DESIGN OF A MIXING TANK

INDIVIDUAL PROJECT

presented to
Professor John Sheppard

by

Thomas Krause _
I •.D. ~~ 8 9 1 1 18 2

Faculty of Agricultural and Environmental Sciences

Department of Agricultural Engineering
Ma~donald Campus of McGill University
.ste-Anne de ·Bellevue, Quebec
Apr i 1, 1993
 ABSTRACT

Disposal of waste is a problem for many swine producers, since

they often do not have the necessary land base to dispose of the
waste. A possible solution, would be to us~ cement kiln dust to
-e nhance the s€ttlement of the manure. This would permit more
economical transportation of the solids and disposal of the
supernatant liquid on available land.

Some studies have been made on this subject. The purpose of

this study is to design a mixing tank, for future studies that need
to be done. The square-baffled tank and flat-bladed impeller
design has improved settlement of manure over the previously used
cylindrical vessel without baffles and propeller impeller.
Preliminary tests -suggest that, a ratio of 1:8 cement kiln dust to
manure, seems to be a promis-ing ratio. Of the mixing speeds
· investigated, 500 rpm -gave the ·best ~esults.
 ACKNOWLEDGEMENTS

I would like to take this opportunity to thank my wife Line

Cote, for her help, support and understanding.

I would also like to thank Dr. Suzelle Barrington for her

support and for giving me the latitude to act on my own in this
project.

F i n a 11 y , I wo u 1 d 1 i ke to thank Sams on A• So to c i n a 1 f or
building the apparatus and suggesting modifications where it was
needed.
 TABLE OF CONTENTS

Page
Abstract
Acknowledgements
Table of contents
List of figures and tables
Introduction 1

Literature review 3

Materials and Methods 8

Results and Discussion 11

Conclusions 14

References

Appendix A: Sample calculations

Appendix B: Graphs of settled sludge volume

Appendix c: Impeller des-ign

 LIST OF FIGURES AND TABLES .

Page

Figure 1 Power curve 5

Figure 2 Tank configuration 7

Table l Settled sludge volume 12

Table 2 Effects of mixing speed 12

Table 3 Comparison of systems l3

 /1

INTRODUCTION

Many swine producers do not have the necessary land base to

dispose of their waste. This has led to excessive doses of manure
spread on land and even dumping of waste into rivers.

The use of cement kiln dust, a waste product of cement

manufacturers, has been found to enhance the fertilizer value of
swine manure by adding calcium and potassium as well as by

stabilizing the waste, this preventing organic N- degradation into

NH 3 ·(Barrington and Mackenzie, 1989).

It has also been shown that the incorporation of cement kiln

dust to swine manures does improve their -sett 1 i ng and produces

supernatant liquids of lower N, COD and P levels (Barrington et

al., 1990). However, the study also reported high levels of

potassium in the supernatant liquids, limiting the possibility of

land disposal.

Further studies need to be done and literature on wastewater

treatment suggest that the mixing of the product into the

wastewater is an important part of the coagulation and flocculation

process.
 /2

In this previous study a cylindrical vessel without baffles

and a propeller impeller were used to mix the slurry with the kiln
dust. This equipment cause the formation of a vortex during
mixing, thus diminishing the ·effectiveness of the mixing.

The objectives of this project are to design a suitable mixing

tank for laboratory tests invol·ving the use of cement kiln dust for
the treatment of swine manures, to improve their settling. The
tank should be designed to permit it's use for both the rapid mix
operation and the subsequent slow mixing operation. Some
preliminary tests will also be done to study the · effects of mixing
speed on settlement . .
 /3

LITERATURE REVIEW

Coagulation is a process in which charged particles are

destabilized to permit their aggl-omeration into floes, which can be
removed by gravity (flocculation).

