaining experimental series. Figure 1 shows three examples of the floc size distributions expressed as the frequency of occurrence in Series I. Histogram A presents floc size distribution at the beginning of the experiment, just before increasing organic loading was gradually introduced. As the loading changed from 0.17 to 0.50 kg COD/kg MLSS/day, the average floc size increased from 121 to 149 #m. During this time the floc size distributions were similar to that shown in Fig. l(a). The highest frequency corresponded to the size range 75-100/~m and varied from 0.25 to 0.21 as the load increased. For loadings greater than 0.5 kg COD/kg MLSS/day a clear trend was observed, with the peak shifting towards larger sizes (up to 150--175/~m), but with a gradual decrease in frequency to 0.16 [Fig. l(b)]. Simultaneously, the upper limit of the floc size range gradually shifted from 500 to 9 0 0 g m [see Fig. I(a) and (b)]. At this stage the sedimentation properties deteriorated quickly and the sludge could be described as filamentous bulking sludge. The flocs were large, coherent, with predominant filamentous microorganisms. The sludge characteristics were similar to that described by Sezgin e t aL (1978) as Case I. Reversing the gradient of sludge loading changed the trend in floc size distribution. The highest frequency peak kept shifting towards smaller floc sizes, with gradual increase in frequency. The amount of filamentous microorganisms decreased, and the flocs became less coherent. After loading dropped to 0.48 kg COD/kg MLSS/day, a temporary stabilization of floc size distribution was noted, the distribution being similar to the initial one [see Fig. l(a)]. The maximum frequency of 0.19-0.20 corresponded again to the range of 75-100/~m. Further reduction of loading to the level of 0.26 kg COD/kg MLSS/day caused rapid decrease in the average floc size. At the same time some undesirable changes in the quality of the activated sludge were observed, The flocs became very

(4)

0.30"-~ (a)

0.25

0.20

0.15

0.10

0.05

0.25 t 0.20

0 . 1 5 o

O.lO (b)

10 25 75 125 175 250 350 450 550 650 750 850 15 50 100 150 200 300 400 500 600 700 800 900

0.05

10 25 75 125 175 250 350 450 550 650 750 850 15 50 100 150 200 300 400 500 600 700 800 900

0.35 -~ (c)

0.30

0.25

0.20

o o~

0.15 tL

0.10

0.05

I 0 25 75 125 175 250 350 450 550 650 750 850 15 50 100 150 200 300 400 500 600 700 800 900

(l~m)

1706

Fig. 1. Influence of bidirectional changes in organic loading on floc size distribution (FSD) in Series I.

(a) Initial FSD observed after completing stabilization stage but before starting an increase phase in sludge loading (organic loading= 0.17kg/kg/day); (b) FSD corresponding with maximum organic loading

(1.68 kg/kg/day); (c) FSD corresponding with pin-point floe.

(5)

Floc size in activated sludge process ₁₇₀₇ weak and would break up following even slight

changes in turbulence in the bioreactor. The outflow from the clarifier also had increased turbidity and suspended solids levels. Moreover, the flocs sedi- mented very quicldy while the sludge volume index (SVI) was very low. Activated sludge with these properties is terraed pin-point floc, meaning the sludge in which the flocs breakup (Sezgin et al., 1978;

Pipes, 1979; Palm et al., 1980). Hence, it can be claimed that the poor quality of the activated sludge was due to the breakup of the flocs. In this case, 44%

of all the flocs fell into the size range of 10-25/~m and 70% into the size range of 10-50 # m [Fig. l(c)].

In Series II, a similar trend in floc size distribution was observed. Although, due to the shorter experiment duration and narrow range of loadings, no floc breakups were noted. The size distribution by the end of the experiment was similar to the initial one. The only difference was the size range, with the maximum frequency, of 75-100/~m shifted to 100-125 #m. This remained in agreement with the fact that average floc size was larger at the end of the experiment. However in Series III-V, the floc size distributions did not differ as much as in Series I and II. The size range with the maximura frequency was 75-100#m, this only occasionally shifted down to 50-75/~m range or up to 100-125 # m range.

