Gravitational sedimentation of flocculated waste activated sludge

(1)

Gravitational sedimentation of ﬂocculated waste

activated sludge

C.P. Chu

a

, D.J. Lee

a,

*, J.H. Tay

b

a

Department of Chemical Engineering, National Taiwan University, Taipei 10617, Taiwan b

Division of Environmental and Water Resources Engineering, School of Civil and Environmental Engineering, Nanyang Technological University, Nanyang Ave., Singapore 639798, Singapore

Received 31 August 2001; received in revised form 28 May 2002; accepted 3 June 2002

Abstract

The sedimentation characteristics of flocculated wastewater sludge have not been satisfactorily explored using the non-destructive techniques, partially owingto the rather low solid content (ca. 1–2%) commonly noted in the biological sediments. This paper investigated, for the first time, the spatial-temporal gravitational settling characteristics of original and polyelectrolyte flocculated waste activated sludge using Computerized Axial Tomography Scanner. The waste activated sludge possessed a distinct settling characteristic from the kaolin slurries. The waste activated sludges settled more slowly and reached a lower solid fraction in the final sediment than the latter. Flocculation markedly enhanced the settleability of both sludges. Although the maximum achievable solid contents for the kaolin slurries were reduced, flocculation had little effects on the activated sludge. The purely plastic rheological model by Buscall and White (J Chem Soc Faraday Trans 1(83) (1987) 873) interpreted the consolidatingsediment data, while the purely elastic model by Tiller and Leu (J. Chin. Inst. Chem. Eng. 11 (1980) 61) described the final equilibrated sediment. Flocculation produced lower yield stress duringtransient settling, thereby resultingin the more easily consolidated sludge than the original sample. Meanwhile, the flocculated activated sludge was stiffer in the final sediment than in the original sample. The data reported herein are valuable to the theories development for clarifier design and operation. r2002 Elsevier Science Ltd. All rights reserved.

Keywords: Activated sludge; CATSCAN; Flocculation; Elastic; Plastic; Constitutive equation

1. Introduction

Slurry sedimentation characteristics are essential to the design of thickener in water and wastewater industries. Traditional flux theory ignored the role of the bottom sediment [1–3]. Dixon [4] discussed the compression of sediment that traditional flux theory does not account for. Tiller [5] and Fitch [6] refined the Kynch theory to consider the effect of increasing sediment duringsedimentation. Font [7,8] reviewed the works consideringthe compression zone effects in batch sedimentation.

The way the solid fraction (volume fraction of solid phase, f) changes with solid pressure (Ps) controls the rheological characteristics of the consolidating sediment. As Koenders and Wakeman [9] revealed, most related works assumed an elastic ﬁlter cake havingthe following power-law type constitutive equation:

f ¼ f_g a þPs P0

b

; ð1Þ

where a can be taken as 0 or 1. f_gis the null-stress solid fraction or the gel point if a ¼ 1 [37]. P0represents the characteristic solid pressure [10]. b denotes the compres-sibility of the sediment subjected to external stress. Eq. (1) assumes that f depends sorely on the local solid pressure Ps: Hence the solid fraction in a consolidating

*Correspondingauthor. Fax: +886-2-2362-3040. E-mail address:[email protected] (D.J. Lee).

(2)

sediment would vary continuously alongthe vertical axis from the settler bottom.

Buscall and White [11] proposed a consolidation model consideringthe ﬂocculated slurry as a purely plastic body that possesses a yield stress ðPyðfÞÞ: If the local pressure of sediment exceeds Py; the network structure would yield. Buscall and White [11] proposed the followingconstitutive equation:

Df

Dt ¼ kðf; PÞ½P PyðfÞ; ð2Þ

if P > PyðfÞ: Otherwise, Df=Dt ¼ 0; that is, no solid consolidation would occur in the sediment. k is called the ‘‘dynamic compressibility’’ of the suspension. This model had been extensively applied to sedimentation processes of ﬂocculated suspensions (see, for instance Refs. [12–15]). If the sediment were purely plastic, there should exist a constant solid fraction portion in the sediment in which the solid stress has not exceeded the yield stress.