H.E. Hudson (1965) showed that rapid mixing during coagulant

addition prior to slow mixing was important for effective
flocculation and that the · floes formed following rapid mixing
contain appreciably more solids than those formed without a rapid-
mix operation.

o'Melia (1969) reported that insufficient mixing may result in

poor floc formation, while intense mixing for too long may break up
previously - formed floes causing restabilization of repulsive
forces.

Met c a 1 f &. ·Eddy ( 19 9 1 ) s tat e that as a genera 1 · r u 1 e , the h i g her

the velocity and the greater the turbulence, the more efficient the
mixing. They also stress the importance of restricting vortexing,
in which the fluid gains rotational momentum thereby reducing
turbulence and frequency of collisions between particles and
resulting in poor mixing. This can be inhibited by the use of
stators (baffles). Baffles restrict rotational motion and promote

vertical mixing.
 /4

In the design of a mechanical rapid mix reservoir, literature

agrees on the following:

better Per f o rman c e of a s q u a r e v e ss e 1 over that of a

cylindrical vessel (Letterman) ;
baffles improve performance - (Camp);
a flat bladed impeller performs better than a propeller
type impeller (Letterman);
coagulants should be introduced at impeller to improve
coagulation (Hudson, 1974).

~\>"' ~w~~Jl
In order to 1 imi t shear on ~lee-u--le polymers,
paddle tip speed should not exceed 5 m/s (Awwa, 1990).

Metcalf and Eddy (1991) suggest paddle tip speed of 0.6 to 0.9
m/s to achieve sufficient turbulence without breaking up floc
during slow mixing process.

Manure slurries are non-Newtoni an fluids. Conventiona l

equation from wastewater treatment literature for power requirement
and velocity gradient cannot be used because they assume Newtonian
fluids. Chen (1981) studied impeller power consumption in mixing

livestock manure slurries.

 /5

In this study Chen proposes a method of calculating the power

requirement, using the following equation.

Where p = power (W)

D = impeller diameter (m)
N = impeller rotational speed (rps)

Np = power number

c::
u.J LO SYW80t. TURBINE: OIAW£TER ClO T

~
lcml
10 .2 0.34
7. 6 0 26
6.-4 0.21

01
10 .ol 10J 10 4 1()6

RE'J'NOLOS NUMBER, NR•

. ·-

Figure l : POWER CURVE TAKEN F-ROM

CHEN (1981)

The power number is found -from Figure l which relates the

power number to the Reynolds number. Although this curve is for

beef cattle slurries, it is the only reference available. For the

purpose of this project, it will be assumed a valid approximation

of the power number for swine manure slurries.

 /6

The Reynolds number can be calculated using

Where p = fluid density, kg/m 3

D = impeller diameter, m

N = impeller rotational speed, rps

lle = effective viscosity, Pa.s

The effective viscosity can be found using

(Chen, 1981)

Once the effective viscosity and power requirements are known

for different mixing speeds, a velocity gradient (C) can be found
using the following equation, adapted from Metcalf & Eddy (1991)

Where p = power, w
-1
c = velocity gradient, s

11e = effective viscosity, Pa.s

V = volume of fluid
 TANK CONFIGURATION USED
IN THE DESIGN OF MIXER

Notes: 1. The agitator is a six-blade flat

~
.,.....,rank • turbine impeiler

nlsatfle 2.. lmpeller diameter, d1 • 1/'3

tank diameter
V
-,-.- -
r
----:-1
'\,
0. lmpel!er height from bottom,
H; • 1.0 impeller diameter
4. lmpe!ler blade width, q • 1/5
II Six·blade impeiler diameter

__ __
II flat t:lade
1 turbine
.._
- ·- ·-
5. lmpefler blade length. r • 1/4
Impel! er Ciameter
6. Length of impeller blade

i'lr~
mount~ on the central cisk •

Hr.. r12 .. 1/8 impeller diameter

7. Uqt:id heighL He.. 1.0 tank
\ I diameter
8. Number or batfle.s • .t. mct:nted
[ITS I ·I p17 vertic;.Jly at tar.X wall ar~d
I ~~-·-~It
r
e:xtencing from the tank bottom

