McMaster University

DigitalCommons@McMaster
Open Access Dissertations and Theses

Open Dissertations and Theses

2-1-1974

Performance of a Multistage Fluidized-Bed IonExchange Contractor with Fluid-Diode

Downcomers
Conrad Yiu-Fook Law

Follow this and additional works at: http://digitalcommons.mcmaster.ca/opendissertations

Part of the Chemical Engineering Commons
Recommended Citation
Law, Conrad Yiu-Fook, "Performance of a Multistage Fluidized-Bed Ion-Exchange Contractor with Fluid-Diode Downcomers"
(1974). Open Access Dissertations and Theses. Paper 554.

This Thesis is brought to you for free and open access by the Open Dissertations and Theses at DigitalCommons@McMaster. It has been accepted for
inclusion in Open Access Dissertations and Theses by an authorized administrator of DigitalCommons@McMaster. For more information, please
contact scom@mcmaster.ca.

, .,

PERFORMANCE
OF

A MULTISTAGE FLUIDIZED-BEQION-EXCHANGE CONT ACTOR

WITH
FLUID-DIODE

DOWNCOMER~

By
CONRAD YIU-FOOK LAW

""
A Research
Report
Submitted (to the School of Graduate Studi~s
in Partial Fulfilment of the Requirement>
for the Degree

Moster of Engineering
;)

I.

McMaster University
February, 1974

Conrad Yiu-Fook Law

1974

..

MASTER OF ENGINEERINJ 0974)

{Chemical Eng ineering)\,-

McMASTER UNIVERSITY
Hamilton, 9ntario

TITLE:

Performance of a Multistage Fluidfz.ed-Bed lon-Exchange Contactor

with Fluid-Diode Downcomers

AUTHOR:

Conrod Yiu-Fook Law, B.Eng.

SUPERVISOR: Dr. L. W. Shemilt

NUMBER OF PAGES: xiv, 203

ii

(McN\aster University)

. ABSTRACT

A study was carried out on the operational and mass-transfer performance of

10 multistage fluidiz:~d-bed continuous countercurrent ion-exchdnge contactor with

fluid-diode downcomers
No +/H+ wi th Dowex-

For the mass-transfer study, Jhe ion-exchange system

W-x8 resin was used. The tobl solution concentr6tion used
L--

'.

waS 0.06 meq/ml .

flui~-diode downcomer design~gnificantlY reDuces the bypassing of liquid

J
h-

up the do Nncomer.

Easy start-up and shut-down were observed and the resin was

ev'enl~stributed on each mass-transfer stage.

displace~t, weir height,

The effects of fluid-diode lateral

woir shape and resin particle size on the operational

performanc) (in terms of liquid-bypassing and resin fluidization) of the solid-liquid

--_./
contactor were investigated.
A one-quarter repl icate. of a factorial experimental design to 7 factors at 2
levels successfully and effi~iently revealed the main factors and interactions which
significantly affect the overall stage efficiency of the ion-exchange-column
operation. The significant main factors are resin feed-rate, process solution

noi,,. ."

rate, weir length and weir height. The significant interactions are between resin feedrate and weir length, process solution flow-rate and weir length, and weir height and.
weir length.
The mass-transfer study has shown thot the solid-liq~id contact time in the
contactor determines the overall stage efficiency. The composition dependence of

iii

the

a-transfer coefficie'nts in both the solid and liquid sides also pip)" a role In
I

Ing the overall mass-transfer.

The rela tive specific volumes calculated fro,n

justify

fUr"thc dev 110pment of the present contactorin the consideraiian for in ustrial
scale applications. The present type of contactor is simpl.e in design and requires

simple control strategy, providing operational slability, and large scrlf"aand liquid
handling capacity.

/
r

.,

IV

,-

ACKNOWLEDGEMENTS

The author wishes to express his sincere appreciation to his research

supervisor, Dr. L. W. Shemil t, Dean of Engineering, for his assistance a l1 d
co-operation during the cOurse of the completion of this study.
Thanks are extended to Dr. A. V. Bapat for contributing same of his ideas to
the present study.
Thanks

areals~nded to Mr.

G. De Bare of the Facul ty lvIochine Shop

and Messrs. J. Newton and B. Dunn of the Chemical Engineering McrctN,ne Shop for

their workshop assistance.

'

Special thanks are extended to Mrs. I. Underell for her

secretarial help.
Finally, the author wishes to thank his wife, Stella, for her help in preparing
the draft for this thesis and her patience and encouragement throughout the COlXse
of the research program.
j

""

~ABLE

OF CONTENTS

LI ST OF FI GURES

IX

LIST OF TABLES

XIV

Ch"ptcr I

INTRODUCTION

Chapter 2

LITERATURE REVIEW
Fluid Diode
2.2 Continuous Countercurrent Ion Exchonge

2.1

Ch"pter 3

Chapter 4

THEORY
3.1
Fluid Diode
3.2 lon-Exchonge Resin
3.3 Ion Exchonge
3.4 Fluidized-Bed Ion Exchonge
3.5 Contil'uous Countercuorent Mult,stage Ion Exchange
3.6 Factorial Experimental Design
EQUIPMENT DESIGN AND APPARATUS
4.1
lon-Exchange Column System
4.1.1
General
4.1.2
lon-exchange column
4.1.3
Liquid feed system
4.1.4
Resin feed system
4.1 .5 ~ Resin sampl ing system
4.1.6
Exhausted resin collecting system
4.2 Resin Drying System
4.3 Resin Regeneroting and Woshing System

Cbptcr 5

FRACTIONAL FACTORIAL EXPERIMENTAL DESIGN

Ch"ptcr 6

EXPERIMENTAL PROCEDURES
6.1
An"lytical Techniques
6.1.1
Displacement of hydrogen-ions from resin
by sodium chlorldc solution
6.1.2
Concentration af sodium chloride solution
for resin sample titration
6.1.3
Determination of resin cap"city

vi.
l

10
13

~~21
23

28
28
30
31
32
32

c;J
35

39
39
40

..-'/

.:Y

""- ........:

-;
6.1.4

")

...

<'"

6.2

6.3
6.4
6.5
6.6
6.7
6.8
6.9

Chapter 7

Chopter 8

D2terrnination of chloride concentration

in solution
6.1.5
Deterrn ination of hydrogen-ion concentration
in solution
6.1.6
Determ ination of pocked resin bed voidage
Calibrations
6.2.1
Process solution feed rotameter
6.2.2
Resin feed rotameter
6.2.3
Variable speed Masterflex tubing pump
6.2.4
Resin feed
6.2.5
Water in resin feed
6.2.6
Chloride-ion concentration
Resin Drying
.
;
Resin Size Scr-eening
Resin Regpneration
Resin Wpshing
Factdrs~ Levels for the Fractional Factorial Experimental

I
Deslgnv
Operational Performance of the lon-Exchange Column
Moss-Transfer Experiment

DATA REDUCTION
7.1
Mass-Transfer Experiment
7.2 Fractional Factorial Experimental Design
RESULTS AND DISCUSSIONS
~./
FI uid ~iodes os Downcomers
,I ,
8.1.1
Effects of the fluid-dioJe designs on resin flow
8.1.2
Effects of fl u id-diode lateral displacement on
resin flow
8.2- Operational Performance of the lon-Exchange Column
8.2.1
Effe~ts of resin and liquid flow-rates
8.2.2
Effec ts of resin particle-size distribution
8.2.3
Effects of fluid-diode lateral displacements
8.2.4
Effects of weir height
8.2.5
Effects of conical weir
8.2.6
Effects of cylindrical weir
8.3 Fractional Factorial Experimental Design
8.4 Effects of the Main Factors and the Significant Interactions on the Moss-Transfer Performance
8.4.1
Resin feed-rote
8.4.2
Process solution flow-rate
8.4.3
Weir height
8.4.4
Weir length

41
42
42
43
43
43
44
,~

44
45
45
46
47
47
50
50

54
54

8.1

vii

56
57

59
61
61
62
62
64
65

66
68
69
70

804.5
804.6

8.5
8.6

~-E!.7

8'.8

Chapter 9

Fluid-die>ee lateral displacement

Downcomer cleara9ce
Resin particle size "',
~A.7
Resin
feed r~te - weir length interaction
~,A.8
804.9 Process solut~on flow rate - weir length
interaction
804.10 Weir height - weir length interaction
Effects of ~ Initia'l Sodium-Ion Concentration in the
Process FeedSofiJ"tion on Mass-Transfer Performance
Effects,of Weir Shapes on Mass-Transfer Performance,
Mass-Transfer Performance Factor
Comparison of the Performance of Different lonExchange Contactors

CONCLUSION

71

72
72
73

74
75
76
77
79
80

84

NOTATIONS

87

REFERENCES

90

APPENDICES
A

C
D

E
F
G
H

Equipment Design and Appora tus

Experimental Resul Is
Operational Performance Diogroms
Miscelloneous Figures
Factorial Experimentol Design
Half-Normal Plot
Computer Program for Mass-Transfer Experiments
Computer Program for Fractional Factorial
Experimental Design
\

viii

93
110
148
156
160

172'
176
192

-LIST OF FIGURES

A.I

Theoretical contour of cusp-cavity shape

94

A .2

Modified cantour of cusp-cavity shape

94

A .3

Flo

95

patte

- low pressure-drop direction

AA

w pattern - high pressure-drop direction

95

A.5

sp-cavity desl n

95

A.6

lon-exc

96

A.7

column

97

A.8

Sec tions 7 and 8 of ion-exchange column

98

A.9

Liquid distributor plah!----

99

A.IO

R('s i n sampler

99

A .11

Resin sampler's plun

A .12

Sec tion I of io -exchange column

100

A .13

Tray design wi h 4" straight weir

101

A.14

r'
Tray design wi

3 3/4"straight weir

102

A.15

Tray design with:3Y4" straight weir

103

A.16

Tray design with conical weir

104

A.17

Tray design with cylindrical weir

105

A.18

Typical f1uid-diode downcomer

106

A.19

Fluid-diade downcomer in section 6 of ion-exchange column

107

100

IX

A.20

Resin drying system.

IDS

A.21 --

Resin regenen:,tion and washing system.

109

B.I

Change in the pH value of mixtures of resin and sodium

chloride solution wi th time.

III

B.2

Equilibrium hydrogen-ian concentration versus initial

sodium-ian concentration in solution.

112

B.3

Titration curves.

113

B.4

Titration curves.

B.5

Measured electrode potentials versus hydrogen-ion

concentrations in solution.

115

B.6

Measured electrode potentials versuS pH values of solution.

116

B.7

Calibration curve of the rotameter for process solution feed.

117

B.S

Cal ibration curve of the rotameter for resin feed

liS

B.9

Calibration curve for Masterflex tubing pump's speedcontroller.

119

B.IO

Cal ibration curves for resin feed

120

B.II

Volume fraction of resin in feed versus resin-feed

rotameter's reading

121

114

.~

B.12

( d l ibration curves for water flow in resin feed.

B.IV Calibration curve for chloride-ion concentration

122
123

measurements

Change in resin capaci ty during regeneration

123

B.15

Change in PJtHlueof liquid effluent with time

during resin
h' g.

124

B.16

Change in pli> alue of liquid effluent frOm the ionexchange column as a function of time.

125

B.17

B .18

.;

Ranges of liquid flaw-rat~ which gave "perfec t'''', opera tiona I

,behaviour in the ion-exchange column'
'

128

Effects of resin feed-rate

129

B.19

Effects of resin feed-rate an the operating line

130

B.20

Effects of process solution flow-rate

131

B.21

Effects of process solution flow-rate on the operating line

132

B.22

Effects of weir height

133

B.23

Effects of weir_ height on the operating

line
-r_.

134

B.24 a

Effects of resin feed-rate on the overall stage efficiency

at different weir lengths

135

B.24b

Effects of resin feed-rate on the composition of resin

effluent

136

B.24c

Effects of resin feed"rate on the composition of liquid

,effl uent

136

Effe~'is of process solution flow-rate on the oyerall stage

efficiency at different, weir lengths

137

B.25 b

Effects of process solution flow-rate an the composition

of resin effluent

138

B.25c

Effects of process solution flow-rate' on the composition

of liquid effluent

138

B.26 a

Effects of weir height on the overall stage efficiency at

different weir lengths

139

B.26b

Effec ts of weir height on the composition of resin effluent

139

B.26c

Effects of weir height on the composition of liquid efflue!",t

B.27a

Effects of initial sodium-ion concentration on the ,?verall

stage efficiency

140

B.27 b

Effects of initial sodium-ion concentration on the operating

line with solution flow-rate at 40.0 ml/sec

141

B.25 a

'.

XI

'.

140

B.27c

Effects of ini tiol sodium-ion concentration on the operating

line
with. solution flaw-rate at 79.2 ml/sec
,

142

B.27 d

. Effects of ini tial sodium-ian concentration on the

c.ompo'sition of resin effluent

143

B.27 e

Effects of initial sodium-ion concentration on the

composition of liquid effluents

143

B.27f.

Effects of initial sodium-ion concentration on the overall

ra te of ion exchanged

144
"

B.27 g

Effects -of initial sodium-ion concentration an the fraction

of input sodium-ion adsorbed by the resin phase

145

B.27h

Effects of ini tial sodium-ion concentration on the fraction

of input hydr03en-ion extracted by the liquid ptfuse

145

Effects of the resin feed-rate on the overall stage efficiency

for different types of weir

146

B.29

A log-lag plot of total flow velocity (u +v) versus f:HU or

HETS'
. .

147

C.I

Straight weir - effects of resin and I iquid flow-rates

149

C .2

Straight weir - effects of resin particle-size distribution

149

C.3

Straight weir - effects of fluid-diode lateral displacement

150

CA

Straight wei; - effects of weir height at 1;20 inch fluiddiode lateral displacement

150

C.5

Straight weir - effects of weir height at 3/20 inch fluiddiode lateral displacement

151

C.6

Conical weir - effects of resin and liquid flow-rates

151

C.7

Conical.weir - effects of weir d.iameter

152

C.B

Conical weir - effects of resin particles-size distribution

152

C.9

Conical weir..- effects of fluid-diode lateral displacement

153

... ,

xii

C.IO

Conical weir -effects of weir height at 1;20 inch fluiddiode lateral displacement
..

C.II

Conical weir - effects of '(Veir height at 3/20 inch fluiddiode ra~r,?1 displacement

154

C .12

Cylindrical weir - effects of resin and liquid flaw-rate

.154

C .13

Cyl indrical weir - effects of resin particle-size distribution

155

D .r

A plot of pressure-drop ratios (h igh/I ow) versus Reynold

numbers with data from the work of Egan (7)

157

D.2

A plot of ini tial sodium chloride concentration in solution

versus ihe equil ibrium constan\ for hydrogen sodium
e?,change on Dawex-50-x8 resin

158

D.3

Continuous countercurrent multistage operation

158

DA

Equi librium diagram fpr the ion-exchange system showing

the o~g line and the number of theoretical stages
for a typrcal contihuous countercurrent operation

159

LI

Confounding pattern for the 2-factor interactions

167

E.2

Hal f-norrr~al plot for the absolute overage' effects

171

Standardized half-normal plot for the absolute aV,~rage

effects

171

G .1

Flow, chart of the computer pro~am for analysing the

results of mass-transfer experiments

177

H.I

Flow chart of the computer pro3ram for onblysing the

results of faelorial experimental design

193

153

.. E.3

'-..,

"

xi~i

!,

"

LIST OF TABLES
Page
B.I

Opera tiona I behaviour of the ion-exchange column at

different liquid flow-rates (-20+30 mesh resin used)

126

B.2

Operational behaviour of the ian-exchange calumn at

different liquid flow-rates (-35 +50 mesh resin used)

127

E.I

Effect matrix generated for a 2 3 factorial design with

foe tors: A, Band C

161

E.2

Yates' technique for generating the total effects for

a 2 3 factorial design with factors: A, Band C

161

E.3

Effect matrix generated for a

x 2 3 factarial design
with A, Band C as the factors and ABC the defining
can trast

E.4

All the possible sets of defining cpntrasts for splitting'

the 27 factorial design (with Factors A,B,C,D,E,F,
and G) into four equalblocks

E .50

Treatment combinations of the principal block for the

'I4-replicate of a 27 factorial design

E.5 b

The four blocks of effects for the

x 27 factorial
design with factors A,B,C,D,E,F and G and defining
contrasts ABC DE, ABCFG and (DEFG)

165

E.6

Alias groups in the 1/4 x 27 factorial design with

factors A,B ,C,D,E,F and G and defining contrasts
ABCDE, ABCFG and (DEFG)

166

E.7

Order of performing the 32 experimental runs for the

1/4-replicate factorial design

168

E.8

Analysis of variance table

169

'12

'14

xiv

162

163

164

I. INTRODUCTION

/
The study of ion-exchonge is over a century old and the technology has been
opplied to industry for more thon sixty years. For many years, ion-exchange WaS
commonly used for water softening. The introduction of synthetic resins of
reliable physical and chemical properties ha~ greatly broadened the application of
ion-exchange. Various successful applicotions of the ion-exchange process are well
known in demineralisation of boiler feed-water, in ~'ffluent treatment, in
demineralisation of sugar solution and in hydrometallurgy and in the purification of
pharmaceuticals. (
The fixed-bed ion-exchange technique has been used for more than fifty
years and is essentially unchanged except for qutomation. The fixed-bed process
,

operates on a semi-batch basis and is easy to achieve on a large scale. The

advance in this technology has been confined almost entirely to the chemistry of
the process, such as through the development of new-and better ion-exchange
resins. Some of these resins approach theoretical limits with respect to their
chemical capacity, efficiency and stability (I). Further improvement on the
performance of ion-exchalJge technology will be in the development of moc:e
efficient solid-liquid contactors.

\,

In addition, there is still need to solve the

problem of a fixed-bed ion-exchange process economically for treatment of a

continuous liquid-feed of high concentration where a short cycle-time is required.
The developme~t of continuous countercurrent contacto,( for ion-exchonge processes

..\..... J

has been emphasized in recent years. Various cantactor designs employ solid resin
(

"

flaw countercurrent to process solution flow in the form of a moving packed-bed, a

fluidized bed or a pulsed bed (3, 4, 5). Progress here has also been encouraged

by the development of synthetic ian-exchange resins. Continuous countercurrent

ion~nge processes have advantages, inc luding maximum utilization of the"

ion-exchange resin inventory during the exhaustion and the regeneration cycles.
Further, the resin inventory is low, and equipment size and floor space requirements are reduced. The regenerant and wash water requirements are also lowered,
indicating a reduction in waste disposol expenditure. The continuous mode of
operation requires no shutdown For regeneration, and automatic control can be
implemented to adjust the resin and liquid flaws in obtaining effluents of desired
composi tions.

Hence overall lobaur requirement may be reduced. In general, the

continuous countercurrent ion-exchange processes can ofFer savings in capital and

operating costs per unit output, above some threshold capacity, when compared
ta the con"ventionel Fixed-bed operation. For the treatment of dilute solutions

(0.0005 gm/cc), this threshold has been put between 6000 and 30,000 gal/hr (15)
which varied For diFFerent operable contactor designs. A large number of cantactors
have been developed but very few, have become commercially successFul (2).
The pr~sent research is on the development of a truly continuous countercurrent ion-exchange c6htactor. This

i~ a multistage Fluidized-bed contactor.

The resin flows under gravity from one stage to the next through a fluid-diode
downcomer. Unlike other countercurrent ion-exchange contactors, this one handles
truly continuous solid and liquid flows within the optimal range of operation
(

conditions. The development of the present conmctor originotes rrom the research
of Souhrada (6) who has applied the controlled cycling technique in the design of
o multistage fluidized-b.ed ion-exchange column with downcomers similar to those
in

gos-liquid canmctor.