Researchers have adopted several non-destructive techniques to obtain the relevant variable in the solid/ liquid separation processes as a function of space and time, such as g-ray [13,16], NMR/MRI [17–19], and X-ray [12,20–26]. Buscall [27] reviewed the non-destructive techniques in solid–ﬂuid system.

Polyelectrolyte conditioninghas longbeen utilized to pre-treat the sludge in order to increase its filterability [28]. However, few works in literature investigated the sedimentation characteristics of flocculated wastewater sludge using the above-mentioned, non-destructive techniques, partially owingto the rather low solid content (ca. 1–2%) commonly noted in the biological sediments. We built up the solid fraction distributions in this paper for the original and flocculated waste activated sludge with repeated scanning using a Com-puterized Axial Tomography Scanner (CATSCAN). A kaolin slurry was used as the reference system for comparison.

2. Experimental 2.1. Sample

Waste activated sludge sample was collected from the 69th Street Wastewater Treatment Plant,

Houston, TX on December 14, 1998. The pH

of the sludge was 6.5. The weight fraction was 1.3% (w/w). An Accupyc Pycometer 1330 (Micro-meritics) measured the solid density as 1630 kg/m3. Therefore the initial solid fraction of the sludge was around 0.8% (v/v).

Kaolin slurry was prepared by mixingprescribed amount of powders with 0.1 M NaClO4 solution at a solid fraction of 1% (v/v). The addition of NaClO4 provided a high ionic strength to prevent the inter-ference of any ions that might be released from the particle surfaces. The pH value was ﬁxed at 7.0. An Accupyc Pycnometer 1330 (Micromeritics) measured the solid density as 2730 kg/m3.

A cationic polyelectrolyte denoted as polymer T-3052 was obtained from Kai-Guan Inc., Taiwan. The polymer T-3052 is a polyacrylamide with an average molecular weight of 107 and a charge density of 20%. The polymer solution was gradually poured into the original slurry with 200 rpm of stirring for 5 min, followed by 50 rpm for another 20 min. A small quantity of floc-polymer aggregates was transferred carefully into the fresh electrolyte at the same pH and electrolyte concentration as the original electrolyte after mixingand prior to settling. The z potentials of the floc-polymer aggregates were measured by the Zetameter (Zeter-Meter System 3.0, Zeter-Meter Inc., USA). Table 1 lists the surface charge and the floc size data. For both sludges, the floc size was increased monotonically with the flocculant dose, and the z potential neutralized at 0.8 g/kg dried solid (DS) for kaolin slurry and 15 g/kg DS for the waste activated sludge, respectively.

Nomenclature

a parameter in Eq. (1) (dimensionless) g gravitational acceleration (m/s2₎ K hindered coefﬁcient (dimensionless) k dynamic compressibility in Eq. (2) (m s/kg) Lc height of sediment starts to yield (m) P0 characteristic pressure in Eq. (1) (Pa) Ps solid pressure (Pa)

Py yield stress (Pa)

t settlingtime (s)

U0 Stokes velocity (m/s) uS solid velocity (m/s)

y the height of supernatant-slurry interface (m)

z vertical distance from the coordinate position-ingdownward with its origin located at the top of sediment (m)

Greek letters

b empirical constant in Eq. (1) (dimensionless) k dynamic compressibility in Eq. (2) (m s/kg) f solid fraction (dimensionless)

f_g null-stress solid fraction (dimensionless) f₀ initial solid concentration of clay slurry

(dimensionless)

Dr density difference between solid and liquid (kg/m3₎

(3)