- - I H
to above tl'le liquid surface
PS •
f.-;
I
I -
-.
--- . --.
--_1: '- l
9. 6aHia width.
diameter
1/10 tank

10. Cantral c'isk diameter, s • 1/4

tank C:amater
r------- o_ ______,_. I

Figure TAKEN FROM METCALF & EDDY

Figure 2 p·..f·ovides dimensions- .of tank components ahd impeller in terms of

rat i os w i t h r e s p e c t to tank d i ame t er • T h i s -sour c e was us e d i n· d i mens i on i n g
the tank and impeller. However some· modifications where made and will be
explained in the following -section.
 /8

MATERIALS AND METHODS

Construction of mixing unit:

The tanks dimensions were chosen to be 6" X 6" X 14" (152 X

152 X 356 nm) giving the reservoir a capacity up to 5 litres
(sufficient for laboratory use) and still leaving enough freeboard
to control splashes

The tank is made of 3/8" (95 rrm) acrylic and the baffles of
3/16" (5 mm) acrylic. This material provides sufficient strength
and permits observation of the fluid as it is mixing. Parts were
~oined using methyl chloride, ·which provides good bondage without
disturbing the materials transparency.

A c 1 ear an c-e o f l I 2 " ( l 2 . 7 rm1) was 1 e f t be t wee n the b a f f 1e s and

the tank bottom, to facilitate cleaning and permit the installation
of an outlet in the future.

The impeller was machined from stainless steel and the blades
were cut from a 24 gage stainless steel plate (3/8" or 9.5 nm).The
size of the impeller blades was chosen to be 3/4" ( 19 mm). This is
slightly larger than what is ·suggested by Figure l. According to
the reference the blade dimensions for a 6" ( 152 rrm) tank diameter
would have been 10.2 rrm X 12.7 nm or 0.4" X 0.5". This did not
leave enough blade surface for effective sweeping at low speed.
 /9

The shaft ( 3/8" or 9. 5 mm) was machined from brass. This

material was chosen because it is easy to machine and
. permi~ted a

tight fit of the impeller onto the shaft. The impeller was pressed

onto the shaft eliminating the need for set screw.

Due to the length of the shait, it was mounted on a pin at the

bottom of the tank in order to limit it's translational movement,
but permitting its rotational motion. A rubber coup! ing was
installed between the motor and the shaft to reduce vibrations and
correct for misalignment.

Preliminary tests:

Materials: - mixing unit

- 1 1 and 4 1 cylinders
- electronic scale
- electronic rpm meter
- stop watch -

Methods:

Swine manure was collected from the grower barn of the

MacDonald Campus farm. The manure was then diluted to 6% T.S.

sin~e rheological data was available for swine manure at 6% T.S.

( Zhang and Day, 1990) .

 /10

Batches of 3 1 of manure were mixed with cement kiln dust.

Ratios of cement kiln dust to manure of 1:2, 1:4, 1:8 and 1:12 were
tested. Each batch underwent a rapid mix period of 2 min followed
by a 10 · min period of -slow mi-xing. Then the mixture was

transferred to cylinders were it was allowed to settle for 48 hrs.

For testing the effects of mixing speed on manure settlement,

a ratio of cement kiln dust to manure of 1:8 was used.

An electronic rpm meter cwas used to measure the shaft

velocity. Speeds of 500, 750 and 1000 rpm were tested using the
same procedure outlined above.
 /11

RESULTS AND DISCUSSION

The mixing tank functions well~ ft promotes vertical mixing and

even aerates the fluid.