Egan (7) reported on the application of the fluid diode

in designing a downcomer to eliminate the problem of liquid bypassing. Chaos (8)

investigation has extended to'tlemonstrating the ability of the contactor to perform
under a continuous liquid-flow. Although these intermittent and continuous liquid-

,
flow operations showed approximately the same mass-transfer performance, the
simplicity of the continuous liquid-flaw operation appears economically more
advantageous.
The purpose of the present study is to extend the investigation an the
operational and mass-transfer performance of the multistage fluidized-bed contactor
with fluid-diode downcorriers when operated under continuous liquid-flow. A
further aim is to compare the performance (in terms of the chemical efficiency and
the volumetric throughput) of this ion-exchange contactor with others, including
some which have been used industrially.

2 . LITERATURE REVI EW

2.1

Fluid Dicxle
Consideroble interest in fluidic (fluia logic) systems and devices hos arisen,

largely fro,n the work done ot the Harry Diamond Laboratories (14). These devices
are generally characterized by having no mechanical moving parts and rely on the
inter"actian of entering fluid streams with each other and with shaped walls to
produce the desired response in the output fluid stream (9, 10, II).

The fluid diode was invented by Telse (12) in 1920. The device consisted of
a channel with many branching and returning loops which behaved I ike an electrical dicxle in offering for more resistance to flow in one direction than to flow in the
reverse direc tion.
Ringleb (13) investigated the flow patterns of on incompressible inviscid fluid
in various single cusp-cavity shapes. The general theory of standing vortices based
on potential-flow consideration was applied to the computation of velocity
distributions along the walls of various cusp-cavities. Ringleb also suggested the
general form of a mapping function to gererate these cusp-cavity shapes.
Thompson (14) has utilized the principle of the fluid diode to develop an
intermittent-flow countercurrent liquid-liquid contactar. The contactor makes use

of the flow-interference effects of the designed fluid-diode structure to convert an

01 ternating applied pressure into a net countercurrent motion of the two liquid
phases.

"

Egan (7) implemented the idea of the fluid diode in his design and
construction of a new downcomer structure to overcome the undesirable downcomer
liquid-bypassing in.the intermittent-flaw multistage fluidized-bed ion-exchange
-::";..

column developed by Souhroda (6).

t,

Egan tested three cusp shapes for the fluid-

diode dawncomer design. The design is essentially similar to that used in

Thompson's fluid-diode contactor but each fluid-diode downcomer consisted only
of twa or three pairs of cusp-cavities.

2.2

Continuous Countercurrent Ion Exchange

A large number of continuous countercurrent ion-exchange (CCIX) contactors

. have been proposed and tried in pilot-plant scale. However few of the designs
have'
to over. met with any degree of ".commercial success (2). The main problem
.
come in the designs is to achieve smooth, even and continuous counterflows of
solid resin and process solution where there is a small density difference between the
twa phases. Various ingenious methods employed to overcome this problem include
solution downflow, resin agitation and vari~us other mechanical and hydraulic
techn iques.
Various mixer-settlers have been patented for CCIX processes. The
horizontal mixer-settler system patented by Marris (16) was tested as a continuous
countercurrent ion-exchange contac tar. Hiester (17) has developed a mixdsettler device using air agitation in the mixer cells. Other mixer-settler systems
have been examined by Iwasyk and Thodes (18) who claimed favourable results for
these contactor systems in the extraction of phenol by ion-exchange.

There are some unusual designs in the mechanical transport af the ionexchange material continuously and countercurrently through the liquid phase.
These are McCormack and Howard's (19) endless tube of cloth filled with resin,
Muend'el and Selke's endless belt of phosphorylated ion-exchange cotfon and the
rotating helical tubes of Miles (21) and Schulman (22). These designs were only
for laboratory operations.
Weiss, McNeill and Swinton (23) have adapted the ore dressing operation
af jigging and developed multideck jigs for CClX, The authors felt that the large
capacity, ruggedness, flexibility and ease of mointenance of the contactor more
than balanced the lower efficiency.
Various contoctor designs have employed the moving fixeil-ll,,-d technique in

" exhaustion, regeneration and washing stoges. Some

the transfer of resin through the
of these designs have become very succe6sful in industrial applications. These
systems use the principle of periodically removing a portion or slug of exhausted
resin from one end ofp fixed bed, and replacing it with an equal volume of
regenerated resin at the other end. Such systems are usually not truly continuous
since the flow of process solution is momentarily stopped during the transfer of resin
in and out of the column. Peeblesand co-workers (24) hav!! developed one such
system where a single colum~ of resin is divided into zones for treatment, rinsing
and regeneration, and resin is pumped from' the bottom to the top of the column
forming a complete loop. A serious difficulty with this design is the'balancing of
differential p!'essures within the unit.
One of the most successful moving fixed-bed designs is the Higgins contoctor

(25) which is also a single-loop system. The resin is transferred around the locp by
regular hydraulic pulses. Further development of the process by Higgins and

Chopra (26) at the Chemical Separations Corporation results in the addition of the
"down flow" design to the single-loop system and has enhanced its versatility as a
process in both water and chemical process demineralisation applications. The
process is now marketed as the Chem-Seps CCIX contoctor. The Chem-Seps process
is claimed to require a resin inventory as low as 15% and rarely over 30% of that
required for conventional fixed-bed systems. The countercurrent flow efficiency
of this continuous process has also given a reduction of regenerant consumption by
I

onehalf and the rinse-water requirement by one-tenth. The Chem-Seps process

also has a lower equipment cost than the conventional fixed bed process. The
comparison is based on the treatment of a feed flow of 380,000 U.S. gals/day and
a removal of TDS from 1800 ppm to 5 ppm.
Ij

The Asahi process and the Permuitt CCIX systems employing the moving fixedbed technique have also been successful in large scale industrial operations
throughout the world (27). Levendusky (3) recently described the Grover

c.1.

process which also employed the moving-bed and intermittent-flow techniques. The
process has been successful in pilot-plant scale after eight years of research. The
design has the versatility in controlling the adsorption and desorption and washing
operations separately. Another version of the samo design can handle a mixedresin bed. The new process is expected to challenge conventional fixed-bed and
existing CCIX processes.. ,
Recent research has been in the development of multistage f1uidized-bed

contactors for CCIX. The fluidized-bed ion-exchange contactors ma~ have a lower
chemical efficiency than that of fixed-bed and moving-bed contactors. However,
the reduction in capital and operational costs due to simpler design requirements
in fluidized-bed contactors may compensate Tor the slight loss in efficiency. The
mavement of resin from one stage to anather countercurrent to the solution Flow has
been achieved in a variety of ways. The horizontal contactcr design of Cloete and

Streat (4) uses a momentary reversal of liquid-flow. Grimmett and Brown (5) useJl
,

the pulsing technique to assist the downward movement of resin through the tubular
down~cmers in their vertical mul.thtage fluidized-bed column. Slater (28) has

studied a similar pulse column. Church and Turner (29) developed and studied a

16 plate fluidized-bed column with stage and downcomer structures similar to those
of a sieve-plate distillation column. Results indicafed that the column waS relatively
easy to operate at steady state but there were difficulties in starting up and
allocating equal amounts of resin in each stage.
The principal difficulty in stagewise contactor columns with downcomers has
been the prevention of Iiquid bypassing up the downcomers. Externally controlled

rl

'. ,

Fluidized seal 'Pots under the downcomers (30) or, rubber flaps or valves in the downcomers (31) ta reduce liquid bypassing have been suggested. However, individual
seals of this nature cannot be regulated to give equal solid flow from each stage
without a ccmplex control system.
Souhrada (6) h:ls applied the principle of controlled cycling to achieve a
stable operation in a CCIX column similar to that of Turntr (29). The intennittent
flaw of liquid through the mul tistage column did provide easy start-up and shut-

dawn, and increased the stability of the steady state operation.

However problems

were still encountered in liquid bypassing up the downcomers resulting in a lower

contacting efficiency.
Egan (7) applied the fluid dynamic properties of a fluid diode in designing
a new type of downcomer for.the intermittent-Flow multistage fluidized-bed column
of Souhrada. The Fluid-diode downcomer succeeded in reducing the undesirable
li"quid bypassing. Cha (8) has showed that the multistage Fluid-bed column can be
operated satisFactorily not only by anintermittent but also by a continuous liquidFlow. The mass-transfer performance of the two modes of operation WaS shown to be
approximately the same. However, the. t~uly continuous operation should be able
to process resin and liquid Teed streams of larger volumetric throughput. The
elimination of the intermittent-flow control is a Fur~her economic advantage,
especially when the contactor is applied to large scale industrial operations.

'
3.1

3. THEORY

Fluid Diode

.
'Ringleb (13) investigated the formation of standing vortices by a two-

dimensianalin~ompressible

inviscid fluid in single

c~sp-<:avities.

He suggested a

general form of mapping function to generate the cusp-<:avity shape in which a

single standing vortex is automatically created when a parallel flow passes over it.
Wall profiles and flow streamlines for families of cusp-<:avity shapeS: were obtained
by integrating equations of the form

dF(a}
-=

a(a-a } ....

(3.1.1)

do

where F(a} is the mappi~g f~nction and a the complex mapping variable; along each
streamline its imaginary part is constant. The complex parameters a ,a , a3' etc.
l 2
determine the nature of the flow field.
Cusp-<:avity 6f the simplest form requires only one complex parameter in the
mapping function (3.1.I). Ringleb (13) and Th?",pson (14) have suggested expressions
of such mapping functions:

(3.1.2)

(3.1.3)

10

II

where 01 = m l + inl' The two adjustable parameter~Qre ml and n which determine

l
the shape of the cusp-cavity. When ml is held constant to fix the scale, a family
of cusp shapes ca,n be generated for different non-positive values of nl by plotting
y versus x in the following expression for the shape of a streamline,

x + iy

F(a)

(3.1.4)

as a is moved along the real axis. The cusp point corresponds to a = O.

Ringleb's analysis is only helpful in designing the channel of Thompson's
contactor (14) for flow in the low pressure-drop direction and serves mainly to
establ ish the kind of cusp-cavity shape that will permit a stable vortex to be
establ ished. When fluid flow is reversed, the inviscid model would predict
unchanged streamlines.
The mathematicat shape gener~ted from the mapping function was modified
,"

by Thompson in order to obtain turbulence during reverse flow. The original arid

Thompson has
/

adapted shapes are shown in Figures A.I and A.2 in Appendix A.

found that if the radius of curvature of the wall C is less than 4 inches, the flaw
may separate and short-circuit to the cusp endpoint, causing less turbulence in the
flow pattern.
Tho fluid flow patterns in Thompson's fluid:-diode contactor are shown in
Figures A ,3 and A.4. The contactor is constructed of two identical horizontal
fluid-diode sections but oppositely oriented. In the low pressure-drop direction
(A to B), the mOin stream follows a direct path through the channel, and it separates
from the wall at each cusp point D and a standing vortex is fanned in each cavity.

12

,:
i
I
I

While some turbylence may be g~erated between these two-flows, the general
appearance to the flow pattern is very smooth. W~en the direction is reversed and
the stream flows in the high pressure-drop direction (B to A), the stream follows a
highly tOrtuous Path. The high vetocity portion of the stream becomes attached to
wall C and is diverted into the cavity, where it is forced to partially reverse in
direction and deflec~o the opposite side of the channel.
Egan (7) tested fluid diodes of three different cusp-c:ovity shapes which are shown in Figure A.S. Results for water flow showed that the friction factor of the
fluid diode in the' high pressure-drop direction ton be 20 times more than that in
the low pressure-drop direction as the Reynolds number of the water fjow in the
~

chonne'l is i~creased. The pressure drop for water flow in either direction of the
. diode is higher than that for pipe 'flow but the characteristics of flow in the high
f"essure-drap direction is entirely different from that of pipe flow. ' Figure D.I
. (
::
shows that the pressure-drap ratio, A Phig!lA Plow' inc:reo~s with the Reynolds
number. The ratio is independent of the fluid-<liode lateral displacement in the
flow region' investigated by E9an. Note that only the ratio is constant but the
pressure drap in both directions decreases in the some proportion as the diode loterol
displacement is increased.

. The fluildiodes with cusp-covity shapes A and B of Figure A.S behove

almost identically. However, fluidic: element C also has a pressure-drop
chorac.teristic comparable to shapes I ond.2. This observation of Egan indicates that.
the i'ncreased number of cusp-c:avities per unit length. has counteracted the short-

" of curvohJre at paint C of shope C is less than

circuiting expected since the radiUs

13

four inche"-. .;
When the fluid diode is odapted as a downcomer structure in the multistage
fluidized-bed contin.uous countercurrent ion-exchange contactor developed by
Souhrada, the characteristics of fluid flow .through the fluid-diode downcomer in
the high pressure-drop directian created by the bypassing process solution during
startup still hold, and serve the purpose of discouroging liquid bypassing up the
downcamers. At steady state a continuaus stream of salid resin particles flows
dawn the dawncomer in the low pressure-drop direction. The flow pattern of the
solid particles with a density slightly higher than that of water is expected to be
different from that of an incompressible inviscid fluid. The characteristic of the
I

resin flow through the downcomer will be discussed in Section 8.1.1.

3.2

lon-Exchange Resin
lon-exchange resins are insoluble spherical solid particles, whose framework

is an clastic matrix consisting of an'irregulor macromolecular, three-dimensional

network of hydrocarbon chains, 'carrying fixed ionic groups associated with
exchangeable cations or anions. These exchangeable ions can be exchanged for a
stoichiometrically equivalent amount of other ions of the same charge when the
resin is in contact with an electrolytic solution. Carriers of exchangeable cations
are called cation exchangers and carriers of exchangeable anions, anion exchangers.
/\

There are Four main types of ion-exchange resins in commercial use, namely
strongly and weakly ocidic (cation) resins and strongly and weakly basic (anion)
rcsi ns.

14

The Dowex-50Wx8 resin used in the present study is

strongly acidic cation

exchanger. It is prepared by the'nuclear sulphonation of styrene-divinyl benzene

beads ond the fixed ionic group is

-SO;.

The resin is stable chemically and

physically. The di~inyl benzene is a cross-linking reagent, the. percentage of

which determines the degree of crasslinkage in the resin matrix. The degree of
crosslinkage determines the physical and chemical properties of the resin. A
decrease in crosslinkage increases the matrix poros;ty and hence its ability to swell
and tokes up more water. The increase in porosity also increases the mobility af the
counter-ions in the resin and hence the rate of exchange.
A detailed discussion on ion-exchange resins can be found in Helfferich (32).
~

3.3

Ion Exchange
lon-exchange is a mass-transfer process between two phases. Ions are
./

transferred from a solution (usually an aqueous phase) to an ion-exchange moterial

(usually a solid and cammonly a polymerized resin), while an equivalent number of
ions of the same charge, but of different species, are displaced and transferred in

the apposi te direction. An example af ion-exchange is the Na +/H+ exchange on

Dowex-50Wx8 resin. The equation of exchange is given by

(3.3.1)

l
~

This is an interdiffusion between counter-ions, H+ and No-+:. The

does not participate in the ion-exchange process; however, there is still som sort
of distribution of the chloride ian between the resin and aqueous phases as an be
I

15

found in the work by Whitcombe et 01 (33).

For on ordinary ion-exchonge system, the ion-exchange rate controlled by
chemicol reaction has been ruled out (32). Ion exchange, as a rule, is purely a
diffusional phenomenon. Hence the rate of exc.hange is governed by the some
general considero tion as for any mass-tronsfer process. There are mass-tr"lnsfer
resistances in both phases, and the overall rate may be controlled by either (or by
both) of these resistances. For either phase, the mass transfer may be' described by
a rate equa tion of the form

dq
= k'A (c...:.)

dt

(3.3.2)

If the two exchanging ions have markedly different diffusion coeFficients,

then on electric Field will be set up in each phase and these fie,lds will affect the
diffusion process. One of the consequences is that the rate aF exchange can be
markedly composition dependent. This phenom'enon in the resin phase can be
illustrated by the present ion-exchange system, Equation 3.3.1. At the start of on
exchange, the hydrogen ions, which have a higher mobility, tend to move ofF more
rapidly than the sodium diffusing in the opposite direction. This leads to a charge
separation which proceeds only to a minute extent beFore an electric field is set up
in such direction and of such strength that the hydrogen ions are slowed down, and
the sodium ions are speeded up until their opposing fluxes are equal. The efFects of
these electric fields are allowed for in the Nemst Planck model For diFfusion in a
resin particle and the interdifFusion coefficient of two monovalent ions, as a Function

16

.7'

of resin composition is given by the following expression

0=------

(3.3.3)

Note that the model assumes the mobility of an ion to be independent of its
surroundings,' which is a reasonable assumption for the exchange of ions of equal
charges (32).
Turner et 01 (34) has verified the Nernst Planck model for diFFusion in
Zeo Karb-225 resin for the exchange of H+ and No + ions. The self-diffusion
coefficient of the hydrogen ions can be more than four times that of the sodium ions
depending on their concentrations.
Turner and Snowdon (35) have also applied the Nernst-Planck behaviour in
the study of the I iquid-film mass-transfer coeFficient in the exchange of H+ and No +
ions between aqueous chl,?ride solutions and Zeo Karb-225 ion-exchange resin. It
has been shawn that where the liquid film plays a significant role in ion-exchange
kinetics, the variation of exchange rate with liquid composition should not be
ignored. The liquid side mass-transFer coeFFicient can vary by a factor of more than
three as the liquid composition changes.
For the present ion~xchange system, H+-Dowex-50W-x8-Na +, Duncan and
Lister (36) have shown that the equilibrium constant, K, remains practically

constant for a given tolal solution normality, Co, and is given by

17
(

y(l-x)
K

=---

(3.3.4)

x(l-y)

where x and yare equilibrium mole-fraction compositions on sodium bases in the

aqueous and resin phases respectively. The temperature dependence of the
equilibrium constant is usually negligible.
Gilliland and Baddour (37) have determined the equilibrium constant for
sodium chloride solutions of various normality equilibrated with Dowex-50-x8 ionexchange resin. The plot of log (Co) vs. K in Figure D.2 gives the following
analytical expression

-log (Co) + 17.75

(3.3.5)
12.5

Note that the present experimental study uses the Dowex-50W-x8 resin which
is claimed (38) to have stronger physical properties than the Dowex-50-x8 resin.
However, the correlation (3.3.5) is assumed to be the some for Dowex-50W-x8
resin since the chemical properties of the two resins aro identical (38).

3.4

Fluidized-Bed Ion Exchange

M~ss transfer in fluidized-bed ion-exchange is a volume process'. With the

Bee.k (42) h~s made a distinctioh betwee~ surface and volume processes in
fluidized-bed operations. Cooting, wetting and agglomeration are surface processes,
and drying, cooling and freezing are volume processes.

18

small dependence on temperature, the physical factors which determine the rate of
the process include inside and outside particle moss-transfer, the fluid and solids
mixing, and the probobility of particles encounter.
The mass-/ransfer rate in a shallow fluidized-bed of ion-exchange resin is
commonly controlled by both the sol id- and 'I illuid-sides mass-transFer resistances.
Turner (39) has shown that the solid-side resistance can account for 20% or more of
the total resistance depending on the compositions. When the liquid flow was

increased, the liquid boundary layer around the' resin particle decreased, the liquidside moss-transfer resistance was reduced and consequently the overall mass-transfer
coefficient was increased. The ion-exchange resin is at a state of particulate
fluidization when the incipient fluidization velocity is exceeded by the solution
flow. For a shallow fluidized-bed, th~ resins are perfectly mixed and the liquid
passes through the bed in plug flow.

