See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/28602623

Relative Performance of Different Shaped

Surface Aeration Tanks

Article in Water Quality Research Journal of Canada January 2007

Impact Factor: 0.91 Source: OAI

CITATIONS READS

5 123

3 authors, including:

Bimlesh Kumar Ajey Patel

Indian Institute of Technology Guwahati National Institute of Technology, Warangal
140 PUBLICATIONS 351 CITATIONS 3 PUBLICATIONS 8 CITATIONS

SEE PROFILE SEE PROFILE

Available from: Bimlesh Kumar

Retrieved on: 07 July 2016
 Water Qual. Res. J. Canada, 2007 Volume 42, No. 1,
Copyright 2007, CAWQ

Relative Performance of Different Shaped Surface Aeration Tanks

Achanta Ramakrishna Rao,* Bimlesh Kumar and Ajey Kumar Patel
Civil Engineering Department, Indian Institute of Science, Bangalore, 560012 India

Surface aeration experiments were conducted in tanks of length to width ratio (L/W) 1.0, 1.5 and 2 as well as circular tanks
to study their relative performance on oxygen transfer and energy consumption while re-aerating deoxygenated water. An
identical geometric similarity of various linear dimensions of aerators, rotor blades and rotor diameter was maintained for
all sizes and shapes of aeration tanks tested. The power consumed per unit volume concept to simulate the oxygen transfer
coefficient k was found to be valid for all three shapes of aerators. Simulation equations to predict the oxygen transfer
coefficient k for any given dynamic parameter governing the theoretical power per unit volume, X, were developed for
rectangular tanks of L/W ratios equal to 1.5 and 2. Results were compared with simulation equations (correlating k and X)
for square and circular tanks. A simulation criterion correlating the oxygen transfer coefficient k with actual and effective
power consumption per unit volume for three shapes of tanks was developed. Energy consumption per unit volume of water
was also analyzed. Examples illustrating the application of results are presented. It has been found that the circular tanks are
the most energy efficient, i.e., they produce maximum k for a given effective/actual input energy, followed by square tanks,
rectangular tanks of L/W = 1.5 and rectangular tanks of L/W = 2. This suggests that the circular tank performs better as far
as power requirements are concerned and hence provide better economy.

Key words: aeration, aerators, oxygen transfer, power, wastewater treatment

Introduction power per unit volume is an established criterion to simulate

the oxygen transfer coefficient (Rao 1999 and Rao et al.
Aeration is one of the important processes employed in water 2004), while studying such a relative performance, one may
and wastewater treatment to reduce biochemical oxygen consider the relationship of oxygen transfer rates in terms of
demand. The basic phenomenon behind the process of the theoretical power per unit volume:
aeration is gas transfer, in which gas molecules are exchanged
between the liquid and the gas at the gas-liquid interface (Fair 3 2
et al. 1971). Aeration processes are also used to remove = 4 1 (1)
volatile substances present in water and wastewater and 3n 3
improve the dissolved oxygen content.
Many types of aerators are used in practice, such as
cascade aerators, spray nozzles, diffused or bubble aerators, where N is the rotational speed of the rotor, D is the diameter of
and surface aerators. Among them, surface aerators are the rotor, is the kinematic viscosity of the water and (P/V) is the
popular because of their comparable efficiency and ease of actual and effective power per unit volume. Rao (1999) and
operation. Rao et al. (2004) developed the simulation criteria for unbaffled
A typical surface aerator with six flat blades, used in this square and circular surface aeration tanks, correlating the
study, is shown in Fig. 1. The main component of these
surface aerators is an impeller or rotor, to which the six flat
blades are fitted. The rotor is rotated to create turbulence in
the water body so that aeration takes place through the
interface of atmospheric oxygen and the water surface. The
rate of oxygen transfer depends on a number of factorssuch
as intensity of turbulencewhich, in turn, depends on the
speed of rotation, size, shape and number of blades, diameter
and immersion depth of the rotor and size and shape of
aeration tank, as well as on the physical, chemical and
biological characteristics of water (Rao 1999).
The performance of surface aeration systems is
measured in terms of their oxygen transfer rate; hence, the
choice of a particular surface aeration system depends on its
performance and the efficiency of its oxygen transfer rates.
Therefore, there appears to be a need to study and analyze the
relative performance of these aerators. As the concept of

*Corresponding author; [email protected] Fig. 1. Schematic diagram of a surface aeration tank.