For all the experimental series activated sludge flocs were less coherent after completing the period of loading changes than initially. It may be supposed that after overloadls had occurred, the flocs became potentially more prone to breakup. At the beginning of each experiment, filamentous microorganisms were present in small quantities in the biocenosis of activated sludge. Each time, the filamentous content increased rapidly soon after organic loading exceeded approximately 0.5 kg COD/kg MLSS/day. However only in Series I, the filaments were clear predominant in flocs at high loadings. In remaining cases, the filamentous and tic,c-forming microorganisms stayed in balance at high loadings. As organic loading decreased the number of filaments was significantly reduced.

The distribution of floc size in terms of surface area and volume was also analyzed. The results are shown in Table 2 in four size categories; they correspond to Series II with long-term and gradual overloads and

Table 2. Distribution of activated sludge flocs Distribution

Size (/tm) By nul~ber By area By volume Series H

10050 3-15% < 1% Negligible

50-125 24-72% 645% 2-30%

125-250 21~4% 48~i9% 58-70%

>250 1-23% 6-53% 12-72%

Series I V

10-50 5-1,1% < 2% Negligible

50-125 49-76% 21-52% 9-38%

125-250 18-3'~% 39-58 % 28-70%

> 250 0-7% 0-30% 0056%

ee

t~

e~t

,..1 0

180

155

130

105

80 0.1

(I) Initial stage (If) Final stage _ Decrease

in loading J J

-

I ~ ~ . ~ ~'~ ~ i n loading

I I I I

0.3 0.5 0.7 0.9

Organic loading (kg COD/kg MLSS/day)

Fig. 2. Hysteresis effect of the floc size under bidirectional changes of organic loading.

Series IV with short-term but rapid overloads. Flocs larger than 50/~m were the major source of the surface area and volume of activated sludge biomass.

This remains in agreement with the observations by Li and Ganczarczyk (1991). This regularity was also true when flocs smaller than 50/~m were predominant in number, as a result of pin-point floc [Fig. l(c)].

Although 50 v m or smaller flocs constituted 70% of the total number in this case, larger flocs represented 92% of the surface area and 98% of the volume. The greatest differences between the distributions presented in Table 2 may be seen for size category over 250 #m. This was due to much larger sizes of the flocs in Series II, as opposed to Series IV. In all cases flocs with sizes within the 10-50/~m range contributed less than 2% of the surface area and negligible volume.

Floc size

Since flocs size has been shown to be governed by a log-normal probability distribution, further discussion of the results will use the logarithmic rather than arithmetic average values.

The experiments proved that the sizes of activated sludge flocs changed in direct response to the organic loading. The degree of mathematical correlation of the results, however, was not too high. In all the cases the floc size curves for increasing and decreasing organic loading were not the same but revealed a significant hysteresis effect (Fig. 2). For increasing rate of loading changes in successive experiments both curves approached each other. In the graphs, clear initial and final stages of flocs size settings may be distinguished, corresponding to the steady state directly preceding and following loading changes.

Sometimes the areas overlapped, meaning that the flocs returned to their initial size. This occurred in Series I l l - V , when short-term but rapid substrate overloads took place (Table 3). In case of long-term loading changes (Series I and II) the flocs size settled at a new level, much higher than the initial value.

The hysteresis effect in the floc size change curve as well as in other activated sludge characteristics,

WR 29/7-~G

(6)

1708 Krzysztof Barbusifiski and Helena Ko~cielniak

Table 3. Changes in activated sludge floc geometric parameters, according to log-normal distribution Series

Parameters I II III IV

AFS (pm) 113~211---~127 (38) ~ 95---~166----,113 88--.126---90 88---~116---.92 82---,103--,78 LD (,um) 153--.305----,165 (44) a 128---~225--~ 151 1 1 0 ~ 1 6 4 ~ 1 1 1 114--.153---115 117---~144--, 109 SD (/~m) 50---~ 98---, 64 (30)" 54---,78---,60 5 6 - - - ~ 8 1 - - - , 6 2 56----78---,56 42~51--~46 E (--) 2.3---~ 3.1---~2.0 (I.4) a 1.843.4---.2.3 1.84-.2.3--.1.7 1.8---.2.1---.1.9 2,0---~2.6----2.1 AFS---Iogarithmic average floc size; LI~-Iogarithmic longest dimension; SD--Iogarithmic shortest dimension; E---elongation (LD/SD ratio).

"Breakup of flocs.

(dehydrogenase activity, sludge volume index) result- ing from bidirectional loading changes had been reported previously by Barbusifiski (1992).