However, longitudinal dispersion of liquid

cannot be neglected in deep fluidized beds.

The moss transfer study of Petrovic and Thodos (40) on a gas-solid fluidizedbed system contributed to understanding the moss-transfer process in a liquid fluidizedbed system since they worked with an almost homogeneous bed with the fluidizedbed voidage given by

~ = (.) 2.5
Vob

(3.4.1)

Eb

which is similar to that by Richardson and Zaki (41) for liquid fluidized-beds. The
following moss-transfer relations' were reported for the fluidized-bed system,

19

Vo

0.57

-Vob)
and

(_E )

(3.4.2)

0.44
{3.4.3}

Eb

Equation 3.4.2 ~hows that kL is proportionol to

V0 is proportional to

0.57

E - and IVO

In Equation 3.4.1 .

E 2.5. Hence kL is proportional to E O43 . Within the

. range of voidage (0.5 to 0.8) commonly used, the values of

E 0.43 varied by a

maximum of 18%. A consequence of this result is that the kL value is almost

independent of the superficial fluid velocity.
The above conclusion coincides with the findings of Rowe ond Claxton (46)
that the tendency of increasing Re to increase Sh with fluidized beds is largely
nullified by the resulting increase in vaidages. For fluidized-bed ian-exchonge,
the moss transfer between resin porticles and the aqueous solution flow is complica. ted by the composition dependence of both the solid- and liquid-sides moss-Ironsfer
coefficients . However, the experimental results of Turner and Snowdon (35) for
liquid fluidized-bed ion-exchange shows thot the change in liquid-film masstransfer caefficien!, k L, with bulk solution c"'l)positian is identical at Reynolds
number 7.2 and 8.8 for a total solution normality of 0.01 N.
Snowdon and Turner (39) claimed thot less mass-Ironsfer occurred in fluidized
beds thon in fixed beds when compored at the 7me

R~ynolds number,

Uofd/P.

{based on I iquid superficial velocity}. The~Orrelated the moss-transfer results of

\

20

both fluidized and fixed beds by

1/3

0.81

=""'E

Sh

where Sh

(3.4.4)

ReSc

= kLd/D,

Re

= Uo~

d/p.. , Sc

=/l/f

D.

However, comparisons of the mass-transfer rate in fixed and fluidized beds on

different bases can give different conclusions in the relative effectiveness of the
two processes. Beek (42) shows that a correlatian between SI.Sc 2/ 3 and U d/V
o
can bring the mass-transfer results for packed and fluidized beds close together,

St.Sc

where Re

2/3

+
-0.5
(0.81 - 0.05) (Uod/'J )
for 5

= lJod/(I- ~ ),

Sc

= "" /D,

Sh

<

Uod/V

=E kLd/(I- ~ ) D

and St

<

50

(3.4 .5)

= E kL!Uo '

The correlation is obtained from the experimental data of Snowdon and Turner (39)
for liquid-fl\lidized ion-exchange beds (100

<

Sc

<

1000, 0.43

< E<

0.63). A

comporison between both beds gives for equal va lue of U d/" equal values of
o

(kL/U )
o

E ond consequently lower kL values in the fluidized bed. Beek believes

thot this expression is the most accurote and reliable presentation of today's knowledge for the most important range of Reynolds number, for all Sc values met in

.,
I

practice and for porosities up to 0.75.

The study by Hanna (51) on the residence" time distribution of sol id particles
,

'

in a countercurrent fluidized-bed system shows that the mixing of solid particles can

,~

be increased by increasing the liquid flow-rate (superficial liquid velocity), by

decreasing sol id particle throughput and by decreasing the bed height.

II
I

21

3.5

Continuou, Countercurrent Multistage Ion Exchange

The continuous countercurrent mul tistage fluidized-bed ion-exchange

contactor under present study operates on continuous solid resin and liquid flows
with equal distribution of resin in each stage when steady state is reached. The
theory of continuous countercurrent ion-exchange is analogous to that of solvent
extraction and the diffusional mass-transfer column operations. The performance is

governed by the equilibrium relationship, the moss balance, the rates of mass
transfer, and the c~iciency of the contactor.

The performance of the ion-exchange column can be analysed by the McCabeThiele diagram. This method has been suggested by Moyle (43) and used by Turner
and Church (44). A material balance across a section of the column and the battom
of the column (Figure D .3) gives the following component balance

(3.5.1)

The ion-exchange operation is a truly equi-molar counter-diffusion process. The

volume flow-rates of the solid and liquid flows in the present system are constant.
Hence, we have

L =L =L = - - - - - - -=L =L
o
I
2
N

(3.5.2)
and

=c =c :::------;::;c =c
I
2
.
N

q =q

=q

.3

=------=q

N+I

=q

22

Equation 3.5.1 can be reduced to

(3.5.3)

or

Lc
=- x
n
Rq

RqYI - Lcxa

+-----

(3.5.4)

Rq

Equation 3.5.4 relates the compositions of the resin phase, Yn+I' and the liquid
phase, x n ' flawing past each other between stages. The locus of (x , y I) on the
n
n+
equilibrium diagram plot is a straight operating line as shown in Figure D.4. The
linearity of equation 3.5.4 is displayed much more vigorously for ion exchange than
the corresponding distillation or absorption, in which the usual simplifying assumptions (45) are usually not strictly true.
The number of theoretical stages (NTS) can be determined from this diagram
by a stepping off procedure involving alternate use of the operating line and the
equilibrium curve as shown in Figure D.4. Tho height equivalent to a theoretical
stage (HETS) is then obtained by dividing the actual contactar height by th- number
of theoretical stages.
However, the actual height of the present ion-exchange column useful for
mass transfer is that of the weir height in each stage. The calming zones between
stages and the stage separation necessary for the installation of the fluid-diode
downcomer is determined by the physical operational requirement and should not be
considered in the mass-transfer performance study." Hence, it is only proper to
calculate the HETS using the total active height of the ion-exchange column; and
HETS is thus given by

..

23

HETS =

Z/NTS = WH. N/NTS

(3.5.6)

In the actual countercurrent multistage contact, the solid and liquid phases
leaving a stage are not in thermodynamic equilibrium. An overall stage efficiency
may be defined to describe the degree of approoch to equilibrium in the contacting
operetion. The overall stage efficiency is defined as the ratio of the number of
equilibrium stages required for a given separation to the number of actual stages
requi red and can be expressed in percentage as

E = (NTS/N) '. 100%

os

(3.5.7)

This stage efficiency is simple ta use in calculations, although it does not allow for
the variations in efficiency which may occur from stage to stage. The overall stage
efficiency depends on many factors such aS,the time of contact and degree of

,l

dispersion of the phoses, the geometry of the stages, the rate of moss transfer, and
the physical ~s and flow rates of the solid and liquid phases.

3.6

Factorial Experimental Design

Experimental design is the planning of data collection from on experiment or

a series of experiments so thot the desired information can be obtained with

sufficient precision. The response of the dependent variable to the independent
I

variables (factors) and to the levels of the factors

~re of

interest to the experimenter.

For a given number of factors and levels of the factors, a complete factorial design
can evaluate jointly and efficiently the effects of all the possible combinations of

factors and levels on the response. Sk.ndard works (47,48) deal extensively with

24

the theory of experimental design. The following is adopted from Peng (47).
For a design of three factors (A, B, C) at two levels, the eight (2 3) possible
combinotions (treatments) are

I, a, b, ab, c, ac, be, abe.

The "I" represents the treatment with all the factors at !heir low levels. In each of
the remaining trealment combinations, the high level of any foetor is denoted by the
correspondi ng letter and the low level by the absence of the corresponding lotter.
The effect of changing A when the other two factors are at their low levels is
a - I

(3.6.1)

Similarly, the effect of changing A when B is at the high level and C at the low
level is
ab - b

(3.6.2)

The effect of changing A when B is low' and C is high is

oc - c

(3.6.3)

Finally, the effect of changing A when Bond C are both at the high level is
abe - be

(3.6.4)

The main ?ffectof A for the whole experiment is defined to be the mean of these
four effects, namely

.\ (0 + ab + ac + abc - I - b - c - be)

(3.6.5)

When C is at the low level, the interaction AB i'-defined to be one-half the

..,
~

difference between the effects of A a,t the twa levels of B, (3.6.1) and (3.6.2), i.e.

25

;; ab - b) ..: (0 - I))

(3.6.6)

Similarly for C-llt the high level, the interoction of AB is

1 ({abc

- bc) - (ac - c))

(3.6.7) ,

Hence the AB interoction for the' whole experiment (over both levels of C) is the
averoge of (3.6.6) and (3.6.7)
AB = ~ (ab + I + ab.:: + c- b - a - bc - ac)

(3.6.8)

The ABC interaction is defined to be one-half the difference between the AB

interactions for the two different levels of C,
ABC = ~ (abc - bc - ac + c - ab + b + a - I)

(3.6.9)

Thus if AB is the SOme for both levels of C, ABC is zero.

For a two-Ie~el factorial design at higher levels, the equations for forming
the main effec Is and interactions can be generated systematically by one of two

.\

methods: effect matrix (47) and Yates' method (48). The methods for generating
the equations are illustrated in Tables E.I and E.2 in Appendix E.
In a complete factorial experimental design the required number 'of measurements is often beyond the resources of the investigator, or it gives more precision
in the estimates of the effects than necessa4, or estimates of

high-~der interaction

effects are 'of less interest. For example, in 02 8 complete,Jactorial design,

treatment combinations are required and each main factor is an average ave

combinations of the other factors. In many practical sitlXltions, a certain degree

of continuity and regularity can be expected with regard to the manner in which the

j.
~

response vories ond the, higher-order interaction effects involving thr/!e or more

[
l

f
factors are often negligible. Then design requiring only odefinite frection of a

26

complete factarlpl
may be used ta estimate the main effects and two-factor
,
,

interaction effects in such a way thot only the information on the highei-order
J

"

interactions is lost.

'-,

, A fraCtioool factorial experimental design can redu~e the required treatments

of a 2'K comple'te factorial design by a factor of 2-r, where r = I, 2, 3 .. and
r is I ess than k. The information on orie or more interaction effects is sacrificed
in order to spl it up the whole experimental design to 2 r bl ocks ~f treatment
combiootio~.

These interaction effects are called the defining contrasts. The

bl,ock which contains the 'treatment combination "I" i~ called the principal block and
is no~allY used for experimental purpose. The ather ~s can be obtained from
the principal block by multiplying (modulo 2) each of its effects with the defining
contrasts selected. Exominatian of'\he effect matrix s~nerated for all the treatme~t

'

combiootions (2k) shaws that the estimate of a factar or factor-interaction in a block

is

id~ntical

to that of

anothe~ facior

or factor-interaction from each of the other

blocks. The'se factors o~ factor-inr'eroctions which have identical estimates are

\;)

called aliases. The effect matrix will. also show thot the defining contrasts have
, estimates identical tot~e general, mean, ond hence Fheir information'is lost<. I,he
idea of confounding (confusing) in the fractioool
"

'

illustrated by the effect matrix generated 'for ci

~actorial

experimental design is
?

1/2-replic~te' of a

3
2 foctorial

design in Table E.3 of Appendix E.

The determination of the total effects from the experimental resul ts for the
fractional experimental design can be achieved by the effect matrix above or 'by a
modified Yates' technique as demonstrated in tne present li4 x 27

f~ctional

il

27

factorial design.
Tho analysis of varianco (ANOVA) is a systemmotic method for making a
useful invostigatlon evon when there are rosldual variations. An analysis of
varianco can be performed on tho calculated total cifecn obtained by the Yatos'
analysis. The high-order interactions which are assumed to be zero are pooled
N?lgether to form the residual variation (random error). The statistical F-ratios can
'[.

then be used to distinguish the significance of the effecn at various conft'dence

levels. The detailed pro~edure 9f th,e analysis of variance can be found in Table E.B

for the present study of a 1/4-replicate of a 27 factorial experimental design.

r.

The magnitude of the effects, in on experimental design can be oxamined by

another statistical method. Doniel (49) proposes the use of tho h~lf-normal plot and
standardizod half-normal plot. This approach removes the quostion as to which
effects can reasonably be assumed to be zero and pooled togother inta the residual,
as required in the ANOVA. The method is discussed in Appondix F.

I,

/:,, .

4. EQUIPMENT DESIGN AND APPARATUS

4.1

lon-Exchange, Column System

4.1.1

General
The detail flow sheet of the ion-exchange column and ancillary equipment is

shown in Figure A.6.

Every part of the system in contact with the sodium chloride

process solution, distilled water or ian-exchange rosin is made of stainless steel,

PVC or Plexiglas materials. Process solution from the staroge' tanks (T , T ) is
I 2
pumped into the bottom stage of the multistage ian-exchonge column (CI)' The
liquid coming out from the top of the column flaws through a cyclone for resin
remov~1 before going to tho drain,

Fresh resin is fod to the top stage of the column

and the resin flows through tho stages dawn the column by gravity. Before_entering
,/-

into the exhausted-resin collector (BI) resin from the last stage of the column can be
sampled by the sampler (S).

4.1.2 lon-exchange column,

The 4 inches (10.2 cm) I.D. ion-exchange column is 3 feet and 8.5 inches
(1.13 m) high. The material of construction is almost entirely Plexiglas. The
column consists of 9 sections as shown in Figure A.7 and are joined together at
the flanges by 1/4-inch (6.4 mm) nuts and bolts. Sections I to 6 are the stages for
fluidized-bed ian-exchange. Section 7 is a calming zane for the liquid feed.
O{'

.,

Section 8 is the liquid feed stage which has three perforated Plexiglas plates to

28

29
I
improve liquid distribution. The detailed structure of sections 7 and 8, and of the
liquid distribution plates, are shown in Figures A.8 and A.9. Section 9 is the last
section which connects the exit resin-stream"'"'to the exhausted-resin collector (BI).
The resin sampler is built into this section and the detailed design is shown in
Figures A .10 and A .11.
The stages (sections I to 6) of the column for moss transfer are identical in
structure except for section I whi"h has a cylindrical trough to direct the liquid
effluent into the exit hose as shown in Figure A.6. The detailed structure of section
I is shown in Figure A.l2. Each of these stages.l\as identical troy structures:
fluidized-bed support, weir shape and fluid-diode downcomer. In the present
study, the performance of three different weir shapes is , studied, nomely straight,
conical ond cylindrical shape weirs. The performance of ' the straight .,:.~ is also
studied ot three' different weir lengths. The straight weir type also has four
different heights. The designs of these weirs and their corresponding tray-support
structures are shown i'Origures A.13, A.14, A.15, A.16 and A.17.
The three flu~d-diode designs used for the downcomers in the preliminary
study ore shown in Figure A.5. These fluid-diodes were prepared by Egan (1971).
The deta iled design of the fluid-diode downcomers used in the present mass-transfer
study is shown in Figure A.18 and these downcomers are identical in each'sectian
for mass transfer except that of section 6, which is shown in Figure A.19. The last
fluid-diode downcomer directs the resin flow into a 3/8-inch (9.5 mm) tygon
tubing which leads into the sampling section of the column.

nl~:.:..-flom.fl!r ~('ctions O~ !.h0Wrl in FiUurl'

A.I.

~fudit::' undL'r voriou::. fJcornt .. trical combinations

Th" cnlumn io mode lip of

,~ction'

inkractions observed in sL'ctiorh

I,

011

the .. loge ~tructlJrt:s.

I, 2, '6, 7, 8 und 9 only.

TI". qUCllilnliv('

2 and 6 are -rt!presl!nlative of

th.)

L'ntirl! collJmn

o;wration. TI", water cHlu"nt fwm the top of the' column i. r,cycled to tl", f""d

4 .1.3 Liquid feed system

The li'1uid feed 'yst.'m i, illclucLd in Figure A.6. The Iwo 45-rlaIJon (liO lilr,,)
poly,thylen(' lanks (TI' T ), connected in ,,('ries br
2

(l

11/2-inoh (:1.8 em) I,,,, ..,, ,"'"

,odiuln chlorid" feed->ol"tion >Iorage tank>. The solutioll is purnp.,d throuGh"

"otumekr (fIll ond into the ion-exchange column at section 8. TI", liquiri f"ed-mtl!
con lH< controlled by the stoinlC'S' stecl ncedle valves on th"
('.'3, "4)' The mixing lin" is used fN prep~ring

(1

rc~yclc

"nd feed I in"s

ho:noJL'neous 501"tion b,for,' on

expLrimental mms-transfer run is storfd. The temperoture of th" f"ed solution elln
be mointain .. d above rOO;n t"mp~raturo by passing hot water throunh thJ stuinl"ss
,.te,,1 coil in the storage tonk. The temperature is monitored by the tl",rmostat in
the tnnk and the water flow by the on-off solenoid valve

015 ) on

the hot water I inl'.

Ti", purpose of heating is to.displaco dissolved air from the solution so os to prcvpnt

31
''--..../

Cnllidmilldtil'\rl nf tJ11~ f"t,d scdutivil i~ ovoidt'd

hy tllt~

ll:.{'

llf II

PvC

pllillp

I'

k',j,l, ty~j\1I1 !,ipin~l, st.li'''t~s\ ~fl'd Cllld PVC \"lh~'~:~ ~tl1illll'"'' "Il,,~l Ulld Sllll.'.'

flltl1-

\I
I

I.::i,,-[,.d 0.>111.> (BZ) alld Ull "'lUll I vofum. of re,in und wat"r is displaced out of tl",

h.,III" Ihr.,,!:)h th. I "',in f""d lim' and into th" first section of th" ion-exchonn"
.
/'

clllumn.

Durin!) c.tart-lIp, air Lubblus trapped in th" bottle c(ln ba ",loosed throunh
"

r,d",,,,,t .. r (1(;, Pl""1' (1'2) is

(I

Mast",flox tubina pump. Tho piping c"mists of

V:-illch (el .,1 nlln) tra,)n und slc,;nl"" st"cI tubing' to ovoid ionic contamination of th"
di.lill,d wukr (lild hydr0D"n-form ,,sin. All other rnateri"ls of construction in

The 'osin f".,d system produces negligible resin ottrilion becaustl no mechanical
1':111'. (II" inv01ved in the.ph).sical tronsport of the lesin.

0ri<Jilnily dt'sign"d as the solid f""d system.

Howev""

A solid screw fcedt'r waS

grinding of the resin

b Iw,.n th" scr.w threads and the :ocrew housing was considNabl".

Increasing the

CI,"I.lnc', b.tw"..,n the screw and its housin] rl'duccd the grinding but thosensitivity

32

of the resinJlow-rote to the screw rotation was greatly reduced.

4.1.5 Resin sampling system

The detailed design of the sampler is shown in Figure A.lO and A.II. The
sampler's teflon plung.,.. has a cavity for the collection of resin coming down inta
section 9 of the ion-exchange column. The sample cup is screwed onto the sample
collector and can be remO/ed. The oscillatory pump (P ) pumps distilled water into
3
the sample collector and displaces the air from the sample collector and sample cup
through the outlet line. A bypass line is provided to adjust the water flow into the
sample collector. The piping is made up of 1/4-inch (6.4 mm) tygon tubing.

4.1.6 Exhausted resin collecting system

The bottom end of the ion-exchange column is connected to a 3-gallon
(11.4 Iitres) Pyrex glass bottle (B I) by a I-inch (2.54 em) tygon tube. The downflow
of resin and solution can be cut off by closing the PVC globe valve 019}' From the

neck of the bottom through the rubber stopper is a 1/4-inch (6.4 mm) bleeding line.
The water Flow-rate through the bleeding line is adjusted by a stainless steel needle
valve

01 10)'

. The purpose of the bleeding is to m<lintain a smooth Flow of resin

down the 3/8-~nch (9.5 mm) tygon tube in section 8 and also out of section 9 of the
ion-exchange column. When the bleeding of water is at the some volumetric flowrate as the resin feed-rote, resin will not back up or block the r~sin exit at the
bottom of th.e column.