26
 Relative Performance of Surface Aeration Tanks

oxygen transfer rates with the theoretical power per unit Dimensional Analysis
volume X. Kozinski and King (1966) summarized twelve
studies on cylindrical baffled tanks and reported that the Many investigators including Schmidtke et al. (1977), Udaya
overall oxygen transfer coefficient at 20C (KLa20) is the et al. (1991) and Rao (1999) have successfully used the theory
a
function of the power input per unit volume as KLa20 = (P/V) , of dimensional analysis to describe the surface aeration
where a is a constant parameter ranging from 0.2 to 0.4. process, which depends on geometric, physical and dynamic
Hwang (1983) developed a scale-up equation which indicates variables as explained below:
that volumetric mass transfer coefficient in cylindrical baffled
tanks can be related to the power input per unit volume as Geometric variables. The geometric variables include cross-
0.97
KLa20 = 0.332(P/V) . Hsieh (1991) has developed a sectional area of the tank (A), depth of water in the tank (H),
relationship between mass transfer rates and effective power diameter of the rotor (D), length of the blades (l), width of the
per unit volume as KLa20 = 0.167(P/V)
0.483
for cylindrical blades (b), distance between the top of the blades and the
baffled tanks. Forrester et al (1998) has predicted the horizontal floor of the tank (h) and the number of blades (n)
relationship between the mass transfer rates and effective as shown in Fig. 1.
0.80 0.06
power per unit volume as KLa20 = (76 21)10 (P/V)
-6
.
Generally, unbaffled tanks are being employed for surface Physical variables. The physical variables include density of
aeration, because these give a higher fluid-particle mass air (ra), density of water (rw), and the kinematic viscosity of
transfer rate for a given power consumption (Johnson and water (n).
Huang 1956; Grisafi et al. 1994; Rao and Jyothish 1997),
which is of paramount importance in designing an aeration Dynamic variables. The rotational speed of the rotor (N) is the
system. dynamic variable associated with the surface aeration process.
The objective of this research is to compare oxygen Variables influencing the oxygen transfer coefficient at
transfer rates and the theoretical power per unit volume (X) 20C (i.e., KLa20) for a given shape of an aeration tank are
for the three shapes of aeration tanks. An additional aim is to therefore given by:
compare their relative performance using a simulations
equation (Rao 1999 and Rao et al. 2004). Lastly, this study KLa20 = f (A, H, D, l, b, h, n, N, g, ra, rw, n) (4)
aims to develop a relationship between oxygen transfer rates
with the actual power per unit volume (P/V) for three shapes
Equation 4 may be expressed in terms of non-dimensional
of unbaffled surface aeration tanks.
parameters as follows:
Theory and Background Information
k = f (A/D, H/D, l/D, b/D, h/D, n, ra/rw, R, F) (5)
Oxygen Transfer Coefficient
where:
While more advanced oxygen transfer models have been
developed in the recent past (McWhirter et al. 1995,
Mahendraker et al. 2005), the two film-theory developed by k = KL a (n g ) is the non-dimensional oxygen
transfer parameter;
Lewis and Whitman (1924) seems to be satisfactory for clean
water, and such a model is used in the present study. = ND 2 n is the Reynolds number; and
According to this theory, the oxygen transfer coefficient
at TC, KLaT may be expressed as follows:
F = N 2D is the Froude number.

=
( s - 0 )- ( s - t ) (2) Alternately it can be expressed as (Rao 1999 and Rao 2004):
L T

k = f (A/D, H/D, l/D, b/D, h/D, n, ra/rw, X) (6)

where ln represents a natural logarithm and Cs, C0 and Ct are
dissolved oxygen (DO) concentrations in parts per million
4 1
(ppm), such that Cs = the saturation DO concentration which where X = F 3 R 3 is the parameter governing the power per
can tend towards very large values, C0 is at t = 0 and Ct is at unit volume. The first six non-dimensional parameters
time t = t. The value of KLaT can be obtained as slope of the represent thegeometric-similarityof the system and the last
linear plot between ln(Cs - Ct) and time t. The value of KLaT parameter represents thedynamic-similarity.
can be corrected for a temperature other than the standard The experiments in the present study have been
o
temperature of 20 C as KLa20, using the van't Hoff-Arrhenius conducted by using six flat blades (n = 6) fitted to a rotor in a
equation (WEF and ASCE 1988): symmetrical manner as several investigators have followed
the same number of flat blades (Schmidtke et al. 1977; Udaya
(T -20) et al. 1991; Rao 1999).
L T = L 20q
(3)
Therefore, the number of blades, n is constant in the
present experiments, also the parameter ra /rw is considered as
where q is the temperature coefficient 1.024 for tap water. invariant. Thus, these two parameters are omitted in the