2.2. CATSCAN

Batch settlingexperiments were conducted in a settlingcylinder that had a height of 20 cm and a diameter of 9.5 cm, and were continuously monitored with the CATSCAN (DeltaScan 2060; 100 kV, 75 mA). The X-ray attenuation, which is normally presented as the CT number [29], could be obtained from a horizontal section of settlingsludge at a thickness of 0.5 mm. Bergstrom [13] adopted similar procedures usingtheir g-ray apparatus. Preliminary tests revealed that the CT number for the solution that contained ﬂocculant (no solid particles) was essentially zero since the organic compounds did not absorb a signiﬁcant amount of X-ray. Furthermore, the CT-number was found to increase monotonically with the solid fraction in the sludge (Fig. 1). The uncertainties in axial position measurement and the local solid volume fraction were 2.5 mm and 0.1%, respectively, close to those reported in Auzerais et al. [12].

Fig. 2 demonstrates some cross-sectional CATSCAN images for the kaolin and the biological sludge sediments. The luminescence intensity (0–255 in gray scale) increased accordingto the solids fraction, with dark portions in Fig. 2 denoting the pore and the bright ones the solids phase. Apparently the X-ray adsorption of the kaolin slurry was much stronger than the activated sludge, mainly being attributed to the greater solid contents of kaolin slurries. Multiple tests were required to enhance the resolution of the activated sludge tests. Moreover, ﬂocculation led to rather non-uniform distribution over the cross-section of the sludge sediment, as evidenced by Fig. 2c.

3. Results and discussion 3.1. Time evolutions

Figs. 3 and 4 depict the transient settling behavior of kaolin slurries and activated sludges, respectively. A supernatant and a constant-solid fraction regime were noted in the ﬁrst 1000 s settlingof original kaolin slurry

(Fig. 3a). Sediment compaction became apparent after 3000 s of settling. The highest solid fraction at the bottom was 21%, 21 times higher than the initial solid fraction (ca. 1%). Flocculation accelerated the settling while sediment compaction started within 100 s of settling (Fig. 3b for 2 g/kg DS and Fig. 3c for 4 g/kg DS). All curves with time greater than 1000 s coincided with each other, indicatingthat the compaction had completed. Meanwhile, the ﬂocculant reduced the compressibility of the sludge cake, accompanied with which the maximum achievable solid fraction at tube bottom decreased to about 11%.

For original kaolin slurry test a zone-settling regime at f ¼ f₀was evident at to1000 s of settling. However, we could not identify the constant-f_g regime as proposed by the purely elastic model [11]. No such identiﬁcation was possible for the ﬂocculated slurries owingto their extremely fast settling.

The original activated sludge settled slower than the kaolin slurry (Fig. 4a). Many ﬂocs suspended in the upper liquid layer that made supernatant-slurry

inter-Table 1

The z-potentials and ﬂoc sizes of the original and ﬂocculated sludges

Kaolin slurries (f0¼ 1%) Waste activated sludge (f0¼ 0:8%)

Dose (g/kg DS) z potential (mV) Floc size (mm) Dose (g/kg DS) z potential (mV) Floc size (mm)

0 17.7 6.3 0 16.3 86.9 0.8 12.1 31.5 2 11.2 92.1 2 10.8 41.5 8 10.1 103 4 1.2 48.5 15 +1.4 202 12 +5.5 118 30 +4.8 269 CT number 0 100 200 300 400 Solid fraction 0.00 0.05 0.10 0.15 0.20 Activated sludge Kaolin slurry

(4)

Fig. 2. The cross-sectional CATSCAN images of the sludge sediment (0.5 cm from bottom): (a) original kaolin slurry, (b) kaolin slurry, 2.0 g/kg DS, (c) kaolin slurry, 12 g/kg DS, (d) original activated sludge, (e) activated sludge, 8 g/kg DS, (f) activated sludge, 15 g/ kgDS. 0 5 10 15 20 Solid fraction Solid fraction 0.0 0.1 0.2 Equilibrium 100 sec 1000 sec 30 00 sec 6000 sec 0 5 10 15 20 0.0 0.1 0.2 Equilibrium 100 sec 1000 sec 3000 sec 6000 sec

Distance from bottom (cm)

0 5 10 15 20 Solid fraction 0.0 0.1 0.2 Equilibrium 100 sec 1000 sec 3000 sec 6000 sec (a) (b) (c)

Fig. 3. Transient settling of original and ﬂocculated kaolin slurries: (a) original, (b) 2 g/kg DS, and (c) 4 g/kg DS.