33

4.2

Resin Drying System

The resin drying system is shown in Figure A.20. The Plexiglas column

((2), 39 inches (99.1 cm) high, hos on inner diameter of 2 7/16 inches (6.2 cm).
Laboratory air supply is used for drying the resin before size screening. The air
posses through a filter before entering the column at the bottom. A 3/8-inch
(9.5 mm) tygon tubing connects the filter and the air inlet of the column. The top

,opening of the column is covered by a 60-mesh screen to prevent the carry-over of

resin during the drying process.

4.3

Resin Regenerating and Washing System

The system is shown in f'"igure A .21. The column ((3) is 4 feet and 2 inches

(1.27 m) high with a 4-inch (10.2 cm) 1.0. ond the material of cOnstruction is
Plexiglas. The bottom of the column is connected to a Masterflex tubing pump (P4)
which pumps either acid or distilled water into the column for regeneration or
washing of the resin .. Near the lower end of the column is on outlet valve (V )
20
(a stop-cock with a 5-rnm bar'! teflon-plug) for resin removal. Bottle (B4 ) is the
J
.
storage for resin to be transferred into the column. Transfer of resin is by volumetric
displacement as for the resin feed system. Distilled water is pumped into the bottle
and an equal volume of resin and water is forced out of the bottle and into the column
at the top. Air trapped in the bottle during start-up can be released through the
vent by opening valve V

26

B5 and B6 are 5-gallon (18.9 litre) Pyrex gloss bottles. 'Bottle B5 is for the
storage of recycled hydrochloric acid during the lost quarter of the regeneration

34

cycle. Bottle B6 is the storage for fresh 10% hydrochloric acid. Outlets of the two
bottles can be closed by the S-mm bore teflon-plug stop-cocks (V23' V

22

). The

two outlets arc joined together before going through the Masterflex tubing pump
(p.v and finally into the column at the bottom inlet. The suction of the pump is
also connected to the 4S-gallon (170 litre)'distilled water storage-tonk (T3). During
resin regeneration or washing, liquid effluent overflows into the trough around the
L--

top end of the column and flows either to drain or into recycle acid bottle (BS ).

~.

5. FRACTIONAL FACTORIAL EXPERIMENTAL DESIGN

Seven operational and geometrical parameters for the continuous countercurrent

ion-exchange column were chosen as the factors for the factorial experimental
design. The response is the overall stage efficiency. The factors chosen are
(i)

resin f1o,:,"-rate,

(ii)

solution flow-rote,

(iii)

resin partie Ie size,

(iv)

weir lengt h,

(v)

weir height,

l'

(vi)
(vii)

fluid-diode lateral displacement, and

downccmer clearance.

These factors showed significant effects (at least visually) on the operational
bohaviour of the column, including the efficiency of solid-liquid contact. Although
the solid and liquid f1ow-rotes were shown to be significant by Egan (7) and Cho (8),
these two factors were included in the experimental design to reveal any possible
interactions among themselves and the other five moin factors. The initial sodium
chloride concentration was not

i~uded,

since the emphasis desired was on

gco.-netrical and physical factors.

A level of two for the experimental design was chosen. Hence a complete
factorial design of seven factors at two levels required a total of 128 experiments.
Experience showed that only two experimentol runs at most could be performed in

35

36

one day. Thus a complete factorial design was considered nat practical for this
investigation and a quarter-replicate selected. This required only 32 experimental
runs .

A quarter-replicate of a complete factorial design causes every effect to be

confounded with three ot.her effects. The three defining contrasts which divide the
complete design into four blocks were carefully chosen such that the main effects
and 2-factor interactions were not confounded with each other. All possible
combinotions of defining contrasts were tried and those which produced the leost
undesirable confounding were selected. Table E.4 is a list of 011 the possible
defining contrasts. The system of defining contrasts that seemed to ,,-,it best for the
present study was ABC DE, ABCFG and DEFG, The effects of the complete factorial
design are split up into four blocks by these three defining contrasts. The principal
block is generated by the modulo equations

(5.0.1)
(5.0.2)

The principal block generated is shown in Table E.50. The other three blocks
and their corresponding pair of modulo equations'tare shown in Table E.5b. The four
olios groups are shown in Table E.6. The four effects in each row are confounded
with each other due to the blocking. The three defining contrasts (ABCDE, ABCFG
and DEFG) are confounded with the general mean,!, hence their information is I,ast.
The loss of information about
interactions with more thon two factors
is of no concern
or
. '

37

because interactions involving more than two factors are often negligible, and
hence can be assumed to be zero. All main factors, and fifteen of the twenty-one
2-factor interactions, are confounded with interactions of second order or higher.
Three pairs of 2-factor interactions are confounded with each other as indicated in
Table E.6. The pattern of confoundinQ for the 2-factor interactions is shown in
Figure E.I. In order to overcome undesirable confounding of the 2-fador interactions, the factors (solid flow, liquid flow, etc.) were assigned to the letter
(A, B, C, etc.) in such a way that one of the.2-factor interactions in each of these
three pairs was very likely to be zero. Thus, the assignment of factors to the
lettersare as follows
A

resin flow-rate

solution f1ow...-ate

resin particle size

weir height

f1uid-diode lateral displacement

downcomer clearance

weir length

Referring to the 2-factor confounding pattern, it can be seen that DG (interaction

between weir height and weir length) which might be significant is confounded with
EF (interaction between f1uid-diode lateral displacement and downcomer clearance)
which is very likely to be negligible in terms of its effect on the overall stage
efficiency of the ion-exchange column.
{

38

In normal practice, treatment combinations in the principel block are chasen

for experimental wcrk.

Experiments of different treatment combinations of an

experimental design should be carried out randomly. However, in order to save

time and labour, treatment combinations of the principe I block were performed in
the order shown in Table 'E. 7. Each day two experimental runs were performed.
These two runs involved the some internal stage structures in the multistage column,
hence the column was dismantled after every two runs rather thon after every run,
The two experimental runs in each day used a resin perticle size different frqm the
day before because the exhausted resin required one day's time for regeneration and
washing before it could be used again. It is claimed that such practical planning is
not against the principle of randomization in statistical experimental designs.

"

6. EXPERIMENTAL PROCEDURE

6.1

Analytical Techniques

6.1.1

Displacement of hydrogen ions from resin by sodium chloride solution

3 ml (bulk volume) samples of hydrogen-form resin (-30 +35 mesh) were

equilibrated with 100 ml of sodium chloride solution of concentrotions = 0.25,0.5,

1.0, 2.0 and 4.0 N. Interparticle water content in ihe bulk volume of meosured
resin wos removed by suction in

Buchner funnel bef~e the r~sin wos odded to the

sodium chloride solution. The calomel and glass electrodes

connec't~

to the Orion

digital meter were used to follow the change in pH value of the resin and sodium
chloride solution mixture which was agitated by

~a~netic stirrer. The cha~ge of

pH value, i.e. hydrogen ion Iiberated, as a function of time is shown in Figure B.1.
It can be seen that equilibrium was obtained in 2.5 to 4.0 minutes. The more
,concentrated' sodium 'chloride solution required less time for equilibriumro be
reached. Hence five minutes of mixing were enough before the mixture could be
titrated for the determination of resin,capacity.
Figure B.2 shows that the hydrogen ion liberated increases with the initial
sodium ion concentro tion . .

-6.1-.2

Concentration of sodium chloride solution for resin semple titration

V"

A measured bulk semple volume of -30 +35 mesh hydrogen-form resin was
mixed with a given volume af sodium chloride solution of various concentrations,

39

40

"nd III .. cllunoa In p~lvolut1 of Iho Il'lullon mlxlu!" durlnU Illn Illrollon wllh 0.1 N
~"dllJ'" lIydr,1 xldl'l lolullon

W\1I

fo~low .. d ullnll Ihll 1:(llomal (lndgl(1II ol':'clrodol

c.""",cl .. d loliltl Oll'n dlOltol pl-tlmV molor. Tho plottlnu ~f Ih" I'll volun

(II

funcll"" of' III" vo'lum" of Iltranl oddod ylClldl Iho Iltrollon ..IHV". Th" ,,'nd polnl
~

of 1II,olion Is <II Ih" polnl of Inflaxlon. Flour" 0.3 Ihowl Iho Illrollon curv", lor
5 'ml(Ix.lk vol urn,,) of IC1lI~ mlx"d with

160 ml

cl>r\l:"nlrollon, 0.25, 0.10 and 3.0N. Figura

of lodlum chlorld:" lolullo", of

D." showl Ih" Iltratlon curV"1 for

3 ",I 01 r,,~ln In 100 OIl of 4.0 and S.ON 100llum chlorldololullon. ThOle:> nxparl"'"nlal ,;',ul," Indk<lh'l Illal

Ih~ Iltrallon Clnd polnl II Indnpondant oflha concanlrollon

of ,odium chlorldo lolullonusod. ~.O N sodium chlorldn lolullon wal cholllO 10 Ir(lat
(,(1,ln lampl"l In Ihn analYI" of ralln capa'cltlnl.

6.1.:1

Dul",mlnallon of rellin copoclly

lhii c"pllclly of IlIsln con b" obl~lnCld flom Ih" "nd-polnl of

0\

Illrallon curvo

~h,'wn 1n S(lcllorl 6.1.2. A fall(l( wuy 10 obluln Iho rOlln cupoclty Is by Iho

FlshM Al',lornallc Tltrlrnul~r. Aboul 3.0 ml of r~lln wal

Iralllformad with 010 ml
.,...;
~

pipe'l Inloa 10 ml momurin\') cyllndor. Tho pockod bod voluma"'-'<1l r"cordod. 'Thn
r",in wm Ihnn rlnsad will, dlstllind walar In a Buchnnr fun""l, and Intnrporllcla
moill,,;" conlant I"movod by lucllon. Aftor Ihn addilion of Iho rOlln to 100 ml of
.1.0 N ,odium chlorido,olutlon 1n a 250 OIl bnakc>r, Iho bMkur wal placod on a

r,.

Fl.,xo-Mix and tho mixlur" aoltalad by a maonotlc stirrar for obout flva mlnutol.
llw ",ixlum wal thon litral"d with 0:1 N sodium hydroxlda lolullon praparad from
concQnlralnd valu",ulrlc lolullon. Tha~chonoa In pH valuo wal followod by tho

,
.

41
\

calomel/glass electrodes connected to the calibra~ed

.
, automatic titrimetor. Tho
!i trimeter automatically stopped the titrant flow when the end point was rcached.
The resin capacity in meq/m! wet volume was then calculated as

Y NaOH NNaOH ,-'

resin capacity = .
.

(5.1.3.1)

YResin (1- E)
'.

The accuracy of the analytical resul Is using.,the

Fisher Automatic Titrimete'r
,
j

and the Orion pH/mY digital meter were compatible as had been confirmed
experimentally. In fact, the main source of error in the resi" capacity .analysls is
in the measurement of resin sample volume and in reading the titrant volume. \The
')

accuracy can be improved by measuring the resin sample on a weight basis. (

However, the drying of resin sample in an oven is time consuming and thi. slight
,

oJ the

gain in accuracy is likely to'Jbe masked by the random errors in other parts

6.1,4

.<

mass-transfer experiment.

Determination of chloride concentration in solution

Chloride concentrati9n in a sample soluti?n was measured directly by the

Orion digital pH/mY meter with chloride specific ion and double junction reference
electrodes. A relative mY reading was obtained' from the digital meter, and the
chloride concenlration read from the calibration curve prepared with standard
!1
sodium chloride solutions. Before a' series of measurements, the Orion meter was
re-calibrated against the existing calibration curve with a standard solution.

It was later discovered that the mY measurement by the electrodes was a

fu~ctiol1 of the ocidity of the soluion.

Figures B.5 and B.6 show the change in

mV I"c(ldil1J v/ith changing hydro90n-ion concentration aild pH vCllllcS respectively.

fooi, duto point was obtained by pipotting 100 rnl of ocie! solution of th~ de,i .."cJ

h/,dr()~lCn-ion concentration into a wei~Jhed s::Jnlp!c of ~odilJrn chloride

(10.35

Girl

for

" 0.0:, i'< ,olution) one! thl' relative mil readinu w,,; tokon "fter thorough
. 1,1ixin'l.
.
Sulphuric acid and nitric ocid solutions were u,,,d to rrepme ti,e solutions.

Re,ulrs

shnw('d lhat the relative mV ml!asuremcnts Wf'rc inJcp~nclcn~ of the sulphate and
nih'ak unions.

Figure B.6 shows that the r"lative mY roauin!) io comtant from 'a

pH vl1lue of 3.::; to at Icast 7.0.

Hence Ihe chloride concentrations in solutions

will, pH values between 3.5 and 7.0 can still be correctly estimated by cJirc:ct

,
nl<'o,.urement uoin!) the Orion cJigital.pH/mV meter calibrated agoinst

(J

standard

sodium chloride solution having.a pH value in Ihe same range.

6.I,S

Dekrmination of hycJrogen-ion canccr;rtrfJlion in solution

TI", pH valu" of an aqueous solution VIm dct,>rmined by direct measurement

. wilh 11r,' Orion digital pH/mV meter using calomel and glms eI"ctrodes. TI", m"ter
WCl'

culihraled against buffer solutions of pH value, 7.0 and 4.0 with oJju>tment

f"" solution temperature different from 25C.

,

'

II \'1m experimentally confirmed that the pH valul! n"'CJsurement by the cclom,,1

and 01"$0 ,-,"'ctrodes Wos

illd"pendent~he sodium and,c1r1,)ride

ion concentrations

'" Ihe solution.

~

6.1.6

Dotcrminotion of packed resin bed voido!)c

Tire voidages of packed resin beds wcre found cxp,,,imentally.

Thcrcsult far

"II particle-si",c distribvtions (-20+30, -30+35 and -35+50 mesh) is 0.33. Ahout

\,
43

10 ml of resin with interparticle moisture removed by suction were added to about

15 ml of distilled water in a measuring cylinder. The measuring cylinder was
vibrated until the volume of the packed resin bed became constant and the valume
was recorded. Theactual volume of resin added was obtained by the difference
between the volumes indicated by the first and second water levels in the measuring
cylinder. With the resin bulk volume and actual volume, the resin voidage can be
co leu la ted.

6.2

Calibrations
The necessary calibration curves were obtained as discussed below.

6.2.1

Process solution feed rotameter

The calibr~tion .curve of the volumelTic flow of liquid against the rotameter
reading is shown in Figure B.7. The calibration was performed with tap water at
room temperature.

6.2.2

Resin feed rotameter

The volumelTic calibration curve of the resin feed rotameter is shown in

Figure B.S. Water wos used for the colibration.

6.2.3

'.

Variable speed Masterflex tubing pump

The calibration of water volume:-flow against the pump-conlToller's speed

setting is shown in Figure B. 9. The calibration was only a guide to the approximate
pumping rote. The liquid f1ow-rate changed with the

agi~g

of the 1/4-inch

(6.4 mm) tygon tubing in the pump head, and it also.depended on the liquid-head

44

pre",ure.

6.2.4

Resin feed
The calibration curves of the resin volume flow-rate versus the resin-feed

rolomete,.!:.-reoding for the three different resin fXlrticle-size distributions (-20 +30,
-30+35 and -35+50 mesh),are shown in Figure B.IO. The actual resin volumeflow was considered in the calibration curves. The volume flow of resin at each
rotameter reading is higher for smaller fXlrticles, i.e. fraction of water in the resin
feed is Ie",. Figure B.II shows the change in the volume fraction of the resin ,in the
tolol flow as a funcJion of the rolometer reading ~ The curves indicate that at high
to~al flow-rates the volume fraction of resin in the feed becomes constant.

6.2.5

Water in resin feed

The calibration curves of water flow in the resin feed for the three different

resin fXlrticle-size distributions are shown in Figure B.12. For the same tala I
volume-flow (water and resin) the volume fraction of water is higher in the resin

fe~d of larger fXlrlicle-size distribution.

'< ,~
\

6.2.6

Chloride-ion concentration
The calibration of chloride-ion concentration versus the electrode potential

reading (mV) of the Orion digilol pH/mY meter with chloride specific ion and
double junction reference electrodes is shown in Figure B.13. Sodium chloride
solutions of concentrations 0.1,0.08,0.06,0.04 and 0.02N were prepared with
distilled water. The electrode potentials were then measured with the digilal meter.

The logarithmic value of the chloride-ion coocentration was plotted against the
corresponding mV reading as shown. Subsequent measurement of chloride-ian
.

concentration was obtained by reference to the calibration curve, justified only

For solutions having a pH value between 3.5 and 7.0 .
.-'l

6.3

Resin Drying
The resin drying column is shown in Figure A.20. About I litre of resin was

poured inta the column at the tap and then air was turned on at a moderate flaw-rate
to dry the resin. Each botch of resin required about twenty hours to become free
flowing. The free flowing resin was removed at the bottom outlet after disconnecting
the a ir hose.

6.4

Resin Size Screen ing

The Dowex- 50W-x8 hydrogen-form resin hod a wet size-distribution of

-20+50 mesh.

Botches af 500 ml air-dried resin was placed an a column of sieves

with 20, 30, 35 and 50 mesh screens and was vibrated by an electric sieve-shaker.
Each botch of resin required about three hours For a reasonoble separation into
size-distributions: -20+30, -30+35 and -35+50 mesh.
After screening the resin was re-hydrated with concentrated hydrochlaric acid
as a precaution against resin-cracking during hydration. The resin and acid mixture
was then diluted gradually with distilled water so that the resin could swell slowly
without cracking up.

46

6.5

Resin Regeneration
Figure tI.14 shows the change in resin capacity during regeneration by 10%

(by weight) hydrochloric acid. The resin bed was about 40 inches (1.0 m) deep.
This regeneration represented the elution by three bottle, of 6 lb. hydrochloric
acid (37%) diluted to 10%. The acid Flow-rate was 10.0 ml/sec. The elution
curve indicates that 12 lb. of concentrated acfd can adequately regenerate the
column of resin completely when a lower elution Flow-rate is used. As a precaution,
2.5 gallons (9.5 litre) of spent acid at the end of one regeneration run were used

beFore the 5 gallons (18.9 litre) of Fresh 10% acid.

The resin-regeneration system is shown in Figure A.21. Distilled water was
pumped into the exhausted-resin bottle'(6~, at an initial rate of about 4.0 ml/sec
I
and gradually increased to 10.0 ml/sec when a smooth resin Flow was obtained From
the bottle, through the Y4-inch (6.4 mm) tubing and into the regeneration column.
While the column was being filled, a slow Flow of distilled weiter was pumped
through the column to keep tha resin bed loose for the regeneration acid flow later.
When all the resin was transFerred to the column, the pump (P2) was turned off and
the pressure in the bottle released. Recycled acid in bottle 65

was

pumped through

the column of. resin at about 4.0 ml/sec. Then the 5 gallons (18.9 litre) of Fresh
10% hydrochloric acid was pumped through the resin at the same flow rate. When
a~proximately

2.5 gallons (9.5 litre) of fresh acid was left in bottle 6 , the liquid
6

effluent from the regeneration column waS switched over from the drain to the
recycle-acid bottle. The 2.. 5 gallons (9.5 litre) of acid collected' was used for the
n~xt regeneration run.

'1

47

During the regeneration cycle, the acid feed wa~ occasionally increased to

about 11.0 ml/sec to mix the resin bed in order to obtain a homogeneous
regeneration.