27
 Rao et al.

analysis. Therefore, the functional relationship of equation 6 the product of the flow discharge and head loss (Rao 1999).
can now be expressed as: Therefore, one may expect a correlation between the effective
(actual or measured) power per unit volume (P/V) and X.
k = f (A/D, H/D, l/D, b/D, h/D, X) (7) Furthermore, the oxygen transfer coefficient (k) is a function
of P/V because k = f(X) and P/V and X are directly related. In
this functional relationship, P/V can be expressed as the non-
Geometric Similarity dimensional form PV. Hence the relationship between k and
P/V can be expressed as:
Equation 7 suggests that if geometric-similarity of the first
five variables on the right-hand side is maintained for any
given shape of aeration tank, then k depends only on X. k = f (P ) (10)

k = f (X) (8) where PV is the non-dimensional form of:

An optimal solution to geometrical similarity has been P P

= (11)
investigated by Udaya et al. (1991) for a square tank by V Vg ( gn) 13
conducting a series of experiments in different sized tanks
and by varying A/D, H/D, l/D, b/D and h/D to a great extent.
The goal is to solve equations 8 and 10 for three shapes of
They suggested the following values to obtain a maximum
tanks by conducting laboratory experiments.
KLa20, for any rotational speed N of the rotor fitted with six
flat blades:
Materials and Methods
A/D = 2.88; H/D = 1.0; l/D = 0.3; (9) Experiments were carried out with the objective of finding
b/D = 0.24; and h/H = 0.94
the effect of tank shape on the oxygen transfer coefficient (k)
under a wide range of X values (dynamic conditions) and
The quantities listed above represent optimal geometric-
power requirements per unit volume for three shapes of
similarity conditions for square tanks. To compare the
tanks.
performance of aerators with different shapes with square
tanks, the same geometric dimensions have been maintained
Experimental Setup
for all tanks in the present experiments.
The ranges of experimental variables are listed in Table 1.
Dynamic Similarity: Power Per Unit Volume Concept Two sizes of rectangular tanks of L:W ratio 1.5 and 2 and
three sizes of each square and circular tank were tested under
When geometric-similarity conditions are maintained, the laboratory conditions. The geometrical similarity expressed
functional relationship represented by equation 7 is reduced in equation 9 was maintained in all tanks.
to a function of dynamic similarity (Rao 1999) for any shape
of aeration tank (equation 8). However, the functional Determination of KLa20
relationship may be different for different shapes.
The intensity of turbulence and wave action of water are
Standard oxygen transfer tests were conducted with tap
major factors associated with surface aeration. Turbulence
water under laboratory conditions. At first, water in the tank
and viscous effects can be described by the Reynolds number
was deoxygenated by adding the required amount of
(R) and surface wave action can be described by the Froude
cobaltous chloride (CoCl2) and sodium sulfite (Na2SO3)
number (F). Hence, both R and F are important to surface
(Metcalf & Eddy Inc. 2004) and thoroughly mixing the
aeration. It may be noted here that both F and R are implicitly
4/3 1/3 2 water. The deoxygenated water was re-aerated by rotating
expressed in X in equation 8 as X = F R , where F = N D/g,
the rotor at desired speeds and maintaining the variables as
and R = ND /n such that X can be considered as a governing
2
per the data presented in Table 1. When the DO
parameter to simulate k (Rao 1999). Hence k can be uniquely concentration began to rise, readings were taken at regular
related to X as previously reported (Rao 1999, Rao et al. intervals until DO increased up to about 80% of the DO
2004). saturation value.A Lutron Dissolved Oxygen Meter was used
Several investigators Horvath (1984), Udaya (1991) and to measure the DO concentration in water. The DO meter
Rao (1999) stated that if power demand (P) per unit volume was calibrated with the modified Winkler's method (APHA
(V) remains constant then there would be a similarity in 1985).
oxygen transfer between geometrically similar systems in Using the known values of DO measurements in terms
clean water systems. It is also suggested that when power of Ct at regular intervals of time t (including the known value
demand per unit volume remains the same in geometrically
of C0 at t = 0), a line was fitted by linear regression analysis of
similar tanks of different sizes, the oxygen transfer coefficient
equation 2, between the logarithm of (Cs - Ct) and t, by
KLa20 remains the same in all tanks. According to Oldshue
assuming different but appropriate values of Cs such that the
(1985) the power per unit volume can be used as a scale-up
regression that gives the minimum standard error of
criterion for homogenous chemical reactions.
estimate was taken. Thus the values of KLaT and Cs were
In equation 8, the parameter X governs the theoretical
power per unit volume and it can be defined from hydraulics obtained simultaneously. The values of KLa20 were computed
principles based on the concept that the power is related to using equation 3 with q = 1.024 as per the standards for clean

28
 a,b
TABLE 1.