0 5 10 15 20 Solid fraction 0.00 0.01 0.02 0.03 Equilibrium 100 sec 1000 sec 3000 sec 6000 sec 0 5 10 15 20 Solid fraction 0.00 0.01 0.02 0.03 Equilibrium 100 sec 1000 sec 3000 sec 6000 sec

0 5 10 15 20 Solid fraction 0.00 0.01 0.02 0.03 Equilibrium 100 sec 1000 sec 3000 sec 6000 sec (a) (b) (c)

Fig. 4. Transient settling of original and ﬂocculated waste activated sludge: (a) original, (b) 8 g/kg DS, and (c) 15 g/kg DS.

(5)

face blurred in the ﬁrst 6000 s of settling. After 24-hours settlingthe maximum solid fraction at tube bottom was 3.3%, about 3.1 times higher than the initial fraction (ca. 0.8%). Flocculation yielded a satisfactory settleability for the activated sludge (Fig. 4b for 8 g/kg DS and Fig. 4c for 15 g/kg DS). (Activated sludge at 2 g/kg DS exhibited a poor settlingresemblingto the original case, whose CATSCAN data are not shown here for brevity sake.) The supernatant was clear and a distinctive supernatant-sediment interface was noticeable over all testingtime. Resemblingto the kaolin tests, most of the settlingprocess of ﬂocculated activated sludge was in sediment compaction stage rather than in the commonly recognized zone-settling regime.

For the original activated sludge a nearly constant-f regime persisted during settling, but with the corre-sponding f values changing with time. For instance, at t ¼ 100 s, the constant f was 0.95% while at 6000 s the solid fraction became 1.4%. This observation indicates a continuous structural change for the networked flocs during settling. On the other hand, flocculation sig-nificantly stiffened the network structure of the waste activated sludge whose constant f regime was of a solid fraction around f_g(ca. 1.2%).

3.2. Final sediment

Fig. 5a and b display the spatial distributions of solid fractions in the ﬁnal sediments of the kaolin slurries and of the activated sludges, respectively. For kaolin slurries, the sediment height increased with the ﬂocculant dose

and reached a maximum at 0.8 g/kg DS. Restated, the sludge became looser as more flocculant was added. A loose cake is beneficial for sludge dewatering. The solids fraction distributions of kaolin slurries flocculated at different polymer dosages resembled each other, reveal-ingthat intensive flocculation did not affect the packing characteristics of the kaolin particles.

The original activated sludge exhibited a continuous increase in solid fraction from the top surface of the sediment. A rather low solid fraction regime (ca. 0.2%) persisted above the underneath sediment after 24-hours settling. Visual observation revealed that the suspended ﬂocs in the upper layer solution should have accounted for the noted low solid fraction regime in Fig. 5b. The solid fraction rapidly increased at ho11:5 cm, de-notingthe surface of the sediment. Large pores existed in the ﬁnal sediment (as shown in Fig. 2c) indicatingthe occurrence of large water channels duringsettling[30,31]. This occurrence increased the heterogeneity of the internal structure of the sediment.

For the flocculated samples, on the other hand, the sweep of fine particles by flocculated flocs pro-duced clear supernatant and well-compacted sediment. The reproducibility of the CATSCAN measurements was satisfactory. The solid fraction in the final sediment was less than the original one. Similar to the kaolin slurry tests, flocculation led to bulky sediment that contained more interstitial (intra- or inter-floc) water in sediment.