The hysteresis was due to the time delay in the response of a given parameter to the stress conditions introduced (Fig. 3). This means that the activated sludge floc size variability has a certain inertia and hence reversing the direction of the substrate overload cannot produce an instant change in the floc size trends. As it has been observed, the delay between the time of maximum loading and the time the flocs size reached its maximum was 24 h in Series I to III and 6 and 4 h in Series IV and V, respectively. The times corresponded to the frequency of introducing the organic loadings in the individual series (Table 1).

The increase in the rate of loading changes caused shorter time delays due to a faster response in floc size changes. Evidence for this comes from the fact that in experiments with slow changes in organic loading the maximum, logarithmic average floc size was greater than those with faster loading changes. The maximum sludge floc logarithmic average size decreased from 211 p m in Series I to 103 p m in Series V (Table 3).

Linear regression was applied to evaluate the relationship between floc size and organic loading.

Correlation coefficients (which were significant at the 95% level) are presented in Table 4. The results of the regression analyses indicated, that with the time delay accounted for particular experimental series, the linear correlation between these parameters increased considerably. Thus it has been demonstrated that the variability in activated sludge floc size remains in proportion to the organic overloads despite time delays due to the inertia associated with the bio-

1.00

.~

^0.75

~ 0.50

~ 0.25

0 8

140

115

14 20 26 90

Time (day)

19o

165 g

o

3

Fig. 3. Time delay of the logarithmic average floc size--2, in relation to organic loading changes--l.

chemical and physical processes of the sludge. Acti- vated sludge floc size is a parameter dependent on the loading and as such, may be an important indicator of changes in activated sludge biocenosis.

Geometric relationships in the flocs

In all experiments a strict, linear correlation between the logarithmic average floc size and their logarithmic longest dimension was noted (Table 5).

The relationship between logarithmic average floc size and their logarithmic average shortest dimension was more ambiguous. No direct relationship between the sizes of the flocs and the elongation ratio has been found. As it may be seen from the data in Table 3, very large flocs were also highly elongated. As the size of the flocs decreased, their elongation ratio decreased as well. Since however, the longest dimension decreased faster than the shortest dimension, no direct relationship between the floc size and the elongation ratio could be expected.

A linear relationship between the logarithmic average and the logarithmic standard deviation of floc size has been observed. As the logarithmic average floc size increased, the logarithmic standard deviation increased as well (Fig. 4). Measurement points circled in the graph correspond to the floc breakup effect case (pin-point floc effect) in Series I, when the logarithmic average size of the flocs fell down to 38 # m (Table 3). For this reason these points were not taken into account in calculating the correlation coefficient (r = 0.86).

Although the cause of the pin-point floc phenom- enon is not known (Pipes, 1979) it seems likely that it is related to the quantitative presence of extracellular polymers in the activated sludge. It has been generally accepted that the main factor binding individual clusters of microorganisms into small and then larger flocs are extracellular polymers (ECP), facili- tating the biological flocculation and providing a stability of a floc form. For example, Pavoni et al.

(1972) and then Chao and Keinath (1979) have shown that microorganisms produce smaller amounts of ECP at higher loading values of the activated sludge. Series I corresponded to long term, high value organic loadings, at which the ECP level should be low. At the same time, introduction of a negative gradient of loading changes caused the condition of continuous deficiency of the substrate. According to Pavoni et al. (1972) the amount of ECP in the activated sludge should then increase due to the

(7)

Floc size in activated sludge process

Table 4. Linear correlation coefficients between organic loading and logarithmic average floc size

Method of result Series

analysis I II III IV V

Time delay not

accounted for 0.55 0.80 0.68 0.45 0.76

Time delay

accounted for 0.85 0.95 0.90 0.96 0.94

1709

decrease in loading level. On the other hand, under the substrate deficiency condition, the microorganisms may use boca poly-fl-hydroxybutyrate (PHB) and ECP as a storage substance (Richard et al., 1985).

It is thus possible 1:hat in the discussed case the rate of depletion of ECP as a storage material could be higher than its production rate, in consequence leading to the breakup of the activated sludge flocs. Due to floc breakup both longest and shortest logarithmic dimensions of the flocs were rapidly decreased (Table 3). The longest dimension though changed more quickly, cawsing more uniform flocs to be formed, as indicated by the low value of the elongation ratio.