Length (L) or Rotor Water Immersion Blade Rotor

C/S area diameter (d) diameter depth height dimensions speed
Sl. A L d D H H L b Min. Max.
No. Tank Shape (m2) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (rpm) (rpm)

29
Relative Performance of Surface Aeration Tanks

a
b -6 -6 2
 Rao et al.

water (WEF and ASCE 1988). Thus the values of KLa20 were It is interesting to note that each set of data points
determined for different rotational rotor speeds N in all of the pertaining to the given shape of the tank fall very closely on a
geometrically similar tanks. unique curve, suggesting the validity of equation 8; however,
the functional relationships are different for different shapes
Measurement of Power Available at the Shaft as the data fall uniquely on different curves. The simulation
equation for square and circular tanks between k and X has
The current (I) and voltage (V) of the power supply given to been established by Rao (1999) and Rao et al. (2004) and the
the DC motor of the aerator were measured by a digital same is plotted in Fig. 2 along with the data from the present
multimeter. The rotational speed of the rotor was measured experiments to verify the validity of such a simulation
by using a digital speed indicator. The power available at the equation.
shaft was calculated by measuring input power and In Fig. 2 the following equations fit well for all data
deducting the losses occurring in the DC motor. Since points belonging to rectangular tanks with L:W ratio of 1.5:1
rotational speed of the shaft was low (70 to 120 rpm) and 2:1 respectively.
compared to the rated speed (1400 rpm) of the DC motor,
only important losses (such as copper and iron losses) were K 0.25

considered for determining the power available at the shaft. kr = = { 0.75 exp ( 0 .19 X ) + 8 .035
To determine these losses, current and voltage at no load (free R (17)
X
rotation of rotor without load) and loaded (rotation with 2 -6
water) conditions were measured during the experiment. The - 7 .955 exp [- 1 .85 (X - 0 .2 ) ]} 10 X
power available at the shaft was then calculated according to
the method given by Cook and Carr (1947). (Rectangular tank L:W = 1.5:1)
From the measurement of no load current (I1) and
voltage (V1), the iron loss was calculated by: and
2
I L = I 1 V1 - I 1 R a (12) K 0.03
kr = = { 0 .6275 exp ( 0.5X ) + 21 .085
R
where Ra is the armature resistance of the DC motor and X (18)
2
I1 R a is the no load copper loss. From the measured loaded 2 -6
- 20 .955 exp [ - 1.85 (X - 0 .2 ) ]} 10 X
condition for current (I2), and voltage (V2), the copper loss
was calculated by:
(Rectangular tank L:W = 2:1)
2
CL = I2 Ra (13)
where kr is the oxygen transfer coefficient k for the
rectangular tanks. Therefore, it is confirmed that k can be
Total loss occurring in DC motor:
related with the theoretical power per unit volume parameter
X, not only for square (Rao 1999) and circular (Rao et al.
TL = I L + C L (14) 2004) tanks but also for rectangular tanks. However, the
simulation equations 17 and 18 are different depending on
Incoming power to DC Motor: L/W. The equations developed earlier by Rao (1999) and Rao
et al. (2004) for square and circular tanks are presented below
Pi = I 2 V2 (15) for comparison:

Power available at the shaft: K - 0 .3

ks =
1.05
= 17 .32 exp + 3 .68
R X
2 2
X (19)
P = Pi - TL = I 2 V2 - I 1 V1 - R a ( I 2 - I 1 ) (16)
2 -6
- 0 .925 exp [ -750 (X - 0 .057 ) ] 10 X
Results and Discussion

(For square tanks)

Experimental data were analyzed with an objective of
computing three major variables k, X and PV and their inter-
K - 4 .5
relationships, leading to an understanding of relative kc = = 10 .45 exp + 2 .45
performance of three shapes of tanks. R X
X (20)
Simulation of Oxygen Transfer Coefficient, k, with 2 -6
- 0 .7 exp [ - 5(X - 0 .35 ) ] 10 X
Theoretical Power Per Unit Volume, X

(For circular tanks)

Oxygen transfer coefficient in surface aeration tanks was
investigated as a function of dynamic variable (X) to verify
equation 8. The experimental data expressed in terms of
X = N3D2/(g4/3n1/3) (equation 1) and k = KLa20(n/g2)1/3 are plotted where ks and kc are the oxygen transfer coefficients (k) for
in Fig. 2. square and circular tanks respectively.

30
 Relative Performance of Surface Aeration Tanks

Fig. 2. Simulation of oxygen transfer rate with theoretical power per unit volume.