----------
at high speed it is evident that it creates great turbulence and
-
I
-
Re - ·
7

The results of different treatments with cement kiln dust are

shown in Table 1. Although -they are preliminary results, the
results contain some interesting information. It is evident that
greater the ratio of cement -kiln dust, the greater the volume of
settle sludge. More importantly the ratio of 1:8 seems most
promising. It has 23% more -settled sludge than the ratio of 1:12
and only 4% less than that of the 1:4 ratio. This is significant
because between ratios 1:8 and 1:4 the settled volume only
increases by 5% whereas the ratio is doubled. There is an obvious
reduction in the effectiveness of the treatment between 1:8 and
1 : 4.
 /12

TABLE 1

Settling volume of treated swine manure

RATIO AVERAGE SETTLED VOLUME

Dust: manure (%)

Zero 46

l :2 88

l :4 77

1:8 73

1 : 12 56

Ratios were measured on a wet weight basis rapid mix operation was
done at 1000 rpm for 2 min.

TABLE 2

Effects ~f mixing speed on settling volume

I
R.P.M. AVERAGE SETTLED VOLUME
(%)

500 66.5

750 61 • 5

1000 60

A ratio of 1:8 cement kiln dust to manure was used for this test.
 /13

The effects of mixing speed on the sludge settling volume is

shown in Table 2. The results indicate that better settlement is
achieved at 500 rpm than at higher speeds. At higher speeds (750
and 10 0 0 r pm) a l aye r of f l oa t i ng mat er i a 1 was p r e s en t af t er
settlement occurred. Thi s was mo s t e v id en t at 10 0 0 rpm we r e a
layer of 1 cm on average was present, thus -reducing settled volume.
The manure is very high in organic matter and at high velocity the

1 shear stress is probably too high. Also the retention time of 2

min. under such vigorous mixing, is perhaps -too long. This
corroborat-es what O'Melia ( 1969) stated, intense mixing for too
long may break up floes causing restabilization.

It should be noted that experiments done with this mixing unit

have consistently given a greater sett 1 ed volume than previous
experiments done with a cylindrical reservoir without baffles and
a propeller type impelle-r (Barrington, 1990). The comparison of
the two sys terns are shown it Table . 3.

TABLE 3

RATIO SETTLED VOLUME COMMENTS

1: 4 77% square tank baffles and flat

bladed impeller

1·: 4 47% cylindrical tank and

propeller impeller

1: 8 73% square tank, baffles and flat

bladed impeller
 /14

CONCLUSIONS

The results of preliminary tests can be used as a guideline

for future mo-re in depth studies on cement kiln dust. A ratio of
1:8 cement kiln dust to manure seems promising. This ratio and
others near it should be the object of further studies. A mixing
speed of 500 rpm with a retention time of 2 min yielded the best
results in this study. However retention time versus mixing speed
should be investigated further in order to find the best
combination for treatment of swine manure with cement kiln dust.

The superiority of the square baffled vessel using a flat

bladed impeller has clearly been shown. Therefore it can be said
that the objectives of this project have been met.
 REFERENCES

AMIRTHARAJAH, A. "Design of Rapid Mix Units", in R. L. Sanks

(ed.), Water Treatment Plant Design, Ann Arbor Science, Ann
Arbor, Mich., ( 1978).

AWWA, ASCE. "Water Treatment Plant Design- Second Edition",

American Water Works Association, Inc., (New York : McGraw-
Hill, 1990).