6.6

Resin Washing

,r

Re;in washing was performed in the regeneration column as shown.in Figure

A .21
. A resin bed, 40 inches (1.0 m) deep, required a wash time in the order of
12 hours at a distilled water flow-rate of 10 ml/sec as shown in Figure B.15. About
two hours were required to wash out the acid in the bed void and then ten hours
to remove the acid inside the resin particles. An increase in water flow satisfactorily
reduced tho time of displacing the acid from the bed void but did not help much in
reducing the washing time in the diffusion stoge.
During the washing cycle, the acid bottles were isolated by closing valves
V22 and V23. Then valve V21 was opened and distilled water in the storage tonk
(T3) was pumped through the bed of resin. Washing was stopped when the liquid
effluent hod a pH value of about 5.2
To obtain an effluent pH value of 7.0 was not found possible. This may be
due to the fac t that Dowex-50W-x8 ion-exchange resin is strongly acidic and whose
dissociation thus causes the wash effluent to remain slightly acidic.

6.7

Factors' Levels for the Fractional Factorial Experimental Design

The two levels far each factor (dimensions of geometrical structures on each

troy and the resin particle-size distribution) for the fractional factorial experimental

48

design are listed below

resin particle size: -20+30, and -35+50 mesh
weir length: 3.380 and 3.970 inches (8.59 and 10.08 em)
weir height: 1. 0 and 2.0 inches (2.54 and 5.08 em)
fluid-diode lateral displacement: 1/20 and 3/20 inche (1.27 and 3.81 mm)
downcomer clearance: 1.0 and 1.50 inches (2.54 and 3.81 cm).

The two levels for the solid and I iquid flows hod to be chosen such that the
continuous countercurrent ion-exchange column was operable under the various

,
treatment combinations.
To observe the operability of the column under various treatment combinations,
the modified ion-exchange column system, discussed in Section 4.1.2, was used.
Observations of the s~lid and liquid interaction in Sections I, 2 and 6 were enough
for the present study.'
It was believed that the downcomer clearance would not affect the solidliquid interaction exten~ively. Hence, the present study used only one level of
dawnca:ner clearance,I.O inch (2.54 cm),so as to reduce the number of experimental
runs by one-half.
Various combinations of the'two levels of the factors (weir height, weir length
and fluid-diode lateral displacement) were installed on the trays of the column.
For each structural combination, the ion-exchange column operation was started as

.
discussed in Section 6:9, with distilled water as the feed liquid. The "perfectly"
operable ranges of liquid flow-rate were observed at resin flow-rates of 0.6 and

49

2.25 ml/sec. The column is considered "perfectly" operoble when the resin bed';s
evenly fluidized and the resin flow thro~gh the downcomer is not cyc\lic. These
observations were made for both resin particle-size distributions (-20+30 and
-35+50 mesh), arid the results are tabulated in Tables B.I and B',2. The operable
ranges of the liquid flow-rates for the two tabulations are clearly displayed in
Figure B.17. An extremely narrow range of liquid flow-rate around 53.0 ml/sec is
operable for all the treatment combinations with the resin flow-rates at either 0.6
or 2.25 ml/sec.
A study of Figure B.17 shows that the operable range of liquid flow can be
increased by trying to reduce the lower limit of operability of the runs with large
resin size-distribution (-20+30 mesh) and to raise the upper limit of operability of
the runs with small resin size-distribution (-35+50 mesh).
The upper I imit of operable liquid flow-rate for the runs with the small size
resin can be raised by using resin 0(-30+50 mesh size-diS:tribution instead.

decrease in solid feed-rate also helps. Hence overfluidization for operations with
the low size-level of resin will occur only at a higher liquid flow-rate. The
decrease in resin flow-rate also reduces the lower limit of the liquid flow-rate
below which flooding of downcomer will occur. Hence for the two level experimental design the levels of particles size were chosen at -20+30 and -30+35 mesh
and oolid flow-rates at 0.84 and 1.6 ml/sec. The operable range of. liquid flawrate was broadened and its levels for the experimental design were chosen ta be
40.0 and 79.2 ml/sec. The difference between th~se two levels af liquid flaw-rote
was sufficient for the factorial experimental design.

"'""

~\

\.",

50

6.8

Operational Performance of the lon-Exchange Column

The equipment set up for the operational performance study was similor to

that for Section 6.7. Various weir shapes (straight, cylindrical and conical), weir
heights and fluid-diode lateral displacements were installed in the three sections
(I, 2 and 6) of the ion-exchange column. The column wasstarted as discussed in
Section

6~9,

bU,t d i e d water was used as the feed liquid. At a given resin flow-

rate, the operational behaviour in terms of fluidization, flooding and the continuity
of resin downflow in each stage, was noted as the liquid flow-rate was gradually
decreased such that the resin bed changed from the state of over-fluidization to the
state of flooding of the feed stage through sluggish resin movement. Near incipient
Fluidization, the column operation became unstable and a quick decrease in liquid
flow would cause premature flooding of the feed stage.

6.9

Moss.Transfer Experiment
The mass-transfer experiment to determine the overall stage efficiency of ,a.

particulor column operation was performed by the equipment as shown in Figure A.6.
A weighed quantity of s~um chloride was added to the two storage tanks
T2) of distilled water.

Val~eJ-NI
.'

and V2 were opened while V3 and V4 were

closed. The solution was then mixed by pumping in a closed loop. After two hours
of mixing, the chloride cancentration in the two tanks was measured with the Orion
digital pH/mY meter using the chloride specific ion and double junction reference
clectrodcs. Extra sodium chloride was odded when required. Four or five hours of
mixin3 WCfe normally required before the 90 gallons (34O litre) of sodium chloride

51

solution became hamogeneous.

The temperature af the feed solution was raised above room temperature by
the heating coil so as to liberate same of the dissolved air. The thermostat was set
at the desired temperature, and the on-off control wos turned on. The hot water
coil raised the solution in the tanks to the desired temperature, and maintained it
at the set temperatufe. The heating ensured the avoidance of a ir bubbles being
liberated when the solution was rising through the column. Bubbles trapped under
the screen of each stage would have affected the liquid distribution.
During start-up all the Y4-inch (?.4 mm) drain holes and vent holes except
VI6 were closed by spring clamps on the connecting Y4-inch (6.4 mm) tygon
tubing. Valve V2 was closed and valves V3 and V 4 opened, and feed solution was
pumped into the column at obout 3.0 ml/sec. A slow liquid flow was needed to
ovoid trapping air bubbles under the liquid distributors and screens. When the
column was filled to section 6, the solution overflowed into the downcamer area
and into section 9. Valv~s V9and VIO were opened so that the resin bottle (BI)
was also filled. When the bottle was filled, valve VIO was closed. When section 9
was filled, valve Vl6 was closed. While section 9 was being filled, the cavity at
the end of ,the sampler's plunger was turned upward so that air bubbles would not be
trapped in the cavity. Gradually the entire column was filled and the solution
overflowed into the trough and to the cyclone. Since there was no air vent in the
resin collector, B3, the cyclone and the bottle were disconnected to release the air
while the bottle was being filled. Then solution feed to the column was shut off
until the resin-feed system was re(]dy. Valve V 4 was closed and valve

\"J

opened

52

to recycle the feed solution.

The resin-feed system was started up by opening valves V , V , Vs and V
6
7
I2
while closing valves VII, VI3 and V14' The Mosterflex tubing pump, P2' was
activated and distilled water was pumped from storage tank, T , to the resin-feed
3
battle (B2)' When all the air was displaced thraugh the air vent, valve V S was
closed and valve VII was apened. The resin feed-rote to the column was then
adjusted by the needle valves V6 and V7'
Before the sampler could be used, the sample collector was filled with
distilled water by activating pump P3. When all the air was displaced and the
pump was turned off, valves V17, VIS and VI9 were closed. This procedure
prevented the introduction of air into section 9 of the column during sampling.
The moss-transfer experiment was then ready to be storted. The desired resin
and process solution feed-rates were set and the continuous countercurrent operation
usually would start automatically. When a run involved a low liquid-flow (e.g.
40 ml/sec) and a high sol id-feed (e.g. 1.6 ml/sec) flooding of the feed stage would
occur dWing start-up. Hence a higher liquid flow-rate had to be used and gradually
reduced to the desired flow after a continuous and smooth resin flow was obtained
down the whole column.
When steady state was reached, four 100 ml samples, of process feed solution
were taken at the recycle outlet. Four samples of liquid effluent were collected
at the outlet of the cyclone. The assurance of steady state could be obtained by
checking the pH value of the liquid effluent from the column. Figure B.16 shows
the change in pH value of the liquid effluent for a Iypical run. Steady state was

53

obtained in about 10 minutes from stori-up. Resin effluent was sampled at section 9.
Resin wqs drown into the sample collector for two or three times. The plunger was
moved slowly to avoid disturbances to the resin downflow. The sampler cup was
unscrewed and resin transferred to 1000 ml of distilled water. The resin sample
was washed, by two litres of distilled water in a Buchner funnel. The sample was
then stored in 200 ml of distilled water. About 20 ml of resin from the feed battle
was also collected at the feed line forwhe analysis of initial resin capacity. Each
fresh and exhausted resin sample was enough for four chemical analyses.

/
1

".
;

,..
~

7. DATA REDUCTION
/

1,1

Mm,-Tramfer Experiment '.

For <'Cleh man-transfer oxperlmonlal run, Iho resul It from Iho chemical

on"I}'"" Wl'ro usod 10 evaluata Iho mass-tra(lsfar performancn of Iho lon-nxchango

op,'r~ion.
,

lion,

,-

A compuler program was wrillon In Farlran

IV to perform the compula,

Tilt' program was proeonod by Ih" CDC 6400 Compulor al MeMOller

Tho overall Ilage.offleleney, Iho volumetric offlcloncy as well as olhor

--;,

"xpurim,;nlol conditions and rOlul1t w.oro calculatod. Man balances on hydrogen

ion (lnd ehloridil baStll woro chocked, The socond half of .tho cornpulnr program
"~--\

. I" <"l ropln1od Iho Ilons;n-Lehner pi oller 10 plot thn e'qulll b~l ~m diagram for each
t.xp<lrinwnlal run, Tho (1qulllbrlum diagram showod Iho analysis operating Iino and

,,1,0 Iho flow opllraling IIno If Iho dlffertlnco betwoon.lho slopos of tho Iwo oporaling
~

lin'" wm grooter than the tolerance

I"" I"eorotical

101

in Iho computer program. The stopping off of

slagos was Included.

Th" g"ner~1 flow chari of tho EOii1p\Jtor program ISlhown In Flguro G .1. A
lompl" compuler oulput and th" program listing aro shown In Appendix G.

7,2

Fractional Factorial Exporlmonlal Do~n

';-t, . A [Joneral compulor program was wrl lion in Fortran IV to porform Iho analysis
r

~'

54

55

of variance for fractional factorial experimental designs. The program wal divided

into three functianal parts. The pass number in tho data deck cantrolled the section
of the program to be used. Tho following results were obtained in the computer
output whon oppropriate pass number and data were supplied in the dota deck

Poss I:

,Yotes' Analysis, and absolute overage effects listed in

ascend i ng order

Poss 2:

Those of Pass I and standardized absolute ?veroge effects

Iistcd in ascending order

Pass 3:

Those of Pass 2 and analysis of voriance table.

The' general flow chart of the computer program is shown in Figure H.I. A
sample computer output and' the computer program listing arc' shown in Appendix H.

"

8. RESULTS AND DISCUSSION

8.1

Fluid Diodes as Downcomers

8.1.1

Effects of the fluid-diodfl designs on resin flew

The three fluid-diode shapes shown in Figure A.5 were installed as separate

downcomers in the multistage ion-exchange column. The qualitative operetionol

eHec Is of the thre" different fluid-diode shapes (A, [land C) on the resin flow were
observed. The comparisons were performed at zero fluid-diode lateral displacement
for the downcomer in each stage ondowith the 3.78-inch (9.6 cm) straight weirs . .
In the present studt-tbe fluid-diode lateral displacement was defined as the distance
between the cusp end points of the two opposing fluidic elements when observed
lengthwise.
A qualitative comparison shows that fluid-diode types A and B give less
resistance to resin flow than type C and hence resin flows faste .. through these downcomers. The smaller resistance to resin flow enables a column to handle a higher
resin feed-rote before th'; flooding of the feed ~toge occurs.
However, at

I..,w resin Flow-rote, operations with these fluid-diode shapes

(A and B) are more sLceptible to cyclic flow of resin, i.e. alternating solid-flow
down ond I iquid-byposs up the downcomers. This is because the resin goes down
the downcomer so Fast that there is not enough resin to maintain an equilibrium
resin-head in the dawncomer area against.the upward flaw of liquid . Hence liquid
bypasses up the downcomer. The resin then accumulates in tlfe downcomer area to

56

57

such a height that the resin dawnflaw can be initiated again.

The dawncamers with fluid-diode shope C offer higher resistance to the
liquid flow and henc!} require a lower solid-head to initiate resin flow and also a
lower equilibrium resin-head to maintain a contin~aus downflaw of resin. A higher
resistance to resin flow in the downcomer also means a slower response of the column
to any change in resin or liquid flows in re-establishing steady state operation.
"

A co~porison between types A and B also ~i'icatesthot. the extended cusp

ends of the former off,;r higher resistance to Iiquid flow and hence require a lower
resin-head to

initiate~sin flow ~nd to maintain a continuous downflow of resin.

Do~ncomers

of different fI uid-diode shape~ have different ranges of

operability in terms of the r~sin and liquid flow-rates. A change in liquid f1.awrate changes the voidage of the fluidized bed, the pressure-drap across the fluidizes!
bed and hence the pressure drop across the fluid-diode' dawncomer. The operable
range" of the column would be smaller if the fluid-diode downcamers installed in
the column ore more pressure sensitive.

8.1.2

Effects of fluid-diode lateral displacement on resin flow

Experimental runs were performed with the 3.78-inch (9.6 em) straight

wel~'"

and using type B fluid-diode downcomer. The fluid-diode ,lateral displacement los
varied between 1/32 and 3/32 inch (0.79 and 2.38 mm).
For a given liquid flow-rotc, a decrease in fluid-diode lateral displacement
reduces the resin-head required to initiate the downflow of resin in the do~ncomer,
and also the equilibrium resin-head required to maintain a continuous flow of resin.

58
A larger fluid-diode lateral displacement such as 3/32 Inch (2.38 mm) not only
increases the resin-head requirement but also causes cyclic flow of resin from
stage to stage. Increasing the resin feed-rate will reduce the frequency of intermittent resin flow until the maximum solid handling capacity is
~he

reach~~?

effect of fluid-diode lateral cflsplacement on the resin flow can be

qualitatively explained by considering the balance of forces acting on the'bulk of

resin in and immediately above the downcomer. When the resin flow In the fluiddiode is fully doveloped, tho forcol acting on the rasin oro tho weight of tho solid
resin (mg), the buoyancy force (f b), pressure force acroSS' tho fluid-diode downcomer
((P2- P I}A), ond the fluid-diode's force of resistance to the downward flow of resin
(R ). Hence,
I
(8.1.2.1)
where PI and P2 ore the hydraulic pressures at the entrance and exit of the fluiddiode downcomer respectively, and A is the or co of the openings at tho onds of the
downcomer. An increose in fluid-diode lateral displacement increases tho hydraul ic
pressure force ((P2-PI}A) across the fluid-diode downcomer but decreases the fluiddiode's resistance (R ) to resin flow. Apparently, the increase in pressure force Is
I
higher than the reduction in fluid-diode resistance and hence, as observed, a higher
.'J

,~I'

resin-head (mg) is required to maintain a continuous resin flow.

Similarly,

force balance can be written at the moment of impending resin

downflow
(8.1.2.2)
where R2 is the fluid-diode's force of resistance to the bypassing flow of liquid and mg
the resin-head required to initiate resin flow. An incrcose in fluid-diode lateral displace:-

ment increa~cs tho pre~surc force but reduces R , .resul ting in a higher resin-head
2
for the initiotion of rosin flow.

8.2

Operational Performance of the lon-Exchange Column

8.2.1

Effec ts of resin and liquid flow-rates

Figure C.I shows the performance of the cdumn under various combinations of

resin and liquid flow-rotes. Each

s~ge has a

straight weir and the dimensions of the

stage structures oro inc luded in the diagram.

The performance diagram can be divided into six regions. In all the regions,
except the one to tne left of the dolled line and labelled "liquid bypassing", resin
flows through the downc~."ers are continuous and the rates are constant, ,at leost
visually. At tho centre is the region of "perfect" operation. In this region the
resin bed in each stage is fluidized completely and smoothly (Le. homogeneous).
In the fI-,!idizatian orca tho height of the fluidized resin-bed is slightly higher than
the weir height and the resin in the fluidization orca overflows smoothly into the
dOWnCO",er area. In this region above (labelled "over-fluidization of fines"), the
resin particles experience a high superficial liquid velocity. The terminal velocity
is reached far the smaller resin particlos at one end of the size distribution and also
for Ihe crocked resin pieces. The fine particles Float above the fluidized bed an(l
even stay on the lower surface of the screen above. In this region, the column is
perfectl y operabl e. A further increase in liquid superficial velocity over-fluidizes
the majority of the resin particles. The fluidized-bed height becomes equal to the
stage sp~cing and flooding occurs. The boundaries, AB and CD, curve down as
the resin flow-rotc is increased since, of the high resin flow-rates, the fluidized

60

bed is churned up by the strong transverse flow of resin. The momentum transfer
causes the resin bed to become over-fluidized at a lower superficial liquid velocity.
When the superficial liquid velocity is lowered bel.ow the region .of "perfect"
~-

operation, tho resin bed in each stoge becomes partly fluidized due to channeling.
Further reduction in liquid flaw-rate causes sluggish resin movement across the
fluidization area and also down the fluid-diode downcomer which has become filled
with resin by then. Finally the feed stage is flooded because the resin feed-rato is
higher than th.o flow-rato of resin down the column. The boundary for flooding (EF)
goes up as the resin feed-rate is increased, because th!0>igher the resin feed-rato
the sooner the resin downflow fails to cope with tho input feed.
Cyclic flow of resin in the fluid-diode downcomers occurs left of the dotted
line. The resin flow is still continuous and some liquid-bypassing is obvio~s. For
a given resin feed-rotc (e.g. 0.4 ml/sec), liquid-bypassing occurs ct some
intermediate range of superficial liquid velocity (0.7 to 1.6 cm/sec). This must be
the region that the fluid-diode downcomer experiences the highest pressure drop, as
the superficial velohity

increased the fluidized-bed's voidage increases and hence

l'

the resin hold-up decreases, resul ting in a decrease in the pressure drop across tho
fluidized bed. Ncar the incipient fluidization velocity channeling occurs in the
resin bcd, causing a decrease in the pressure drop across the bed. Channeling
frequently occurs at incipient fluidization when screens are used as liquid
distributors (50). Liquid-bypassing docs not occur at higher resin flow-rates
because the plentiful supply of resin is sufficient to maintain an equilibrium resinhead in the downco:net orca to counieract the unfavourobly high pressure drop.

61

8.2.2

Effects of resin particlo-size distribution

. The performanco of the ion-exchange column wi th tho same stoge set-up as

discussed in Section 8.2.1, but operatod with -20+30 moshresin, is shown in'
Figure C.2. Tho basic porformance pattern is tho sarno as in Figuro C.1.
Tho region for over-fluidization is missing boca,use the terminal volocity of
these larger resin particles Is highor. Tho large particle-size and honce the
reluc tonco to mo~bilizo increosos the orcas of the partial mobilization and flooding
regiom.

Tho largor particle-size distribution also glvos a smaller fluidized-bed

voidage and a larger resin holdup rosulting in a higher pressure drop across the
resin bed. The increase in pressure drop is tho reoson for the larger rogion of
liquid-bypassing in the performance diagram.