31
 Rao et al.

Relative Performance of Surface Aeration Tanks Simulation of Oxygen Transfer on the Basis of
with k and X Effective (Actual) Power per Unit Volume

It is clear from Fig. 2 that for values of X 0.5, square tanks As reported in the literature (Kozinski and King 1966;
offer a higher oxygen transfer coefficient k, followed by Hwang 1983; Hsieh 1991), the rates of oxygen transfer can
rectangular tanks and circular tanks. But at higher values of X be correlated with effective power consumption per unit
(say X 4) rectangular tanks with L/W = 2 give marginally volume. Figure 4 shows the behaviour of the oxygen transfer
higher values of k than the square tanks do, followed by the coefficient with input power per unit volume for square,
rectangular tank of L/W =1.5, and with the lowest values in circular and rectangular tanks. It is interesting to observe that
the circular tanks. Thus, while there seems to be no fixed data for an individual shape of an aerator fall on a unique
trend of k for all X values, there seems to be an order for a curve, suggesting that the oxygen transfer rates can be
given pair of shapes: square tanks always produce a higher k simulated with the actual and effective power per unit
for a given X when compared to circular tanks (Rao et al. volume. After conducting statistical analysis of the data for
2004) whereas rectangular tanks of L/W = 2 produce a higher an individual shape, the associated relationships have been
k than the other rectangular tanks of L/W = 1.5. presented in equations 21 to 24.
The relative performance of circular and rectangular
tanks of L/W = 1.5 and 2 with respect to square tanks in terms -
= 4.64 + [ 0.185(
5 0.5
of the dimensionless oxygen transfer parameter, k/ks, for a 10 ) ] (21)

given X is shown in Fig. 3. It is clear from Fig. 3 that for any
given X, the ratio of k/ks for values of a rectangular tank of (For square tanks)
L/W = 2 and rectangular tank of L/W = 1.5 kr kr
(2) (1.5 )
is always greater than one. This means that the performance
- 0.5
of a rectangular tank of L/W = 2 is better than a rectangular 10
5
= 7.384 + [ 0.33( ) ] (22)
tank of L/W = 1.5 as far as k and X are concerned. But at

higher values of X (X > 0.5), the performance of the
rectangular tank of L/W = 2 is marginally better than the (For circular tanks)
square tank. It is also observed from Fig. 3 that, for any given
X, circular tanks produce lower values of the oxygen transfer -
coefficient k than square tanks. At lower values of X, the 5 0.5
10 = 4.64 +[ 0.13( ) ]

(23)
performance of circular tanks is less than that of rectangular
tanks; however, circular tanks produce better oxygen transfer
rates than rectangular tanks of L/W = 1.5 at high X values (For rectangular tanks where L/W = 1.5)
(high speed). Non-uniform variation of the oxygen transfer
coefficient, as shown in Fig. 2, may be attributed to various -
5
parameters such as turbulence intensity, input power per unit = 4.77 + [ 0.1045( ) ]
0.5
10
(24)
volume, non-uniform spacing around the impeller, etc., which
needs further investigation.
(For rectangular tanks where L/W = 2)

Relative Performance of Surface Aeration Tanks

with k and PV

From Fig. 4 it is evident that for a given effective power per

unit volume, circular tanks produce higher oxygen transfer
coefficient (k) than square and rectangular tanks. As far as
rectangular tanks are concerned, k values decrease as L/W
values increase. The reason for such behaviour in rectangular
tanks may be due to the liquid tending to move along circular
trajectories, resulting in decreased relative velocities between
impeller and fluid, and weak radial flows directed towards
the tank walls. The presence of sharp corners (in square and
rectangular tanks) interrupts circular liquid patterns, and
suppresses the free surface vortex, which increases input
power and small-scale mixing (because of stronger axial
flow). This implies that flow energy decreases with an
increasing aspect ratio. In addition, for the same value of the
oxygen transfer coefficient, k, square tanks require less power
per unit volume of water than rectangular tanks of L/W ratio
1.5, and even less than rectangular tanks of L/W ratio 2 as
Fig. 3. Relative performance of (b) rectangular tanks (2:1), (c) shown in Fig. 4. Overall, circular tanks gave better re-
rectangular tanks (1.5:1) and (d) circular tanks on oxygen aeration rates than all other tanks in terms of the input power
transfer rate, for any given X, compared to (a) square tanks. per unit volume of water. It is clear from Fig. 4 that to achieve

32
 Relative Performance of Surface Aeration Tanks

Fig. 4. Simulation of oxygen transfer rate with actual power per unit volume, P .

33
 Rao et al.

the same oxygen transfer coefficient (k), while aerating the

same volume of water, circular tanks utilize less power than
all other tanks. The relative performance of all tanks with
respect to square tanks is shown in Fig. 5.