It has been found that the loading of the activated sludge exerts a significant influence on the floc size distribution. Generally, long-term changes in sludge loading caused larger disturbances to the floc size distribution than more rapid, short lasting ones.

This research has proved that the measurement of floc size and its distribution, may be an important tool in the evaluation of the activated sludge process, under dynamic conditions. As individual biochemical and physical activated sludge properties are also dependent on the organic loading level, there must be a certain relation between those properties and flocs size. Due to the large number of factors affecting activated sludge performance which may obscure this relationships, further research should be carried out under clearly defined, forced substrate overload conditions.

The simplicity of measurement, as well as the low cost of the equipment are both advantages of applying the direct microscopic observations method, for in-field assessment of the activated sludge process.

Some personal experience in this domain would certainly be necessa::y, hence it has to be stressed that applying image arLalysis systems is currently the most adequate and comprehensive way of testing the microbiological aggregates.

Table 5. Linear correlation coefficients between floc size and others geometric parameters according to log-normal

distribution

Series AFS :LD AFS: SD AFS: E

I 0.997 0.949 0.756

II 0.980 0.447 0.635

llI 0.984 0.873 0.383

IV 0.981 0.859 0.088

V 0.995 0.863 0.538

AFS--logarithmic average floc size; LD--Iogarithmic longest dimension; SD--Iogarithmic shortest dimension;

E---elongation (LD/SD ratio).

C O N C L U S I O N S

1. Under the conditions of dynamic loading changes the distribution of floc sizes (expressed as the average diameter and the longest dimension) larger than 10 # m may be approximated by a log-normal model, Only about 8% of samples did not fit to this model.

2. The size of the activated sludge flocs showed a direct proportionality to the changes of the organic load over a wide range of substrate overloads (0.17-1.8 kg COD/kg MLSS/day), even though some time delays of the response were noted due to the inertia of the biochemical and physical processes of the sludge. The size of the flocs increased with the value of the loading.

3. In all the cases, a strict linear correlation between the logarithmic average floc size and their logarithmic longest dimension was preserved. A linear relationship was also found between the logarithmic average and the logarithmic standard deviation of floc size. The standard deviation increased with the increase of the mean floc size.

4. Flocs larger than 5 0 # m constituted the main source of the surface area and volume provided by the activated sludge.

5. It has been found that the organic loading characteristics has an important effect on the floc size distribution in the activated sludge system. Long- term loadings of the activated sludge caused much bigger disturbances to floc size distribution (even leading to floc breakup; pin-point floc effect) than even more rapid but short-lasting changes.

A 200

"~ t 60 o

120 o o o

Pin-point floc

0 | 1 I I I

o 25 75 125 175 225

L o g a r i t h m i c a v e r a g e f l o c s i z e (p.m)

Fig. 4. Relationship between logarithmic average and logarithmic standard deviation of floc size.

(8)

1710 Krzysztof Barbusifiski and Helena Kogcielniak Acknowledgements--The authors would like to thank

Professor J. Ganczarczyk of University of Toronto, Canada, for his helpful comments on the manuscript.

REFERENCES

APHA (1980) Standard Methods for the Examination of Water and Wastewater, 15th edn. American Public Health Association, Washington, D.C.

Barahona L. and Eckenfelder W. W. (1984) Relationships between organic loading and zone settling velocity in the activated sludge process. Wat. Res. 18, 91-94.

Barbusifiski K. (1991) Adaptation of activated sludge to laboratory research conditions (in English). Envir.

Protect. Engng (Poland) 17, 57-65.

Barbusifiski K. (1992) Changes of selected biochemical and physical properties of activated sludge under forced substrate overload conditions. Ph.D. thesis, Silesian Tech.

Univ. Gliwice (Poland).

Bisogni J. J. and Lawrence A. W. (1971) Relation- ships between biological solids retention time and settling characteristics of activated sludge. Wat. Res. 5, 753-763.

Chao A. C. and Keinath T, M. (1979) Influence of process loading intensity on sludge clarification and thickening characteristics. Wat. Res. 13, 1213-1223.

Finstein M. S. and Heukelekian H. (1967) Gross dimension of activated sludge flocs with reference to bulking. J. Wat.

Pollut. Control Fed. 39, 33-40.