Although visual observation revealed a clear super-natant-sediment interface for the settlingsamples, the transition in solids fraction measured usingCATSCAN between the sediment and the supernatant was gradual and exhibited no sudden jump in solid fraction over the interface as proposed by most highly idealized theories. A procedure was adopted herein to estimate the f_g value, as schematically depicted in Fig. 6. Since the ultimate error for solid fraction estimation was 0.1%, the position where f ¼ 0:1% was taken as the upper surface of the equilibrium sediment. The ‘‘interfacial zone’’ between the suspension and the sediment was taken from the surface (f ¼ 0:1%) to 1 cm below. The average solid fraction over this 1-cm zone was defined as f_g; and the results were listed in Table 2. The f_g of waste activated sludge kept at around 1.2%. That is, the large and porous activated sludge flocs (with a size of 87 mm for original sludge and 200 mm for flocculated ones) might entangle with each other, and among these flocs a network formed. Flocculation only mildly influenced the sediment structure of waste activated sludge, as evidenced by the nearly constant f_gvalues in Table 2. For kaolin slurries, on the other hands, f_gwas 9% for the original sludge, The value dropped to less than 4% after flocculation, indicatinga loose, net-worked structure with flocculants.

0 5 10 15 20 Solid fraction 0.0 0.1 0.2 Original 0.8 g/kgDS 2 g/kg DS 4 g/kg DS 12 g/kg DS (a)

0 5 10 15 20 Solid fraction 0.00 0.01 0.02 0.03 Original 2 g/kg DS 8 g/kg DS 15 g/kg DS 30 g/kg DS (b)

Fig. 5. Solid fraction distributions in the ﬁnal sediments: (a) Kaolin slurries, (b) waste activated sludges.

(6)

The local solid pressure (Ps) in the ﬁnal sediment were evaluated as follows: Ps¼ Z z 0 @Ps @z dz ¼ Z z 0 ðDrgfÞ dz ¼ Dr g Z z 0 f dz; ð3Þ

where Dr is the density difference between the solid and the liquid phases, g; the gravitational acceleration, and z; the vertical distance from the coordinate positioning downward with its origin located at the top of sediment whose correspondingsolid fraction is f_g: The local pressure distribution were evaluated by integrating Eq. (3) with respect of z using data in Fig. 5. Fig. 7 presents the curves for solid fraction versus the calculated solid pressure. Large pores in the flocculated sediment could be identified from the CT images, suggesting that the sediment of kaolin slurry had become looser when the flocculant presented (Fig. 2c). Two linear regimes existed for f versus Pscurves on a log–log scale for kaolin slurries, dividing each other by a

threshold pressure (approximately 1–2 Pa). Restated, exceedingthe threshold pressure the sediment of kaolin slurry was more readily compressible.

The solid fraction of the original activated sludge increased from 1% to 1.2% at less than 1 Pa to more than 3% at 3–4 Pa. The original activated sludge was much more compressible than kaolin slurries at the low-pressure range investigated. Moreover, flocculation produced bulky sediment (low solid fraction), which was strongenough to resist yieldingat applied solid pressure. As Fig. 7 displays, the solid fraction of flocculated sludges only slightly increased from f_g to 1:522:0f_g while the solid pressure exceeded 2 Pa. This observation correlates with the sludge compaction studies [32] and for the centrifugal dewatering of polyelectrolyte floccu-lated activated sludge [33]. Whether the sediment would further yield over the investigated pressure range (10 Pa) is not clear.

Eq. (1) correlated the f Ps curves in Fig. 7 using graphical method by Tiller and Leu [10]. The ob-tained parameters accordingto the graphical method were adopted as the initial guess for non-linearly regressing the experimental data. Table 2 lists the results. The parameter P0 was rather low, generally around or less than 1 Pa, indicatingthat these slurries would largely deform at the low-pressure regime. Also, the b values were higher for waste activated sludge than those for the kaolin slurries (0.2–0.3 by Chu and Lee [34] and Chu et al. [35]), statingthat the latter was more compressible than the former. Finally, the higher ﬂocculant dose would yield the lower P0 and b: The sludges after ﬂocculation would deform quickly at the low solid pressure (low P0) but could resist the external applied pressure at a greater solid pressure (low b). Flocculated sludge would pack more readily when it initially formed. But the sediment became rather ‘‘stiff’’ after initial compaction.