Relative Performance of Surface Aeration Tanks

with k and Energy Input

Energy requirements of surface aerators are of paramount

importance while choosing and designing particular types of
aerators to meet the application needs. Energy can be
computed as the product of power and time required to
achieve a desired level of DO concentration. As KLa20 has the
units of inverse of time, one may characteristically express the
energy with the parameter P/KLa20. As k is a non-dimensional
form of KLa20 and PV is the non-dimensional form of P/V, the
energy (parameter) per unit volume (e) may be defined as
follows: e = PV/k.To get a comparative energy requirement for
achieving the same oxygen transfer rates, the energy
parameter e versus k has been plotted for all the surface Fig. 5. Relative performances of (a) circular tanks, (c)
aerators in Fig. 6. The values of PV at k have been generated by rectangular tanks (1.5:1) and (d) rectangular tanks (2:1),
using fixed-point iteration and Newton's method (Curtis and compared to (b) square tanks on oxygen transfer rate, for any
Patrick 2003). It is observed from Fig. 6 that circular tanks given input power per unit volume, P .
consume less energy than any other tank. For the same
oxygen transfer rates, the rectangular tanks of L/W = 2 Application of Results
consume more energy than any other tanks.
The applications of these results on the relative performance
Relationship between X and e of different shaped aeration tanks can be explained under
three special cases of: (1) constant rotational speed of the
The relationship between theoretical power per unit volume rotor; (2) constant power supply; and (3) constant oxygen
X and the effective (actual) energy e of all the aeration tanks is transfer rate KLa20 and they are expressed in Tables 2, 3 and 4
shown in Fig. 7. A common feature of all curves in Fig. 7 is respectively with suitable examples.
that e gradually increases as X increases and attains a
maximum value, then decreases gradually. The maximum Conclusions
value of e is different for each aeration tank. This value
corresponds to the least efficient point, since it corresponds to This paper presented the relative performance of different
the lowest oxygen transfer rate or highest effective power per types of surface aeration tanks. Based on data and analysis
unit volume. The maximum value for square, circular and conducted in this study, the following conclusion can be
rectangular tanks of L/W ratios 1.5 and 2 is 0.05, 0.15, 0.2 made.
and 0.07 respectively. These observations show the existence
of a transitional state in surface aeration systems. To
overcome this transitional state, systems consume more
energy. Based on data shown in Fig. 7, surface aerators should
not be operated near the maximum energy point to
economize their operation.
The relative energy consumption of all the other surface
aeration tanks versus square-shaped tanks has been plotted in
Fig. 8. It is observed from Fig. 8 that the rectangular tanks of
L/W = 2 take more energy per unit volume while aerating the
same volume of water at a lower range of X, but at higher
range of X, rectangular tanks of L/W = 1.5 consume more
energy than any other surface aeration tanks.
Figure 8 shows a descending trend in accordance with
the aspect ratio. Rectangular tanks of L/W = 2 consume more
energy per unit volume than rectangular tanks of L/W = 1.5,
and the energy consumed is least in circular tanks. Based on
these data, it can be concluded that circular tanks are more
suitable for water and wastewater treatment than any other
aeration tank.
Fig. 6. Energy requirements while achieving the same oxygen
transfer rate for (a) rectangular tanks (2:1), (b) rectangular
tanks (1.5:1), (c) square tanks and (d) circular tanks.

34
 Relative Performance of Surface Aeration Tanks

The oxygen transfer coefficient k is related to the

theoretical power per unit volume (X) for rectangular
tanks of L/W ratio 1.5 and 2 as expressed in equations 17
and 18. Such equations developed earlier for square
(equation 19) and circular (equation 20) tanks were also
verified through further experiments.
For a given X, the oxygen transfer coefficient (k) in
square tanks is always higher than that of circular tanks,
whereas the rectangular tanks of L/W = 2 always
performed better than the rectangular tanks of L/W =
1.5. An interesting feature of rectangular tanks of L/W =
2 is that they perform slightly better than the square
tanks beyond a value of X 0.8 as reflected in Fig. 2
and 3.
It has been established that k can also be simulated with
the effective (actual) power per unit volume (PV) for a
given shape of the tank. Simulation equations governing
k and PV were developed for all the three shapes of
surface aeration tanks as given in equations 21 to 24. It is
established that for a given value of PV, the oxygen
transfer coefficient (k) is the highest in circular tanks
followed by square tanks (L/W = 1), rectangular tanks of
L/W = 1.5 and rectangular tanks of L/W = 2 as shown in
Fig. 4 and 5.
Among all the shapes, circular tanks were the most
energy efficient as shown in Fig. 6. The next efficient
shape was square followed by the rectangular tanks of
L/W = 1.5 and the least efficient were the rectangular
tanks of L/W = 2.
Although the square tanks were the best for quick
aeration, they consumed more energy than the circular
tanks, whereas circular tanks were the most energy
efficient but required more time than the square tanks.

Fig. 7. Energy characteristics of surface aeration tanks

Fig. 8. Energy characteristics of (a) rectangular tanks (2:1), (b)

rectangular tanks (1.5:1) and (d) circular tanks, compared to
(c) square tanks.