8.2.3

Effects of fluid-diode latoral displacoment .

The performance diagram for the column operation in Soction 8.2.1 with tho

fluid-diode latorol displacemont Increased to 3/20 inch (3.81 mm) is shown in

Figure C.3. The result is on increase in the regime of liquid-bypassing up tho
downcomcrs and can be explainod by the same principles discussed in Section 8.1.2.
The region of partial mobilization is extended to tho right and its boundary moets
the floJding line at a higher resin feod-rate~ The increase in fluid-diode lateral
L"

displacement lowers tho fluid-diode rosistance to resin flow.

Hence the resin flow

in the downcomer is loss sluggish even at a low superficial li.quid velocity and the
I

column can c080 with a higher resin flow-rote before the flooding of feed stage
! .

takes place.

62

8.2.4

Effects of wair height

Tho performance diagram of the column operation of Soction 8.2.1 but with

I-inch (2.54 cm) height for the weir$ls shown in Figura C.4. Tho basic pattern is
tho same as in Figuro C.1. This sha Ilower fl uidi zad-bed means a lower pressuredrop across the troy and hence no apparent liquid-bypassing in tho downcomor.
The region of partial mobilization is more dominant due to tho poor liquid distribution in a shallow bed and hence serious channeling. A lower weir height also
means a shorter downcomor seal.

Henco tho re$in flow out of tho downcomer is los$

restricted and flooding becomos independent of the rO$in Feed-rato. Flooding of

the feed stago occurs at a vary low superficial velocity which i$ close to the
/

minimum fluidization velocity.

Figure C.5 shows the porformanco diagram of the column whe~ tho fluid-diodo
lateral displacement is also changed. The fluid-diode lateral displacement is
increased to 3/20 inch (3.81 mm). The performance diagram is the Somo ~s that in
,-

Figure C.4 except for tho extra region of intormittont resin flow (in the downcomers).
The intermittent resin flow is portly due to liquid-bypassing, and portly due to somo
stagnant resin particles, blocking the downcomer's exit.

8.2.5

Effects of conical weir

The performance of the column with conical weirs of diameters 1.28 inches

(3.25 cm) and heights 2 inchas (5.08 cm) is shown in Figure C.6. The other
structural parameters of the tray are as shown in the diagram.
The performance diagram can be divided into six regions similar to those for

63

straight weir operations. However, tho orca of the rogion for "porfect" column
operation is reduced due to excessive fluidization at a lower, and partial
mobilization at a higher, superficial liquid volocity. Tho loworing of tho upper
boundary may be caused by tho interaction between the diverging shapo of tho
conical weir and tho upward liquid flow. Tho liquid going up the side of tho
conical weir deflects resin particlos away from tho downcomer area.

Henco flooding

(i .". fluidized-bed height equals stago spacing) occurs at a slightly lower superficial liquid velocity when compared to tho operation with straight woirs. Tho
interaction between the conical woir and tho insido column wall may also havo
caused the occurrence of partial mobilization of tho rosin bod at a higher suporficial liquid velocity. The conical weir is situated ncar the column wall as shown
in Figure A.16. The transverso flow of resin from tho downcome-r sido to the woir
side of the tray results in somo stagnant resin areas on the sidos of tho woir away
from the downcomer (areas laoellccl<'''A'' in Figure A.16). Thoso stagnant rosin areas
became do.-ninant as the superficial liquid velocity is lowered.
The patterns of liquid-bypassing and tho flooding of tho feed stago arc
similar to column operations with straight weirs.
Figure C.7 is tho performanco diagram of tho column with a conical weir of
smaller diameter, 1.03 inches (2.62 cm), but of the same woir hoight. Figures C.6
and C.7 arc similar. The upper limit of superficial liquid volocity for "perfect"
operation is about the same for both sizes of conical weirs. The largor area of
partial mobilization for the operation with smaller conical woir may be due to
larger stagn:mt areaS at the weir and due to the lower effective fluidization supcr-

64

ficial velocity experienced by tho resin as can be understood from tho difference
.

In stcepness of the side walls of the two sizes of conical weirs.

"

The effects of resin particle sizo, fluid-diode lateral displocement and weir
height on the performance of the column operating with conicol woirs oro shown in
Figures C.8, C.9, C.IO and C.II in Appendix C. Tho performance patterns of
these operations can be explained as in the corresponding straight weIr oporations .
The regions of liquid-bypassing for the conical weir operations occur at a lower
resin fced-rate than thoso for tho straight weir operations. This is certainly duo to
the el iminotion of dead downcomer areas insido tho conical weirs. Hence tho
solid heod of resin can ~moro efficiently maintained by the f1u'idized resin particles
overflowing over the weir and into the dawncomers.

8.2.6

Effects of cylindrical weir

The performance diagram of the column operation with cyli~drical weirs is

shown in FigureC .12. The basic pattern of the performance is similar to that for
stroight and conical weir operations. The upper boundary of "perfect" operation
is slightly higher than that for the corresponding conical woir operation. This is
due to the smoother pattern of liquid-flow around the vertical aU,tside wall of the
cylindricol weir. The existence of large dead resin area between the cylindrical
weir and the inside column wall (areas labelled "A" in Figure A.I?) is the reason
for the large region of partial mobilization at lower superficial liquid volocity.
The inner sloping woll of the cyl indrical weir 01 iminates dead downcomer
areas and hence I iquid-bypassing occurs only at a very low resin flow-rotc.

65

Figure C .13 is the performance diagram for the col,umn operation with tho
cylindrical weirs using -20 +30 mesh resin.

8.3

r~actianal

Factorial Experimental Design

,

Th;-32 treatment combinations were performed for- the

~4-replicate

factorial

experimental design fOf 7 factors at 2 levels. The experimental runs which had
mass-balance discrepancies in hydrogen ian greater than 6.5% or an initial
sodium ion concentration deviation greater than 0.0035 N were repeated. A positivemass-balance discrepancy was likely due to errors in the sampling of resin and
liquid effluents before the steady state was reached. The level of the initial
sodium ion concentration in all treatment combinations was chosen to be 0.06 N.
The resul ts of the mass-transfer experiments were processed by the computer
program discussed in Section 7.2. The yield of the experimental design is the
overall stage efficiency of the countercurrent ion-exchange operation. The analysis

of variance on the average effects shows that out of the seven main factors and
twenty-one 2-factor interactions the following are significant:
Main effects:

resin flaw-rate
liquid flow-rate
weir length
weir height

interactions:

liquid flow-rate and weir length

solid flow-rate and weir length
weir height and weir length.

66

The analysis of variance table in Table E.B of Appendix E shows that these main
effects and interactions are significant at the 95.0% and 99.0% levels of
confidence. The other main facto<s {resin particle size, fluid-diode lateral

displacement and downcomer clearance} and other interactions are not considered
to be significant in affecting the overall stage efficiency of the ion-exchonge
column because their effects are nat as large, as those due to random experimental
errors.

The use of half-normal and standardized half-normal plots confirmed the

above results as shown in Figures E.2 and E.3.

"
B.4

Effects of. the lvIain Factors and the Significant Interactions on the IvIassTransfer Performance

8.4.1

Resin feed-rate
The resin flow-rate is significant in determining the overall stage efficiency

of the ion-exchange operation. As shown in Figure B.18a, the increase in resin

flow-rotc normally reduces the overall stage efficiency. Figure B.19 shows the
"~

equilibrium diagram {with operating lines} of the experi,:,ental runs.

The faster the resin travels through a stage' of the column, the sho<ter is its
contact time with the flow of process solution., Hence the degree of approach
towards thermodyn:lmic equilibrium is less. The decrease in solid mixing in' each
stage as the solid feed-rate is increased also accounts fo< some decrease in the
over,,11 stage efficiency. The decrease in contact time decreases the transfor of
sodium ion to each resin particle. The mole fraction of sodium in the exit resin is

67

"",

thus reduced as shown in Figure B.ISb. However, the increase'inresin feed..,.ote

increases the ov~~all rate of counter-ions (No + and Hl exchanged in the cantactor
V

between the solution and resin phases as shown in Figure B.ISd.

Hence the mole

frac t ion of sod i urn ion in the liqu id phase decreases as shown in Figure B.18 c .
Under the present experimental conditions (Re p = 10.0), both ,'he
solid and
~
liquid sides playa significant role in the ian-exchange kinetics. In the multistage
ion-elichange column, the steady state compositions ,of the resin and salution
phases in each stage vary as fhe resin feed-rate is increased. Examination of
Figures B.ISb and B.ISc shows that the hydrogen-ion content both in the resin and
the liquid phases increases with the increase in resin feed-rate.

Hence mass-

transfer coefficients in both the solid and liquid sides decrease as the resin feedra te inc reases. This increase in overall moss-transfer resistance a Iso causes the
decrease in overall stage efficiency as the resin feed-rate is increased.

When the ion-exchange operation is concerned with the extraction'of ions

'

(Na+) from a solution (Noel), Figure B.ISe shows that' it is not economical to
increase the extraction yield beyond 65% by increasing the resin feed-rate. When
the opera'tion is concerned with the extraction of material (Hl from the resin
particles, Figure B.ISf shows that thc fraction of ions extracted from the, resin
,

decreases as the resin feed-rate is increased. Figure B.ISe is equivalent to

Figure B.ISd and Figure B.ISf to Figure B.ISb.

68

...
8.4.2

Process solution flow-rate

.,

The process solution flow-rate is significant in determining the overall stage

efficiency. Figure B.20a shows tbt the overall stage efficiency decreases as the
process solution feed-rate increases. Figure B.21 shows the equilibrium diagram
(with operating lines) of these experimental runs.
It hos been shown that the increase in mass~transfer rate by increasing the
liquid flow in fluidized-bed moss-transfer is nullified by the increase in bed voidage

(46). However, in the present

operation of a multistage
fluidized-bed column the
.
.
r
increose in voidage also resul ts in a decrease in resin holdup in e'ach stege and
Q

'hence a decrease in.;:asin r'esidence time in the column. For.a give[l resin feedrate, the increase in liquid flow-rote a.lso decreases the contect time of a given
vol"ume of process solution with the resin. On the other hand, the increase 'in liquid
"

flow improves the resin 'mixing in each stage. Apparently the degree of approach

,,~

..

to chemical equilibrium is reduced due to the more significant effect of decreasing

.the resi n and I iqu id con tec I lime
. The increase in solution flow-rate results in an increase in the mole fraction
of sodium in the resin phase, as.shown in Figure B.20b, because each volume of
resin is exposed to more sodium-ion contect. Hence the higher flow-rate of sodium
ion compensates For more thon the decrease in. the residence time of the resin.
For the solutio~asef the decrease in residence t.ime of the solutiorueduces
the' tronsFer of sodium ion From the solution'per'unit volume as shown in Figure B.20c.
However,' the overall rate of ion-exchange is i~creased as the solution flow is.
v~,

increased as shown in'Figure B.20d. The moss-transfer coefFicients in the solid

69

and liquid sides increase when tho liquid flow-rota increases becauso of tho
corresponding decroose in hydrogen-ion malo fractions in both phases.
When the ion-exchange procoss is designed for tho extraction of ion (Na+)
from solution, Figuro B.20e shows that tho increaso in solution flow only reduces
the fraction of ions romovod. ,Howover, when the, extraction of ion (H+) from tho
resin is of concern, the increase in solution flow does increaso tho fraction of ions
removed from the resin as shown' in Figure B.20f.

8.4.3

Weir height
The weir height is a significant factor in affecting the overoll stage efficiency.

Figure B .220 shows that the overall stage efficioncy increases as the weir hoight is
increased from 0.5 to 2.0 inches (1.27 to 5.08 cm): The equilibrium diagram
showing the changes in the operating lino with tho changos in woir hoight is shown
in Figure B .23.
'0

For given resin and process solution flow-rates,' the increase in weir height

in~r~as'es the re<1r:.holdup in each stago. Thus the resin residence timo in the column

The mole fraction'oF sodium ion in the resin effluent is increased and
\

consequently the- mole fraction of s~dium ion in the liquid effluent i's decreased as

'

shown in Figures B.~2b and B.22c,'respectively. With the increase'in weir height,
the compositions in the resin and liquid phases change in the direction of increasing \
the solid-side moss-transfer coefficient and docreasing the liquid-film mass-transfer
coefficient, respectively.
The increasing lock of ideal solid mixinfj with the increase in weir height' ,

\
70

apparently docs nat seem to override the other facters in reducing the overall stage
effie ieney.
Figure

B.22 d

shows that as the weir height is increased the overall rate of

ion-exchango increases and again can be uxplalned by the longer residence time of
tho rusin in the cantactar. Figure B .22 d also indicates that the increase in the
overall rato of ian-oxchange approaches a constant value as the weir haight is
increased.

Henca thare is a maximum wair haight bayand .which tha overall rata

of ian-exchanga dous nat increasa significantly.

: The frac tion of sodium Ion adsorbed f~. the solution by tha rasln flow and
the fraction of hydrogan ion axtracted from the rasin by the liquid flow follow tha
some trend as the overall rata of ian-exchanga whan tha wair height is-increasad.
These arc shown in Figures B .22 u and B.22 f.

8.4.4

Weirlcngth

;-

The straight wair length is a significant factor in affecting the overall stage
,efficiency. Figuras B.24a, B.25a, and B.26a show that for a given rasin faedrate, liquid flow-rata and wair height, respectively, the overallstaga efficiency
of the ion-axchange column increases with a decrease in the length of the straight
weir.

In the present tray desisn, the decrease in tho longth of the straight weir

corresponds to an increasa in fluidization area. The increase in fluidizetian orca

corresponds to an increase in resin holdup and hance an increase in resin residance
time. For a given liquid -flow-rate, the increase in fluidization area also dacreases
the fluidization superFiciol velocity and hence

1;

further increosa in resin holdup duo

71

..
I"

II

deere,,,,, in fluidized-bud voldogo. On tho othor hand, tho docrtlmo I~

"'I"riie,illl liquid voloclty causus an Incroaso In tho liquid-film mms-transfur

1,,j"I"ncu. i\pparontly, tho oxporimontol rosults of tho prasont study'lndlcoto that
Ih. .. fi,cl on thu ovorall stago officioncy duo to rosin hOldup I, dominant ovor that
dtJ(~ 10 incr{lo~inu

li(lUid-fill11 mass-transfor rosistanco.

Since tho chang" In ;"oir longth affect, tho procon offlcloncy mainly through
IIIl' ,,,,in holdup, tho change in compositions of'tho ro,ln and ,olutlon offluonts and
tl,, overall ratn of ion-exchango will follow tho samo trond, m tho$o discussod for
wei, hl'ioht changes in Suelion 8.4.3.

11..1.5

Fluid-diodo lateral displocemont

Th" chango of fluid-diodo latoral displacoment betweon 1/20 and 3/20 Inch

(1.21 und 3.81 rnrn) (a 3-fold chango) doos not show a significant offoct on tho ma,,, II ",dt'r performance of tho ion-oxchangu column. Tho absonco of

signifi~a~i

inl{"qctiom b"twoun the f1uid-diode lateral displacomont and tho athor six main
foclor, al,o indicatos Its negligible effeel undor any oporatlng conditions.
I-honce the increase In liquid-bypassing up tho downcomers whon tho flulddiode lateral displacement is Increased from 1/20 t~ 3/20 Inch (1.27 and 3.81 mm)

cia", not aff"ct the mass transfer to a significant oxtont. The bypassing roprosonts
onl y " very ,mall fraelion of tho tota'isolution flow, and tho bulk of maSS transfor '
,till luk", place In thu fluidization "rca on' oach stogo.

72

U.4.6

Downcomcr clearan/
The downcomer clearance as well as its interactions with the other six factors

arc not significOJlt in determining the ovorall moss-transfer stage efficiency. The
levels of tho downcomer clearance wore 1.0 and 1.5 inches (2.54 and 3.81 cm) and
so the downcomer seal ,,:orie. from -0.5 to 1.0 inch (-1.27 to 2.54 cm) (0 three fold
change). Hence, the effect of tho dawncomer .eal on the resin's residenco ti~o in
each stago is not large enough to couse any significant difference in mos.-transfer
performance.

8.4.7

Rmin particle size

The change in resin P9rticle-size distribution from -30+35 mesh to -20+30

mesh is not significant in affecting the moss-t~ansfer efficiency.

The mean particle diameter is varied from 0.507 to 0.718 mm and is 040%
increase ;'n diameter.

However, the result of the experimental design shows that

this change in resin particle-size is not significant in affecting the overall stago
efficiency. A smaller resin particle has a larger surface orca to valumo ratio and
also a lower moss-transfer effect due to solid-side resistance (when compared to
that of the liquid-side). However, for a given weir height and a given superficial
liquid velocity the resin holdup in the fluidization area is lower for the smaller
particl~':";ze distribution.

The d",crease in moss transfer due to a lower resin hold-

up may possibly just cancel out the net increase in mass transfer duo to a larger

surface orca to volume ratio and a lower mass-transfer effect due to solid-side
-

r('sistonce.

Hence the overall stage efficiency is indifferent to th? change in

73

pmticle siLe.

8.4.fl

Rosin food rato- woir longth intoraction

Tho intorac.!ion botwoon tho resin feedrato and tho longth of tho straight

Wl~lr

(i.e. fluidization aroa) on tho oV,orall stago officloncy 1\ shown in Flguro B.24a .

Tho torm "interaction" means that tho eHoct of tho resin foed-roto an tho o~
stClfle efficiency is difforont for difforont woir longths. Tho offocts an tho ov~
stooo efficioncy due to tho chango in ,-osin food-rato and woir longth woro
discussed in Sections 8.4.1 and 8.4.4 respoctivoly. Figuro B.24a shows that tho
respome of tho overall stago officioncy to tho resin foed-rate is very significant at
tho shortest woir length espociolly in tho rooion of low resin feod.ratos.

At a weir longth of 3.38 inches (8.59 cm), the decreasa in resin rosidonce
tillll' Clnd overall mms-transfer coefficient duo to incroasing resin foed-rato can
ClCeQUnt for the cortesponding decreme in ovorall stooo eHicioncy. However I at a
w"i, lonoth of~.97 inchos (10.08 cm) (smaller fluidization orca) the overall staoo
(Officil"ncy is indifferent to tho change in resin residence tlmo and overall masstrCl",fer coefficient. Ono possible explanation for this b<:hoviaur at tho weir length
of 3.97 inch .. s (10.08 cm) is that the initial part of tho ion-oxchango takes place in
o vl'ry ,hart ti'me m C(]n b<: understood by tho graphs in Figura B.I. Tho rosin

".,idl'nce time provid.,d by tho holdup in each stag,: is moro than enough for tho

cOlllpletion of this port of the reaction. The decreaso in resin residence timo by

.-

in(.remino the resin f.eed is not signif1cant enough 10 affect the ;'ampletion of this

reoct ion. As can be seen from Figure B .24,b and B. 24 c, the change In resin ancr"/ .......

'

..

"

,I

74

solution compositions is relatively small as the resin feed-rote is increased. Hence

the ov,srall stage efficiency rcmaim.practically unchanged.
The same concept may be used .to explain the approach toward a comtant
value of overall stage efficiency with increasing resin feed-rate for the operations
wi th slra ight weirs of lengths 3.38 and 3.78 inches (8.59 and 9.60 cm) .
.,

8.4.9

Process solution flow rate - weir length in,teractian

Figure B .250 shows the decrease in the overall stage efficiency as the process'

solution flow is increased for the three weir lengths. Tho effects of the chango in
solution flow-rate and weir length on the overall stage efficiency were discussed
in Sections 8.4.2 and 8.4.4 respectively.
Figure B .250 shows that the effect of the change in process solution flowrate on the overall stage efficiency is slightly more significant for greater weir'
\

lengths (i.e. for smaller fluidization orcas). The decrease in resin holdup is the
main couse .for the lowering of the overall stage efficiency when the weir length is
increased ~

At a weir length of 3.38 inches (8.59 cm), the change in solution flow from

\ 40.0 to 94.2 m I/sec corresponds to a chango in particle Roynolds nuih'ocr from 7.2,

:\.