3.3. Rheological properties

To analyze the rheological properties of the two sludges, we adopted the model proposed by Landman Fig. 6. The schematic description of the fgdetermination.

Table 2

Model parameters obtained for rheological properties of waste activated sludge Waste activated sludge (f0¼ 0:8%)

Dose (g/kg DS) fg(%) U0K (1014m/s) Py(Pa) P0(Pa) b

0 1.2 ca. 0 9.8 3.9 1.1

2 1.3 ca. 0 9.8 2.5 0.23

8 1.2 1.5 5.7 5.8 0.24

15 1.2 1.5 3.4 1.5 0.20

(7)

and Russel [36] as follows: usðf0Þ ¼ dy dt¼ U0Kðf0Þ 1 qPs=qz Dr gf₀ ¼ U0Kðf0Þ 1 Pyðf0Þ Dr gf₀ðy LcÞ ; ð4Þ

where U0 is the terminal velocity of individual particle; K the hindered factor; y the upper interfacial height where f ¼ f₀; and Lc; the height where f deviates from f₀[15]. A regime of constant f₀was assumed to exist in the consolidatingsediment over the range of z ¼ y where Ps¼ 0 to z ¼ Lc where Ps¼ Py: Hence, the plot with dy=dt versus 1=ðy2LcÞ estimated the U0K and Pyat f ¼ f₀with linear regression analysis. Eq. (4) was valid only for the consolidatingsediment with f > f_g: Apparently the constant-f₀regime for kaolin slurries was at a solid fraction less than f_g (9% for original and 4% for ﬂocculated ones). Only the cases for waste activated sludge warrant further discussion.

Here we took the correspondingheight of f ¼ 0:5f₀ as y and of f ¼ f_g as Lc for the calculation. Fig. 8 displays some dy=dt versus 1=ðy2LcÞ data for the waste activated sludges in the first 500 s. Table 2 lists these obtained Pyand U0K accordingto regression analysis of Eq. (4). For the original waste activated sludge and that dosed at 2 g/kg DS, their supernatants were blurred and the determination of the y values beingdifficult. Since their settlingwas rather slow the correspondingU0Ks were close to zero. Furthermore the Py’s should be the net weight of the solid phase since under the previous assumption the whole slurry should neither settle nor compact significantly in the early stage of setting. The Py’s for flocculated sludges over 8 g/kg DS largely

declined to around 5 Pa, with markedly increased U0K: Hence, duringdynamic settlingthe flocculation led to the more readily deformed sediment compared with the original sample. A rather low Pyfor the waste activated sludge (o10 Pa) indicated that the solids fraction could be easily enhanced (at least) to its gel point (f_g) regardless of the original sludge concentrations. Re-stated, in the final clarifier, the waste activated sludge could be easily concentrated to f_g (1.2% in the investigated case) with the sediment height greater than about 1 m if the effective floc density (Dr) were taken as 1 kg/m3. Flocculated samples are more easily to con-solidate than the original sludge.

The Py values possessed a negative correlation with the b’s based on the solid fraction distribution in the final sediment. This occurrence indicates that although the flocculated sludge could be compacted more readily than the original one, the final sediment for the former was stiffer than the latter. Hence, to increase the sediment height could not sufficiently increase the solid fraction in the underflow. The benefit of employingthe ‘‘superthickener’’ by Johnson et al. [14] is limited for the waste activated sludge subjected to polyelectrolyte flocculation, which is a common practice in wastewater treatment plant. The data reported herein were bene-ficial to the theories development for clarifier design and operation.