35
 Rao et al.

TABLE 2. The performance of different types of aerators to re-aerate 0.347 m of water at a constant rotational
speed of the rotor, N

Surface aeration tanks

Parameter Rectangular Rectangular

Step estimation Parameters Square Circular (L/W = 1.5) (L/W = 2)
Area 1m 1m 1m 1m
Length 1m 1.224 m 1.414 m
Geometric
1 Equations 1 to 9
dimensions Depth 1.128 m
Width 1m 0.82 m 0.707 m

2 Equations 17 to 10 k 2.12 0.344 1.042 2.55

20
3 K= KLa (n/g ) KLa20 0.569/min 0.118/min 0.28/min 0.68/min

4 KLa KLaT 0.641/min 0.133/min 0.316/min 0.774/min

5 KLaT Time t required to reach 150.5 s 723.32 s 305.52 s 124.84 s

80% saturation value

6 Equations 21 to Effective power per unit 154 W/m 28 W/m 120 W/m 221 W/m
24 volume of water

7 E = (t/3600) x Effective energy consumed 6.44 5.62 10.2 Wh/m 6.44 Wh/m
(P/V) per unit volume of water Wh/m Wh/m

8 Energy savings using circular tanks vs. others 14% 81% 36%

Constant rotor speed, N = 80 rpm.

The rotor is rotated at a constant speed until DO concentration, Ct, attains 80% of the saturation value. Assume
the initial DO concentration C = 0 at water temperature T = 25C.
Known variables: T = 25C; n = 0.88 x 10 m /s; g = 9.81 m/s ; volume of water = 0.347 m ; X = N3D2/g4/3n1/3 = 1.42

36
 3 a,b,c
TABLE 3.
Surface aeration tanks
Parameter Rectangular Rectangular
Step estimation Parameters Square Circular (L/W = 1.5) (L/W = 2)
2 2 2 2

5
k

37
1/3 K a20
K a20 n g 2
K a20 K a
K a t

3 3 3 3
E t P/V
Relative Performance of Surface Aeration Tanks

a 3
P/V
b
C
C0
c -6 2 2 3
n
 a,b
TABLE 4. The performance of different types of aerators to re-aerate 0.347 m3 of water at a constant oxygen transfer rate, L 20

Area 1 m2 1 m2 1 m2 1 m2
Length 1m 1.224 m 1.414 m
Geometric
1 Equations 1 to 9
dimensions Depth 1.128 m
Width 1m 0.82 m 0.707 m
2 = L 20 (n/ 2) 1/3 105 1.756 1.756 1.756 1.756

38
Rao et al.

3 Equations 21 to 24 Effective power per unit volume 139.92 W/m3 77.67 W/m3 159.83 W/m3 181.9 W/m3
of water

4 = ( /3600) x ( ) Effective energy consumed per 4.49 Wh/m3 2.59 Wh/m3 5.32 Wh/m3 6.05 Wh/m3
unit volume of water

5 From equations 17 to 20 1.09 5.83 3.81 0.96

6 Energy savings using circular tanks vs. others 73% 105% 133%
a
Constant oxygen transfer rate, L 20, is approximately 0.5/min.
b
Known variables: T = 25C; ?n = 0.88 x 10-6 m2/s; g = 9.81 m/s2; volume of water = 0.347 m3; L 20 = 0.5/min.
 Relative Performance of Surface Aeration Tanks

References surface aerated reactors. Prog. Water Technol.