' .

\ t~.
16.S. Both the solid and liquid sides are significant in determining the overall
~

. ,

~s-tramrc~.,coefficient which is increasing with the in~rcose i(lsQlution flow-rate.

d2''-:0.the'fa~oura6Ie
I ;composiJio~ changes.as shown
. .in Figures B.256 and
. B.25c.
he I uid-film resistance also' is decreasing with Ihe
ve'I'
OClt

increa~e

in superficJa!liquid

hi/,

Ti'1e conge In I'Iqui'd fl'ow-role correspon d s I0 a conge

h
.In Ih c super/'

'

II

)
/)

75

ficial liquid volocity from 0.61 to 1.44 cm/soc which would significantly increase
Ihe mixing of tho ra.sin particlos In each stago. Thoso Increasos in ovoroll massIramfer coefficient and in rosin mixing countoroct port of the offect of Iho dec;eoso
in ,olid-liquid contact as tho solution flow .... ato is increased.
Howover, at a greator weir longth, o.g. 3.97 inches (10.08 cm), tho $Omo
chango in solution flow-rato corrosponds to a chango of portlclo Roynolds numbor
from 11.7 to 26.4 and tho superficial volocity from 1.0 to 2.3 cm/sec. Tho solid,

sid" mass ti-onsfer becomos tho controlling stop, espoclolly at tho high rang~ of
liquid flow-rates.

The increaso in superficial liquid velocity no longor has a

significant effect in incroasing tho ovorall mass-transFor rato by reducing tho liquidfillll.mass-tronsFor resistanco. Tho effoct of increasing tho solid mixing by increasing
the solution flow-rato is less significant bocauso of tho high suporficial liquid
velocity already experiencod by the resin bod at tho low rango of liquid flow-roto.
Hence, the incroase in tho overall moss-transfer coefficiont and tho 'resin mixing
with increasing solution flow-roto is less significant at groater woir longths (smaller
fluiJiwtion areas),

The docrease in salid-liquid contact due to increasing solution

flow-rotc becomes m<;>re dominant For operations with smaller fluidization orca.
ft.-nce tl", solution flow-rate has a m'oro significant offect on tho'overall stage
dficicncyas the weir longth is increased. "

8.,1.10 Weir height - weir length interaction

/ ":'

Figure B .260 shows thq response of tho

overalhtog~ efficiency at th'ree

different weir ~engths as tho weir height is increased. The increase in the overall,

76

,Iau" efficiency as the woir hoight and fluidization area are increased was
discussod in Soctions 8.4.3 and 8.4.4 respoctively.
Figuro B.26a shows that tho. rospanse of tho overall stago efficioncy to the
weir height is moro significant at a smaller weir longth, 1.0. larger fluidization
area. For tho oporations with difforent weh lengths, Figures B.26b and B.26c
,I,ow Ihe soma pattorn of composition chango (In the solid and liquid phoses)'with
differellt woir lengths. Honco the composition depondonce of the ovoroll masst",nsfer coofficiont is not tho roasan for tho intoractlan botween 'the woir holght
(lnd till' woir length.

Bosidos, tho solid-sido mass transfer may be controlling at the

weir length of 3.97 inches (10.08 em) (Ro

~ 33.2), and tho chango In resin

cOll1position due to increasing weir ho'ight is Increasing tho solid-sido moss-transfer

caeff i c i el1 t
Ono roason for tho intoractian botwoon tho woir hoight and woir longth is that
~

till' ,osidence timo distribution of the resin, is more significantly increasod by tho
illcrl'oso in we ir hoight when the fluidization area is large (short weir longth).
lienee II", more significant increase In resin rosidence time causes a more significant
iIlGrc'ClSI' in overall stage efficiency.

H.~

Effects of the Initital Sodiu';'-Ion Concentration in the Process Feed Solution

all Mass- Transfor Performance
Fi,luro B .27 a shows the increase in overall stage ~efficiency as .tho initial

,odium-ion concentration is incr'cased. 'Exporiments "';oro porformed at two process

. '

"

, .

.-

solution feed-rates: 40.0 and 79.2 ml/soc. The moan rosin particle-size is 0.735

-'

'

77

""" (-30 I 35 rnu,h). Tho equilibrium dlogrorm showing the operating Iinel oro
,IHlwn in rinurus [I.27b and B.27c.

The equilibrium comtant for tho lon-

l'xchon)" 'y,tem remalm practically c'onstant when the Initial concentrations of

Ihl' ",dium Ion vary botwuon 0.02 and 0.1 N.
Tho porticlo Reynolds numbors for tho oxporlmontal run rango from 5.98 to \
I

''\

'I!

11.9, indicating that tho moss tramfer Is controlled by both the ,olld- and Ilquld-,
,id" ,e,istances. As tho Initial sadium-IJn concontratlon Is Incromed, tho IncremJ
in ,odium-ion mole fraction In both the rrin and liquid phases (as shown In

,
riguros B.27d and [1.270) tunds to inciase the solid-lido and liquid-side ma .. I,o",fer coefficient>. This is ono roo

n for the Incroase In' the degroe of approach

towmd, Ihermo
f shows Iho tho Increase in initial sodium-Ion concentration
II r L r ion-exchango but tha fraction of sodium Ion extracted

hy Ill<' I<',in from t u solution docreoses as shown in Figura [I.27g. If Ihe objective
nf tlu' operation

is

elutc ions from the rosin phase, the increa~o In eluant

concenlloli0n will si nificantly increoso Iho fraction of ion extractod from the resi
phowen ,hown in Fig

H.6

e B.27h.

es on Mass - Transfer Performanco

EXI""irn"nts on moss-I

mfer performanco were porformed on

ion-f'xch'~nul' column with cyll drieal weirs in~tallod in each troy. Tho cylin icol
Wl'i, hm

(J

weir 'height of 2 inches

which giv ....

(J

.08 cm) and a diamoterof 1.69 inches, (4 29 cm)

downcomercatching--arc

",
",

-,

that of tho tray design wi

\

"

78

3 .38-inch (8.59 em) straight weir.

The overall stago efficioncy of tho experimental runs with different rosin
ked-rat"s ore shown in Figure B.28. The overall stago efficiency decreasos with
increming resin foed-ratc as expected. Figure B.28 also Indicates that the ovorall
stage efficiency for an operation with cylindrical weir is always lower than that for
on operation with3.3B-inch (8.59 em) straight weir.

Undcr thc same operating

conditions of resin and liquid flows, the rosin holdup in both tray-designs are the
scune because the dawncamor areas and weir heights are the same. Diffcrcnce in
resin mixing may not be thc main reason for the difference in overall stage
efficiency because small rcsin particle-,ilo (-30+ 35 mesh) and high superficial
liquid velocity (1.2 em/sec) woro used. A comparison of tho two dosigm indicatos
thot the distanco btween the fluid-diodo downcomtlr (from tho tray abavo) and tho
dOWflcomPr area is much shorter for tho tray-dosign with cylindriCal woir. Honea
the le,in residence timo is smaller in the tray-design with cylindric'al woir, resulting
In

lower overall stage effiCiency.

Figure B .28 01 so shows the experimental resul ts for ion-oxchango col umn
~ --~---------

operations with 2-inch (5.08 em) high coniCal weirs. The overall stago efficioncy
for th"se runs is higher than that for the runs with straight woirs and cylindrical
wei". This is because the fluidization area and hence the rclin holdup oro
,li(Jhtly Imger for the tray-design with conical weir.
Among the various weir-shope dosigns, tho resin rcsidonco. time (dctermined
by the resin holdup and the distance/between the downcomcr of ana stago and tho
weir of the next) in each tray determines the overall stage efficiency of tho ion-

79

exchange column operation. The design of weir shapes mainly affects the
operational performance of the column in terms of the flooding and cyclic resinflow characteristics as discussed in Sections 8.2.5 ond 8.2.6.

-------.

8, 7

'

Mass-Transfer Performance Factor

The overall stage efficiency has been used in assessing the mass-transfer

performance of the ion-exchange column under various operating conditions. This

performance factor indicates the Idck of thermodyn:Jmic equi Iibrium. However, it
does not account for the volumetric throughput rate of the operation. With the
resin feed-rate approaching zero, we can have on overall stage efficiency
approaching 100%_ but with little or no output at all. The response of the overall
stage efficiency to the resin feed-rate at 3.97-inch (10.08 em) weir length in
Figure B .240 does not indi~ate that the ion-exchange operation at point A is
actually handling nearly tbree times the resin feed-rate at point B, although the
overall stage efficiency is unchanged.
In Egan's optimization study on the operation of the ion-exchange column
I

the volumetric efficie cy

FS + F

(8.7.1)

-----f--

(HETS) ~ (A)
was used as the m ss-tronsfer performance foe tol. The HETS is calculated from the
physical heigh of the',contactor and A'is the cross-sectional area of the column. In
the contlru us countercurrent ion-exchange operation using the present cantactor and

\I

the other ontactors, the sol id feed-rate is much smaller than the flow of process

",

80

solution. The vore flow-rote of the solution con be more than one hundred
I

times that of the lesin. Hence the response of the volumetric efficiency to

change in resin throughput is extremely sluggish when compared to that for a change
in solution flow. This Jactor accounts for the total volumetric throughput but
neglects the chemicol compositions of the saJution ond resin feed-streoms.
Hence a m'odified volumetric mass-transfer performance factor can be proposed

(E )
vm

(8.7.2)

As discyssed in Section 3.5, only the octive height of the column should be
considered in colculating the HETS. Hence the (HETS)f in Equation 8.7.2 is
calculated from the total of the weir heights or more accurately fluidized-bed,
heights of the-multistage column. By the same argument, ,the fluidization area
(A)f' rather than the cross-sectional area, A, of the calurrm is used in Equation
8.7 .2. Simi lor to the volumetric effic iency, the modified volumetric efficiency
may be interpreted as the moss-transfer volume of the ion-exchange column required
to handle the ion flows for the ,necessary separation. For any separatior), a large
:,value of the volumetric efficiency i~ desirable. For a given column (fixed weir
lengths and weir heights) and handling feed streams of constant compositions

(comtont qo and 'c o), the m~ified valumetr,i~ efficiency (Ey)m is equivalent to the
"..

j/

original volumetric efficiency: E , with the response to resin flow magnified.

v

8.8

Comparison of the Performance of Different lon-Exchange Contactors

Slater (2) has used the sp~cific ,volume in comparing 'existing ion-exchange

81

contactors on the basis of their chemical performance and total volumetric throughput capacity. For stagewise processes,

~ specific volume is given by

HETS

v---

(8.8.1)

u+v
where u + v is given by

FS + FL

u+v

=------=---=-----

(8.8.2)

X-sec tional area of column

The specific volume, for which a small value is desirable, may be interpreted as
the volume of a theoretical stage for unit total volumetric throughput. This
performance index is actually the reciprocal of the volumetric efficiency in
Equation 8.7.1.
The specific volume index also serves to compare the performance between
fixed bed and continuous countercurrent-contactors without solving for the
concentration profile in the fixed bed.

For sodium;hydrogen ion e,\change,

Moisan's data (52) for fixed bed and Turner's (29) for fluidized bed are shown in
Figure B.29. Data from the present study are also included. The

e~sent ((IX
,~

optrations appear to be much more efficient than a similar column developed by

Turner (29). The reoson is that the presentcontactor design using fhlid-diode downcomers makes stable column operation possible at high resin and process solution
throuohput rates.

Our experimental data have much higher (u + v) values. The

HETS values calculated from the active height of the <oolumn are essentially the

....../

"orne because both ore mul tistage fluidized-bed operations having the some

. system,. solution concentration, and resin'particle-size distributi'on.

chemic~1

"

82

",

_ If it may be assumed that the comparison 'Of specific volumes of different

,

. '

I .

contactors for a given total flow velocity and chemical system will hold for other
-~

chemical systems, then all conloctors may be compared wi!h each other and with a
fixed bed of specific val~me, Vf This has been done by Slater (2). The standing
of the multistage fluidized-bed io!'-exchange conloctor with fluid-diode down" comers among other existing ion-exchange contactors is shown below:
Relative 'SPecific
Volume (VrM

Conloctor

1.0

, Fixed bed

0.6

Fluidized moving bed (Selke)

0.6 to 0.3

Pulsed sloges (Grimmett)

Moving bed (force) (Higgins)

0.5
0.5

. Moving bed (gravity) (Se,Ike)

Present study

0.36 to 0.13

Jigged bed (McNeill)

0.25

Fluidized stage (Turner)

0.06

Coil (Schulman)

0.06

The relative specific volume' for the present operations ranges fJexit 0.36 to 0.13
with a mean at 0.18~ Before the calculation of the relative specific volume, the
specific volume dolo from the present study have ~en approximately adjusted to
account for the differenc;:e in solution concentrations be~een Maison's data and our
data. The ~erforrilan~e of the pr~nt contactor is much better than that of
Tu~ner's fluidized-stcige canloctor. The performance is in the same ';pnge as the

\.

"-,

,~

,,,

83
~

oth~r commercially successful moving-bed contactors'.

Slater (2) bel ieves thot the aboye comparison is very opproximate 000 relates
to capital cost principally. The lower the relative specific volume the lower must
,
,
be the total contactor cost to be competitive with a fixed bed. The specific
volume of a Fixed bed concerns only the height of one transfer unit. The actual
volume of a Fixed-bed contactor is in fact several tim~s greater than that of a'
continuous countercurrent contactor with the some duty because most of the resin
inventory in a fixed bed is not active. However, the relative specific volume
would serve well in the comparison among other (CIX processes.
The specific volume of a fluidized-bed contactor would likely be always
larger than that for moving fixed-bed contactors due to the larger bed voidage.
Howe~er, the simplicity in contactor design, the large solid and liquid handling

capacity, and the excellent op~rotional stability of the present de~lgn may
compensate for the slight loss in chemical eFficiency., The required control on the
\

resin and process solution flow is much simpler than that for the moving fixeci-bed

roperations. The" continuous use of the resin for obout two btIndred experiml(lntal
, , runs
has shown that r/i'sin attrition in this continuous countercurrent operation is n~gligible
because no mechanic,al parts are im;olved i~ the resin transport' and' the entire
operation involves only moderate pressure changes.

'"

\
, 1

\I

.'

9. CONCLUSION
-----.,...,,--

--

,",
_0

The multistage fluidized-ged-CCIX column ~ith fluiq,-diode downcomers from'

'

the work of Cha wat'modified and improved. The entire ion-exchange column
system and the regeneration and washin9~YStems were redesigned and rebuilt: The
system of resin feed by volume,trlc displacement proved to be con"trollable and
produced no noticeable resin attrition. The modified ion-e,(change column had
,improved liquid distribution. The effluent resin no longer backed up at the exit.
<

.. \>e

The resin sampling system proved to

vert efficient
(

The study has shown that the implementatio,n of fluid diodes in the design of

,/

'

,downcomers is successful i'n ';'ini,nllzing liquid':'bypassing up the downcomers

-

ensuring a smooth and continV6us flow of resin from stage ta stage down the column.
/' /.J

Easy start-up and but-doWn"'were observed and the resin was evenly distributed on'

\..../.

each mass-transfer stag<e. Fluid-dioCle downcomers are pressure sensitive devices,

"'.

the performance of which can be prediCbted by a pressure-clrop consideration'. The

more pressure sensitive is a fluid-diode design, the mare unstable ';;i11 be the column

operatio~owardS any sudden change in resinia,nd proce~ solut;on flows.

The tray str~tures and resin particle-size distribution can significantly affec't
the contactor:~s operational performance in terms of resin fluidization in the
fluidization area and liquid-bypassing In the.Jowncomers. The increase In fl."id-'
diode lateral displacement increases the possibility of liquid-bypassing. A large
resirl parti~le-size distribution increases the pressure drop across the fluid-diode

....

84
r

85

downccrners resul ting in signiFicant liquid-bypassing. The increase in weir height

improves liquid distributit'"the fluidized bed and ht:mce reduces channeling.
The conical and cylindrica
area in the dawncomer

ir design hos shown thot the elimination of dead

a~~s

the eFfect of minimizing liquid-bypassing up the

downccrner because the equilibrium.'esin-heod is more effeCtively maintained in the

downccrner area. Hence' the operation of the column at very loW resin but
moderately high liquid flow-1"ates b6comes'feosible. The interaction of ihe conical
.

and cylindrical weirs' wall with the inner 'wall of the ca'iumn reduces resin mobility
in the fluidization area. Hence the ~traight weir with sloping inside walls would
be the best design for the present solid-liquid contaetor.
The l/4-t-eplicate of a factorial experimental design to 7 factors at Z lev~1s
has flfficienlly and successFully revealed the operational and structural parameters
as well oslhe interactions among these parameters which are significant in
determining the overall stage efficiency of the present ion-exchange operation. The
experimental design showed thot the resin feed-rate, process solution. flow-rate,
weir length and weir height are signifjc~nt main factors. The interactions between
resin Feed-rate and weir length, process solution Aow-nlte and weir length, and

weir height and weir length are significant interactions. The resin particle size,
f1uid-diode lateral displacement, and downcomer clearance (or seal) are not
significant in affecting the overall stage efficiency between the levels Investigated

Further mass-transfer s~y has shown that resin residence time is the main
factor in determining the overall Slage efficiency of an operation. The dependency
of the solid- and liquid=-Side mau-transfer coefficients on the chemical compositions

6 .

..
86

or" the,solid and I iquidpha'ses also ploys a role In aff~ting the oyerall stage
/.

efficiency.
-

Experimental results from the present stUdy have/demonstrated'thot

the
,.i
_.
.

multistage f1ui1zed-bed ~ClX c~tactar with f1uld~~ ,downcomers has a

.'

favourable chemica efficiency andvolumetrlc.. capaclty ~ht~ compared to other

ion-eKchange c
bed operation

operotion~:

ctars on the 1;C;sls of relative specific volume. -The ftuldizeds a slightly lower chemical emcle~y than the moving flxed:bed

However, the

slmpl1c1t~ln contactor d~sign and In control

require-

monts together with the operatlanal stability, and larg!, solid arftl liquid handling
capacities, may compenstate far. the slightly lawer chemical efficiency In large
industrial applicatIons.

'.

.~.

.
NOTATIONS

, '

complex mopping v a r i a b l e .

cross-s(:ctional ore a of ion-exchange column; openIng

diode downcomer (cm 2)
Af

fluidizati-on area (c.m 2)

concentration (meg/ml)

ar"<J'(r fluid,

f.

interfacial concentration (meg/ml)

C.
I

Ca

inilf>l sodium-ion concentration (meg/ml)

particle diameter (cm)

interdiffusion coefficient in resin (c,;,2jsec)

DA

- self-diffusion ~fficient of species A in' resin (cm 2jsec)

DB

self-diffusion coefficient of speciesB in resin (cm 2/sec)

DC

downcomer clearance (cm)

DLD

fluid-diode lateral displacement

DP

':J
,

'0.

(~m)

resin particle size (mm)

.'

Eos
E

(E)m
.F

. overall stage er!tciency

>

It

volumetric efficiency (Ifor)

modifi~d

volumetric efficiency' (meg/sec-cm 3 )

force (dY!1e)

fb

bouyancy force on resin (dyne)

FHE

fraelion of hydrogen ion extracted

,.
<.