4. Conclusions

This paper reported, for the ﬁrst time, the spatio-temporal gravitational settling characteristics of acti-vated sludges using CATSCAN. The tests with kaolin Solid pressure (Pa)

0.1 1.0 10.0 Solid fraction 0.02 0.04 0.06 0.08 0.20 0.01 0.10 Original 0.8 g/kg DS 2 g/kg DS 4 g/kg DS 12 g/kg DS Original 0.8 g/kg DS _{2 g/kg DS} 4 g/kg DS 12 g/kg DS Activated sludge Kaolin slurries

Fig. 7. The solid fraction versus solid pressure plot in the ﬁnal sediment. 1 /( y-Lc) (1 /m) 0 5 10 15 20 d y/ d t (m/s ) -0.00020 -0.00015 -0.00010 -0.00005 0.00000 2 g/kg DS 0 g/kg DS 8 g/kg DS 15 g/kg DS 30 g/kg DS

Fig. 8. Plot of dy=dt vs. 1=ðy LcÞ of the waste activated

(8)

slurries were conducted for comparison. The settling velocity for the waste activated sludge was much lower than the kaolin slurries. Flocculation markedly en-hanced the settleability of activated sludge, but did not markedly reduce its maximum achievable solid fraction. Since the sediment-supernatant interface was blurred for the activated sludge, an averaging procedure estimated the gel-point solid fractions as ca. 1.2% for the original and the flocculated activated sludge. The solid fraction of the original activated sludge after 24-hours settling increased from 1% to 1.2% at less than 1 Pa to more than 3% at 3–4 Pa. Further yieldingneeded higher applied pressure. Flocculation produced bulky sediment with strength to the applied solid pressure. The graphical method by Tiller and Leu [10] correlated the solid fraction distribution in the final sediment. The parameter P0 was rather low, generally around or less than 1 Pa, indicatingthat these slurries would largely deform at the low-pressure regime. The final clarifier could hence easily concentrate the investigated waste activated sludge to f_g (1.2% in the investigated case). The transient settlingfor the investigated activated sludges was described using the plastic rheological model by Buscall and White [11]. Flocculation would produce a lower yield stress. Therefore, although the flocculated sludge was tougher than the original sample in the final sediment, the latter was more compressible than the former in the transient settlingstage.

Acknowledgements

National Science Council, ROC, ﬁnancially supported this work. The assistance from Prof. F.M. Tiller, Prof. K.K. Mohanty, and Dr. WenpingLi of University of Houston on CATSCAN experiments is highly appreciated.

References

[1] Coe HS, Clevenger GH. Methods for determining the capacities of slime-settlingtanks. Trans A I Min Eng 1916;55:356–84.

[2] Kynch GJ. A theory of sedimentation. Trans Faraday Soc 1952;44:166–76.

[3] Talmadge WP, Fitch EB. Determining thickener unit areas. Ind EngChem 1955;47:38–41.

[4] Dixon DC. Compression effects in batch settlingtests. J Environ EngASCE 1982;108:1171–91.

[5] Tiller FM. Revision of Kynch sedimentation theory. AIChE J 1981;27:823–9.

[6] Fitch B. Kynch theory and compression zone. AIChE J 1983;29:940–7.

[7] Font R. Compression zone effect in batch sedimentation. AIChE J 1988;34:229–38.

[8] Font R. Analysis of the batch sedimentation test. Chem EngSci 1991;46:2473–82.

[9] Koenders MA, Wakeman RJ. Initial stages of compact formation from suspensions by ﬁltration. Chem EngSci 1996;51:3897.

[10] Tiller FM, Leu WF. Basic data ﬁttingin ﬁltration. J Chin Inst Chem Eng1980;11:61–70.

[11] Buscall R, White LR. The consolidation of concentrated suspensions. Part I: the theory of sedimentation. J Chem Soc Faraday Trans 1987;1(83):873–91.

[12] Auzerais FM, Jackson R, Russel WB, Murphy WF. The transient settlingof stable and ﬂocculated dispersions. J Fluid Mech 1990;221:613–39.

[13] Bergstrom L. Sedimentation of ﬂocculated aluminum suspension—gamma-ray measurements and comparison with model predictions. J Chem Soc Faraday Trans 1992;88:3201–11.