Finstein M. S. (1967) Growth and flocculation in a Zoogloea culture. Appl. Microbiol. 15, 962-963.

Ganczarczyk J. J. and Kosarewicz O. (1961) Dispersion of activated sludge flocs. Gaz Woda i Tech. Sanit. (Poland) 35, 19-21.

Glasgow L. A., Pollock R. J. and Barkley W. A. (1983) Particle size reduction by breakage in biological wastewater treatment. Biotechnol. Bioengng 25, 901-918.

Hermanowicz S. and Ganczarczyk J. J. (1983) Some fluidization characteristics of biological beds. Biotechnol.

Bioengng 25, 1321-1330.

Laubenberger G. (1970) Structure and Physical Properties of Activated Sludge Flocs (in German). Karlsruher Berichte zur Ingenieurbiologie, H.3.

Li D. H. and Ganczarczyk J. J. (1986) Application of image analysis system for study of activated sludge flocs. War.

Pollut. Res. J. Can. 21, 130-140.

Li D. H. and Ganczarczyk J. J. (1987) Stroboscopic determination of settling velocity, size and porosity of activated sludge flocs. Wat. Res. 21, 257-262.

Li D. H. and Ganczarczyk J. J. (1988) Flow through activated sludge flocs. Wat. Res. 22, 789--792.

Li D. H. and Ganczarczyk J. J. (1991) Size distribution of activated sludge flocs. Res. J. Wat. Pollut. Control Fed.

63, 806-814.

Li D. H. and Ganczarczyk J. J. (1992) Advective transport in activated sludge flocs. Wat. Envir. Res. 64, 236-240.

Li D. H. and Ganczarczyk J. J. (1993) Factors affecting dispersion of activated sludge flocs. Wat. envir. Res. 65, 258-263.

Magara Y., Nambu S. and Utosawa K. (1976) Biochemical and physical properties of an activated sludge on settling characteristics. Wat. Res. 10, 71-77.

Mueller J. A., Morand J. and Boyle W. C. (1967) Floc sizing techniques. Appl. Microbiol. 15, 125-134.

N~mer J. and Ganczarczyk J. J. (1993) Settling properties of digested sludge particle aggregates. Wat. Res. 27, 1285-1294.

Nov~ik L., Larrea L., Wanner J. and Garcia-Heras J. L.

(1993) Non-filamentous activated sludge bulking in a laboratory scale system. Wat. Res. 27, 1339-1346.

Palm J. C., Jenkins D. and Parker D. S. (1980) Relationship between organic loading, dissolved oxygen concentration and sludge settleability in the completely-mixed activated sludge process. J. Wat. Pollut. Control Fed. 524 2484-2506.

Parker D. S., Kaufman W. J. and Jenkins D. (1971) Physical conditioning of activated sludge flocs. J. Wat. Pollut.

Control Fed. 43, 1817-1833.

Pavoni J. L., Tenney M. W. and Echelberger W. F. (1972) Bacterial exocellular polymers and biological flocculation.

J. Wat. Pollut. Control Fed. 44, 414-431.

Pipes W. O. (1979) Bulking, deflocculation, and pinpoint floc. J. Wat. Pollut. Control Fed. 51, 62-70.

Richard M., Hao O. and Jenkins D. (1985) Growth kinetics of Sphaerotilus species and their significance in activated sludge bulking. 3. Wat. Pollut. Control Fed. 57, 68-81.

Roth M. and Pinnow P. (1981) Determination of the particle size distribution in secondary effluents (in German). Vom

Wasser 57, 309 327.

Sadalgekar V. V., Mahajan B. A. and Shaligram A. M.

(1988) Evaluation of sludge settleability by floc characteristics. J. Wat. Pollut. Control Fed. 60, 1862-1863.

Sezgin M., Jenkins D. and Parker D. S. (1978) A unified theory of filamentous activated sludge bulking. J. Wat.

Pollut. Control Fed. 50, 362-381.

Tambo N. and Watanabe Y. (1979) Physical characteristics of flocs--I. The floc density function and aluminium floc.

Wat. Res. 13, 409-419.

Wanner J. (1992) Comparison of biocenoses from continuous and sequencing batch reactors. Wat. Sci. Technol. 25 (6), 239 249.

Zahid W. M. and Ganczarczyk J. J. (1990) Suspended solids in biological filter effluents. Wat. Res. 24, 215-220.