9:477493.
APHA, AWWA, WPCF. 1985. Standard methods for the Udaya Simha L, Shrma KVNS, Rao ARK. 1991. Effect of
th
examination of water and wastewater. 16 ed. American geometrical parameters for overall oxygen transfer
Public Health Association,Washington, DC. coefficient, p. 15771581. In Proceedings of the
Cook AL, Carr CC. 1947. Elements of electrical engineering. Symposium on Environmental Hydraulics, University of
th
5 ed.Wiley, NY. Hong Kong, Hong Kong.
Curtis GF, Patrick WO. 2003. Applied numerical analysis. 6
th WEF, ASCE. 1988. Aeration and water treatment process. In
ed. Pearson Education, Delhi, India. Manual of practice for water pollution control. Water
Fair GM, Gayer JC, Okum DA. 1971. Elements of water Environment Federation and American Society of Civil
supply and waste water disposal. John Wiley and Sons, Engineers,Alexandria,Va., U.S.A.
New York.
Forrester SE, Rielly CD, Carpenter KJ. 1998. Gas-inducing Notations
impeller design and performance characteristics. Chem.
Eng. Sci. 53:603615. The following symbols are used in this paper:
Grisafi F, Brucato A, Rizzuti L. 1994. Solid liquid mass A cross-sectional area of an aeration tank (L2)
transfer coefficients in mixing tanks: influence of side b width of the blade (L)
wall roughness. Inst. Chem. Eng. Symp. Ser. C0 initial concentration of dissolved oxygen at time
136:571578. t = 0 (ppm)
Horvath I. 1984. Modelling in the technology of wastewater
treatment. Pergamon Tarrytown, New York. Cl concentration of dissolved oxygen in the liquid
Hsieh C-C. 1991. Estimating volatilization rates and bulk phase (ppm)
gas/liquid mass transfer coefficients in aeration systems. CL copper losses
Ph.D. thesis. University of California, Los Angeles, Cs saturation value of dissolved oxygen at test
Calif., U.S.A. conditions (ppm)
Hwang HJ. 1983. Comprehensive studies of oxygen transfer Ct concentration of dissolved oxygen at any time
under nonideal conditions. Ph.D. thesis. University of t (ppm)
California, Los Angeles, Calif., U.S.A. D diameter of the rotor (L)
2
Johnson AI, Huang C-J. 1956. Mass transfer studies in an F N D/g, Froude number
2 2
agitated vessel.AIChE J. 2:412419. g 9.81 m/s acceleration due to gravity (L/T )
Kozinski AA, King CJ. 1966. The influence of diffusivity on H depth of water in an aeration tank (L)
liquid phase mass transfer to the free surface in a stirred h distance between the top of the blades and the
vessel.AIChE J. 12:109116. horizontal floor of the tank (L)
Lewis WK, Whitman WG. 1924. Principles of Gas hf head loss due to rotational movement of the
Absorption. Ind. Eng. Chem. 16(12):12151220. water in an aeration tank (L)
Mahendraker V, Rabinowitz B, Mavinic DS. 2005. I1, I2 input current at no load and loading
Comparison of Oxygen Transfer Parameters from Four conditions respectively
Testing Methods in Three Activated Sludge Processes. IL iron losses
= KLa20(n/g ) = non-dimensional oxygen
2 1/3
Water Qual. Res. J. Canada 40(2):164176. k
McWhirter JR, Chern J, Hutter JC. 1995. Oxygen Mass transfer coefficient
Transfer Fundamentals of Surface Aerators. Ind. Eng. kr non-dimensional oxygen transfer coefficient for
Chem. Res. 34:26442654. rectangular tanks
Metcalf & Eddy, Inc. 2004. Wastewater engineering: kc non-dimensional oxygen transfer coefficient for
treatment disposal and reuse. Tat McGraw-Hill, New circular tanks
Delhi, India. ks non-dimensional oxygen transfer coefficient for
Oldshue JY. 1985. Current trends in mixer scale-up square tanks
techniques, p. 309340. In Ulbrecht JJ and Patterson GK KLaT overall oxygen transfer coefficient at room
(ed.), Mixing of liquids by mechanical agitation. Gordon temperature TC of water
and Breach Science Publishers, New York. KLa20 overall oxygen transfer coefficient at 20C
Rao ARK 1999. Prediction of reaeration rates in square, L size of rectangular tank (L)
stirred tanks. J. Environ. Eng.-ASCE 125(3):215233. l length of the blade (L)
Rao ARK, Bharathi Laxmi BV, Subba Narasiah K. 2004. N rotational speed of the rotor with blades (1/T)
Simulation of Oxygen Transfer Rates in Circular n number of rotor blades = 6
aeration Tanks. Water Qual. Res. J. Canada P 2
power available to the rotor shaft (ML /T )
2

39(3):237244. P/V Power per unit volume (M/LT )

2

Rao ARK, Jyothish S. 1997. Oxygen transfer in circular PV = P/[Vg(gn)] = Non-dimensional power per
1/3

surface aeration tanks with and without baffles.

unit volume
Presented at the International Conference on Industrial
Q discharge of water being pumped by rotor
Pollution and Control Technologies, Jawaharlal Nehru 3
rotation (L /T)
Technological University (JNTU), Hyderabad, India, 17
ND /n, Reynolds number
2
R
to 19 November 1997. 3
Schmidtke NW, Horvath I. 1977. Scale-up methodology for V volume of water in an aeration tank (m )

39
 Rao et al.

W width or breadth of the surface aeration tanks

Ra armature resistance of DC motor
V1, V2 input voltage at no load and loading
conditions respectively
X = N3D2/(g4/3n1/3) = F4/3R1/3 = theoretical power per
unit volume parameter
q 1.024; constant for pure water used in equation 3
n
2
kinematic viscosity of water (M /T)
ra
3
mass density of air (M/L )
rw
3
mass density of water (M/L )
e energy consumption per unit volume

Recieved: May 25, 2005; accepted: February 1, 2007.

40