BARRINGTON, S.F., et al. "Cement Kiln Dust To Improve Swine

Manure Settling", Proceedings of the Sixth International
Symposium on Agricultural and Food Processing Wastes, ASAE
05-90, p.259-268 (1990).
CAMP, Thomas R. "Floc Volume Concentration", J. Am. Water Works
Assoc. 60:656-673 (1968).
CAMP, T. R. AND P.C. Stein. "Velocity Gradient and Internal Work
in Pluid Motion", J. Boston Soc. Civil Eng. 30:219-237 (1943).
CHEN, Y.R. "Impeller Power Requirement in Mixing Livestock Manure
Slurries'', Transactions of the ASAE, 24(1): 187-192 (1981).
HUDSON, H. E., Jr. ( 1965) "Physical Aspects of Flocculation",
· Proceedings Awwa Seminar on Upgrading Existing Water-Treatment
Plants, 157-164 (1974) .
HUDSON, H. E. , J r . · and J. P. Wo 1 f ne r . "Design of Mixing and
Flocculating Basins", J. Am. Water Works Assoc. 59:1257-1267
(1967).
LETTERMAN, et al. "Influence of Rapid-Mix Parameters on
Flocculation", J. Am. Water Works Assoc. 65:716-722 (1973).
METCALF & Eddy. "Wastewater Engineering - Treatment, Disposal,
Reuse", 3rd Ed., ·New-York, McGraw-Hi 11 ( 1991).
O'MELIA, C.R. "A Review of the Coagulation Process", Publi-c Works,
lOO, 87-98 (1969).
PEAVY et al. "Environmental Engineering", New-York, McCraw-Hill
(1985).

SIEVERS, D.M. "Rapid Mixing Influences on Chemical Coagulation of

Manures", Biological Wastes 28:103-114 (1989).
ZHANC, R.H. AND D.L. Da-y. "Rheological Properties of Fresh Swine
Manure .slurry'', Proceedings of the Sixth International
Symposium on Agricultural and Food Processing Wastes, ASAE 05-
90, p. 541-547 {1990).
A P .P E N D I X A

.. . SAMPLE.. CALCULATIONS .
 SAMPLE CALCULATIONS

Maximwm and minimum allowable mixing speeds:

- Maximum paddle tip speed of 5 m/s (Awwa, 1990)

- Imp e 11 er d i ame t er = 6 3 • 5 X 1 0- 3 m
- Impel1er ~adius = 31.75 X 10- 3 m

v = w.r
w = v/r = 5 m/s · 31.75 X 10- 3 m= 157. -48 rad/s

Therefore Nmax = 1500 rpm

-For flocculation process, suggested paddle tip speed--- from 0.6 m/s

to 0.9 m/s (Metcalf & Eddy, 1991).

3
w = vIr = 0 •9 m/ s · 3 1 • 7 5- X 1 0- m -= 28 • 35 r ad Is

Therefore for flocculation Nmax = 271 rpm

Nmin = 180 rpm

Calculation of power requirement at 1000 rpm (16.7 rps):

- Effective viscosity at 16.7 rps

lle =C (Kl N)n-1

Where C = 0.035 and n = 0.76 (Zhang and Day, 1990)

K1 = 10.3 (Chen, 1981)

lle = 10.175 X 10- 3 Pa.s

- Reynolds number becomes

NRP. =--t -·9 o2 N } I lle-

3
Where p = 1000 kg/m

D = 63.5 X 1 o- 3 m

N = 16.7 rps
lle = 10.175 X 1 o- 3 Pa.s

NRE = 6618

From Figure 1, Np - ~ 2.8

- The power requirement

p = p o5 N3 Np

p = 13.5 w
- The velocity gradient

Where

Therefore C = 616 s-l

APPEN - D . ._IX B

GRAPHS .OF .SETTLED VOLUME

 SETTLEMENT
/o VOLUME
0

100 .

80

·...•.

60
·:;;

Series 1

. 40

20

0
ZERO 1:12 1:8 1:4 1:2

tETENTI ON TIME: 48 HR.

EFFECTS OF MIXING SPEED
0
/o VOLUME SETTLED

70

60

50

• _.,. •, •I

40
. '

Series 1 I
30

500 750 1000

RETENTION TIME: 48 HR.

A P P E N D I X C

_IMPELLER DESIGN_
 2 1/2"
'~~--~--------~>

1 1/2"

---- ;s.

3/4"
. . . _ _.......
~~~- - ,..,,~~1
_
3/4"

60
DEG
• • • • , • • \1.. ., •••••• '.~;·~·t~··· ... ·~· ··. -~ •. ~ .. ·-·.~~