[14] Johnson SB, Scales PJ, Dixon DR, Pascoe M. Use of a superthickener device to concentrate potable water sludge. Water Res 2000;34:288–94.

[15] Shen C, Russel WB, Auzerais FM. Colloidal gel ﬁltration: experiment and theory. AIChE J 1994;40:1876–91. [16] Shin BS, Dick RI. Applicability of Kynch theory to

ﬂocculant suspension. J Environ EngASCE 1980;106: 505–26.

[17] ChangD, Lee T, JangY, Kim M, Lee S. Non-colloidal sedimentation compared with Kynch theory. Powder Technol 1997;92:81–7.

[18] Friedmann T, Windhab EJ. Determination of ﬁlter cake properties by nuclear magnetic resonance. Sep Sci Technol 1998;33:2221–39.

[19] La Heij EJ, Kerkhof PJAM, Kopinga K, Pel K. Determiningporosity proﬁles duringﬁltration and expres-sion of sewage sludge by NMR imaging. AIChE J 1996; 42:953–9.

[20] Somasudaran P, HuangYB, Gryte CC. CAT Scan characterization of sedimentation and ﬂocs. Powder Technol 1987;53:73–7.

[21] Bierck BR, Dick RI. In situ examination of effects of pressure differential on compressible cake ﬁltration. Water Sci Technol 1990;22:125–34.

[22] Bierck BR, Dick RI. Mechanisms of compressible sludge cake shrinkage. J Environ Eng ASCE 1990; 116:663–82.

[23] Bierck BR, Wells SA, Dick RI. Compressible cake ﬁltration: monitoringcake formation and shrinkage using synchrotron X-rays. J Water Pollut Control Fed 1988;60:645–50.

[24] David KE, Russel WB, GlantschnigWJ. Disorder-to-order transition in settlingsuspension of colloidal silica: X-ray measurements. Science 1989;245:507–10.

[25] Tiller FM, Yeh CS. Relative liquid removal in ﬁltration and expression. Sep Filtr 1990;27:123–35.

[26] Tiller FM, HsungNB, CongDZ. Role of porosity in ﬁltration XII: ﬁltration with sedimentation. AIChE J 1995;41:1153–64.

[27] Buscall R. The sedimentation of concentrated colloidal suspensions. Colloid Surf 1990;43:33–53.

[28] ChangIL, Chu CP, Lee DJ. Filtration followed by expression characteristics of polymer ﬂocculated clay sludge. J Colloid Interface Sci 1997;185:335–42.

[29] Hounsﬁeld GN. A method of and apparatus for examination of a body by radiation such as X- or

(9)

Gamma-radiation. British Patent No. 1,283,915, London, UK, 1972.

[30] Chen GW, ChangIL, HungWT, Lee DJ. Regimes of zone settlingof waste activated sludge. Water Res 1996; 30:1844–51.

[31] Vesilind PA, Jones GN. Channelingin batch thickening. Water Sci Technol 1993;28(1):59–65.

[32] Chu CP, Lee DJ. Expression of polyelectrolyte ﬂocculated sludge. J Chin Inst Chen Eng 2000;31:321–31.

[33] Chu CP, Lee DJ. Experimental analysis of centrifugal dewateringprocess of polyelectrolyte ﬂocculated waste activated sludge. Water Res 2001;35:2377–84.

[34] Chu CP, Lee DJ. Solids ﬂuxes in consolidatingsediments. J Chin Inst Chem Eng2001;32:547–54.

[35] Chu CP, Ju SP, Lee DJ, Mohanty KK. Batch gravitational sedimentation of slurries. J Colloid Interface Sci 2002; 245(1):178–86.

[36] Landman KA, Russel WB. Filtration at large pressures for strongly ﬂocculated suspensions. Phys Fluids A 1993; 5:550.

[37] Landman KA, Buscall R, White LR. The continuous ﬂow gravity thickener: steady state behavior. AIChE J 1988; 34:239–52.