87

.'

Fs. \

fraction of sodium ion odsorbe~

FL

process solution flow-rate (ml/sec)

FS

resin ft)eci-rate (ml/sec)

>1

acceleration due ta gravity (cm/sec 2)

mTS

h.:-ight,equivalent to a theoretical stage (feet, em)

,UIlTS)f

height equivalent to a theoretical stage based On active

contactor heigh't (feet, em)

I'L

liquid-side mass-transfer coefficient (em/sec)

kLb

liquid-side moss-transfer coefficient at bubble point (em/sec)

mass-transfer rate canst~nt (cm 3/meq-sec)

process solution flow-rate (ml/sec) ,

number of stoges

NTS

number of theoretical stages

PI' 1'2

pr,mure (dyne/cm 2 )

resin capacity (meq/ml)

I:

resin flow-rote (inl/sec)

RI ,

R2

I:IOX

I,

,-

resistances of fluid-diode do.wncomer, to flows (dyne)

rata of ion exchanged (meq/sec')
time (sec)

resin flow velocity in terms ~f fluidization area (ft/hr) ,

superficial velocity (em/sec)

velocity of process solution flow in terms of fluidization area

'(FI/hr)
,
\

"

89
V

sp"cific volume (hr)

-/j

specific volume of

s~p~r!icjal velocity {cm/sec}

"

fixed'bed (hr)

\' ()~

superficial velocity,at bubble point

VIII

weir height (em)

\'/!.

weir length (cm)

", Xl'., XT

mole fraction,of sodium-ion in liquid phose,

stream IT'" top outl e t stream

y, YB, YT

B~ bottom

inl<>f

mole fraction of sodium idn if! resin phase, B = bottle out~et

~ stream, T ~ top inl et stream

active'h:!ight of contactor (cm)

,Di""""ionic:ss groups
<')

Rc

Reynolds number

s-

Schmidt number
/

Yl

Sherwood number

~)t

Stanton number

ere'.,k L"tters

"-

voido9.e

t.b

voidoge at bubble point

"

I(

dyn~mic viscosi ty {gm/sec}

kinematic viscosity (cm2/sec)

density ,(gm/cm 3)

kt/k LB

..
1

REFERENCES

I.

Slot, G.S., Soc. Chern. Ind. Conf. on lon-Exchange, p. 85 (1970).

2.

Slater, M.J., Brit. Chern. Eng.,~, 41 (1969).

3.

Levendusky; J.A., Chern. Eng. Prog. Syrnp. Ser., 65,113 (1969).

4.' Cloete,F.L.D., and Street, M., Nature, 200, 1199 (1963) .

. 5.
6.

Grimmett, E.S., and Brown, B.P., Ind. Eng. Chern., 54, 24.(1962).

Eng.~unive:ity of

Sauhrada, F/,. {n:ublished work, Dept. of Chern.

NewBrunswic~ FrederictOn .(1969).
.
."_

.
7.

Egan, S.M., M.Eng. Thesis, McMaster University,. Ha'!1i1ton (1971).

8.

Cha, C.Y., unpublished work, Dept. of Chern. Eng., McMaster University,

Ham iI ton (1972).
.

9.

Wood, L.O., and Fox, H.L., Intern. Sci. Technology, 23, 44 (1963).

10.

Vario~s

II.

Various Authors,

AuthQrs, Fluid iriplification Symposium, HOL (1964).

~lllid

Amplifier State of the Art, NASA, CR 102 (1964).

12 . . Telse, N., United States Patent 1,399,559 (1.920).

13.

Ringleb, F. 0., Syrnp. on Fully Separated Flows (ASME Meeting), p. 33

(1964)

14.

Thompson, D.W., Can. J. Chern. Eng., 48, 236 (1970).

15.

Stevenson, D.G., Soc. Chern. Ind. Conf. on Ion Exchange, p. 114 (1969)

16.

Morris, W., Brit. Patent 885,503 (1958).

17.

Hiester, N.K., etal., Ind. Eng. Che~., 45,2402 (1953).

18.

Iwosyk, J.M., and Thodos, G., "Chern. Eng. Prog., 54, 69 (1958).

90

l.

~19 .

91

McCormack, R.H., and Howard, J.F.,.Chem. Eng. Prog. 49,404 (1953).

.

20.

Mue7~H.,

J'

' .

and Selke,

~.A:'I~.

--

Eng. Chern." 47,374 (19:5).

21. ' Miles, J.H., Brit. Potent 1,054,578 (1968).

22.

Schulman, H.L, et 01., Ind. Eng. Chern. Proc.Des. Dev., 5, 257 (1966).

23.

Weiss, D.E., McNeill, R., and Swinton, E.A., J. orMetals,

24.

Peebles, F.N., Chern. Eng. News, 34, 5200 (1956).

25.

Higgins, I.R., Ind. Eng. Chern., 53, 635 (1961).

26.

Higgins, loR., and Chopra, R.C., Soc. Chern. Ind. Coof. o~ Ion Exchange,
p. 121 (1969).

27.

Gilw~,

955 (1955).

'.

M,.E., Chern. Eng., 74,. p. 83, Dec. 18 (1967).

28: Slater, M.J., Soc. of Chern. Ind. Conf. on,lon Exchange,

29.

?'

P.

127 (1969).

Turner, J.C.R., and Church, M.R., Trans. Instn. Chern. Engrs., 41, 283
.(1963).
)

30.

Stermerding, S.,et 01., Soc. Chern. Ind. Symp. on Fluidizotion (1964).

31.

Cox, M., Trans. Instn. Chern. Engrs., 36, 29 (1958).

32.

Helferrich, F., 101) Exchange, McGraw-Hili Book Co., Il)c. ,(1962) ..

33.

Whitcom~,

34.

'--

J.A., Banchero,J.T., and White, R.R., Chern. Eng. Prog.

Symp. Ser. iSO, 73 (1954).
.
/
.

Turner, J.C.R., Church, M.R., JohMon, A.(W., and Snowdol), C.B.,

Chern. Eng. Sci., 21, 317 (1966).

.'

35.

Turner, J.C.R., and Snowdon, C.B., Chern. Eng. Sci., 23, 221 (1968).

36.

Duncan, J. F ., and Lister, ~A. J., J. Chern; Soc., 3285 (1949).

37.

Gilliland, E.R., and Baddour, R.F., Ind. Eng. Chern., 45, 330 (1953).

38.

DovielC Ion Exchange, Do\/{ Chemical Co.,' Midland, Michigan (1958).

39.

Snowdon, C.B., and Turner, J.C.R., Proc:. Int. Symp. on Fluidization,

Ei ndhoven, p. 599 (1967).

--

.92

40.

Petrovic, J.J., and Thodas, G., Pr9C. Int. Symp. on Fluidization,

Ei ndhoven, p. 586 (1967).

,
41.

Richardson, J.F., and Zoki, W.N., Trans. Instn. Chern. Engn., 32, 35
(1954).
-

42.

Beek, W.J., Proc. Int. Symp. on Fluidi:tation, Eindhoven, p. 507 Q967).

43.

Moyle, F.J., Australian J. of App. Sci., 10, 433 (1954).

44.

Turner, J.C.R., and Church, M.R., Trans. Inst. Chern. Engrs., 41, 283
(1963).'
-

45.

Foust, A.S., et 01., Principal of Unit Operations, J. Wiley ond Sons,

New York (1960).
.;
.

46.

Rowe, P.N., and Claxton, K:T., Trans. Inst. Chern. Eng., 43,321 (1965) .

. ' 47 .. : Peng, K.C., The Design and Analysis of Scientific Experiments, Addison' , ' Wesley Publishing Co., Massachusetts (1967).
..
,
48.

Johnson, N.L., and Leone, F.C., Statistics and Experimental Design: In

Enginee~ing and the Physical Sciences, Val. 2, John Wiley and Sons, Inc., '""'New York (1964).
. . . .

49.

Daniel,

50.

Leva, .M., Fluidization, McGraw-Hili Book Co., Inc., New York, p. 35

(I 959} .

51.

Hanna, M.R., M.Sc. Thesis, University of New Brunswick, F~dericton Q970).

52.

Moison, R., and O'Hern, H., Chern. Eng. Prag. Symp. Ser., 55,71 (1959)

c.,

Technometric~,

!! 311

(1959).

93

.\:I'

.,

APPENDIX A

Equipment
, Design and Apparatus
,

..

~..,

. 94

"0'

FIGURE A.I

Theoretical contour of cusp-cavity shope

'.

FI GURE />. '!. Modified contour of cU$p-covity shope

9S

~I

-,

FIGURE A.3 Flow pottern -.low pressure-drop direction

FIGURE AA Flow pattern - high pressure~op direction

SHAPEA~
.

SHAP,EB

.'

SHAPE C

......

:.

1f _
Vl,

r-

OW.

.
,

.--11',.C

. I,

-~'Vi

"'-',....

- F-'

t ........

lORAIN .

,.~.
''''~:
R1 'iJ
~. I:.'::::?,'

.T~

.V5

va;' .V4. .

l@ .

FIGURE A.6

DRAIN

J
lorl:e~honge

..

I_

IV '--;'

~
I "-Y

~:=:
---

column system

lr' S 0
"18 ,

V19

'')I

(,

:i:'~'~i
.;~.

-17
j

V7/

TO RESIN_
REGENERA-

~~~M?7
OW
. .
/

R2' ~

~--

~------i"~"'a--.----J
'12

. .'.

1'
.~:;:::

y17

. .

'.-C,:

A~~

;::-

-..i..

'(n~:

..,...1.,.,

va'J tt

--I

::'.::;.~:.

. tb

-_._ T2. .,._-- =n--:

'

- -

V13

..~'
,
~e!
,~:;"
yI ....
~
.~.
.
cc
....
:
.
;
....
DRAIN
.:::; :';'.

' V27

V12

V6

- -- --.

-=..T;3-:.
-~

,.

V14
V15

..'
V16

.
V
9,

~ t

~ --I.."

;~:;\~

(~':.pg

V10

[o:A1 N
~

6"Diam,-.,
LIQUID OUTLET....
1"
44 Diam.

1-;:'

r-------l ~
I

97

65/8

1c'~.. ' ~

7t"Diamt~iI!;~~E;f",~,:].41.8-1, y."
v

'

..

11

II "

II II

3/~ t~,.!,~~!~;~=d=-U~~~f-t
II;,.
II

:.

:1 I:
;,-: :."

"
~-BORE

HOLE'E'E~*=u-=Il=!~'~:'~i:p3

DRAIN .
1!4~srnE
AIRVENT

./'

. ",

FLUID-DIODE
DOWNCOMER

:::;

'I II
II II

d~2" ~

u_u

.' ...::I.

c:
E

~, "
~ I'

II

:>

"'0

"

..2'
u

.1

,I ,I
Il_JI

.2u

.l:!

><

f
c:

'y

WEIR

..
0

-.
III

1~'BOLT~

:;:

"5:

,~

- I:'. "II

3/16" TYGON

TUBING -----..\

:1 ::

II II
,. .\4

"<
....

.Q

..:

:::l

DISTRIBUTOR 'r: _
'
PLATE --tto""'_____ _

UL

I! I
"

til-------

LIOUID INLET - - - ' .-.Rlj+-l'LOi~_-:..-~-l_r-P~,

u.

6"

~ i::-:I ''?~'~..'\ / "

4"' ~9
~
!-1 l-'-~H/r/-'- ~ , Scale 1:4-5
_ _.!R~E~S~IN~OU~TL~ET~=,~.h..=!~t,=:::::.....-t.~J____=-=-=."'-,=_.:...:_.:
.._:.:,._:..
SAMPLER

,=."

'.'L

__

.=~

98

/i'

71
1

It: -,

..

'4~~
II

, I

'

"II

L
T

3/a"

I
I
I

--.I

Is

5"

II

..

1/8'
O-RING

1/4'"
I

.o.: .... ~~'"' .. - ....

r'

II

11~"

118'
.'

"
',.:

Scale 1: 5/3

FIGURE A. &( ~clions 7 and 8 of ion-:exchange column

I <,:"_


99

",

Scala 1:1 .

FIGURE A.9

liquid distri.butor pi_ate

~-

-~

15-

.Q-RUG

6"

'/'2 .
1'"

I
I

Scale 1:8/5

.,,

.1
1
I

FIGURE A.lO Resin sampl~

100
\

L- _ _ _ _ _ - _ ,

Scale 1:12
FIGURE A.II

Resin sampler's plunger

1 - - - - - - 6''-.:....-----1

I:

: lIa"

....II,

1
3 //"
4

.
rr.,------- - - -i~...........--_
t. _".r--

II

~ ,'.

I~

I,

31a' ~ : . :1 ' ':

~ ~'l"--r.!I- - - - - - - -- - - I I
I

,'
!W

t --'-

"

l/S'

I,..

. 1 " :I

.l

2"

I
I

I
I

-r,

I
I

3~'" 1r-~,~!-ij:~,~::-------T!,T!,--~kr-~,nj[J

T
FIGURE A.12

71/2'r-4114'
Sec:tion I. of .

Scale 1:2

------------------,.--------,101

.,

TOP VIEW

.'

I
I

J
J
J
I
I
I

Vs"1j
%2'

FRONT VI8N
~

3132"

111.32 .

3116"

.1

4"
43;8'

J
.

18

( ':}'S')S': :VS')

~j

\1"
~

. , ,%2"
f '

f.i1;, 6

1f.32

II

Va"

SIDE VIEW

118'

20&~

mes
,screens
_

(I-;UREA.13

.,

. ._ ... rIo............

.~

II

Scale 1:1

Tray design with 4" straight weir

-

( .
~

~-


102

iI -cH'I

TOP VIEW

%2"
Lt====V:=~='~==~'11~11~='~'~====~C=~
%2"
I-----~

.,~ 31(6"
33Q."-------l

'-I- - - - - 311;1s"-----'----I-1
I

I
I
I

I
I
'I

I
I
I

"'4',

3}32

118"

SIDE VIEW

J
Scale 1: 1

FIGlJRE A.14 ' Troy d~sign with 3 3/4" straight weir

II

r-~--------------------------

____________

.r---2~6"Rad.

lIs"

TOP VIEW

~6"

~' i 11!a2"
3~"

,.

~,

r
I .'

C,

CENTER

a31s'
3Y4'

:11

FRONT VIEWT

. , '7/'
H's

(,3/8~ 1'a':~i

\f:' ,~'
T'

1!3~" '116"
118"

.'

SIDE VIEW
20&60
mesli

<:...screen~

... ....... - ....

"..,..

~ FI-:;URE A.IS

W'

~Jo"

-I I- ....~16
,'-.I'

Tray design with 3 1j4" straight weir

Scale 1:1 ~

"

I03

.-"-\

104

II

43;8 4"

TOP VIEW

Scale 1:1

T
2"

(also 11,"
v2 1"1'"
'2)

20&60
mesh
screens

~1~~~J~~~6~~~~~~~~~~~~~~~
,-.
~-----"'-. ---3-;lij
"

. f

"

~~u__

SIDE VIEW
FIGURE A.16

.---=~---- ~S'8':

Tray design with conical weir

'.'

1Itf'

----------------~'"~--------~---------, I~

.. 3

II

4 Va 4

..

TOP VIEW

\
\

"/

J;4

I
I

~1!8 i/16

"

T== -=-==-----~-------~

SIDE VIEW
..
/
/

,/

,___ T\-___ .l.,-__ ,1-L3I32

.

It

":-'1It6~
1:-13116'
'
. !...
,9116"

FIGURE A.17

(also' 2,', V2) "

\
\

.y" .. , ..

2"

"

...:

.----- - - - p1

3116'

screens

-r

M'

Scale
/

Troy design with cylindrical weir

20&60

m esh

106

T
"
5Va

"J!!!i=C;::j::J:J:!l,L

~ I
'~-~-,:-L

L~~c=~'t,t'I.J~'

I=- 1~"--I

TOP VIEW

.,

, , -.... '

2'/16

23/4'

,'~
,~

:' -- ...,

"

\
\

.~

-1\

-I

"etb.

5116"

'--t-', -.,.
,

..
17/32

....

........;;

':':'~

'1){
I

-1 191.32'j.SIDE VIEW

FRONT VIEW

\
!': 'JIJRE A.IS

------

114

............
_,

II 215~6"

-----

I
3/32"
I

3116"

- l\

~
--..... \
,~\

3~6

----..., ........
...,..
( :.-J..a

"

I,

I
I

I \

I'

, -............. II

~,

1/4

T/'pical fluid-diodd downcomer

Scale': 1

'

,,

..

,
h ..

raz

19/3i'

'I' I ' ,-,-

-~" ,

I,

5/S'

, I

-L

'

'1-4p

: I
I I
I '

,-

I'

107

5116"

!-=- 1~~
~ .. ~ -

TOP VIEW

3116" ..
~
t-~
, 5/,&"- 1-:
6"
--l l- fto'

'/2'L ~

~\
,,... -- ",.

11,'4'

,I ~

,,

I'

21',6"

't"

23k."

~I'
t.

1'" "1. . .,

., -.,,

~
I

.1, :
' (

,;

I '.
I

'"

~,.

"
~32,

\i

FRON T VIEW

, - ..... I

3116"

~-

I~ ;,I,
. :, l ~
,'1
," , l.. .

'--.

I' '\

,
I

113/e;i
BOTTOM
VIEW

--~----

13/:32..

I I

,;.I

I'

1I8"-i

: i '':;LJRE A .17

--,- - - -

............. '....

'.~

11J~6;;

1-112~

'

'. SIDEVIEW

-- -'- .1 ____t.. ~

3'16':

713i~3i'

Fluid-diod~ downcomer In section 6' d ion-exchange ~olUl'Dn

.,

108

C2

... : ......

.~ .. ~ .., ...
:: ,'. " '.:'

\': : :.='

"

.
"

I "

..
..

.. '
.. .::
" , ..
\

.'

. " ,-

............
.

"

.::.: ..'~~:::,

.. ,' ..

.. :.,'.'
.....

. 0"

L,

FIGURE A. 20

Resin drying system

I: I :I--'.r:....V>c~~8_...
LJ

........ AIR

AIR
SUPPLY .
FILTER,/'_

,-1--

V24

<.

TOB2'

.,

- C3"7

V12

",," ,0,: "

.:"~'.: .. ~.

f":

:: ~0.:0;:

O
.....
Z

T3

0 , "

'-

",

V27
WW.

"II " ". ""

~ .:~0.0:0:;
:
r \.,:0 .

. \flO . 0 : ; 00: ;0-:1

- ... . _"

--

-..... -

T3-

-,.\

0" ,. .;.
FIGURE A.:i( ,Resin regeneration and washing system

'.r'

- 66-:-

;....u_~-..II-I~*Q""'jV22

'"

"

,0,.:.' ...
" .:,~0,-

..:

:65:

..

..

0 ;'0.:-::
...

.~::.:

....,_-1

~
z

.- -."" :"0"":"
... """.,,,":,
. .." " " .
" -"
".,,"

TOT3

.'

"

.,

,,-,1

Ci
'C

110

"

APPENDIX B

Experimental Resulh

..

_ _ _ _ _ . . ~.

_.",...-

.j

_.

__

t.i

. 125

lOO

,
\~ .......
\.....

III

II-

~--.

"-

0'25N

0'5ON .

WON

..--

050

.....

\ .....

OD

___________" _____2~OO~N~.-

4'00N
.~.----------~.~----------.--

100

20

30-0

TIMECMINJ.

FIGURE B.I

Change in the pH alu~ of mixtures of resi

sodium ~hlaride solution with time. 3 milk
volume) samples of H+ farm resin wol ~dd d to
100 ml NaCI solution of various concen tions.

"

112

,
J

03

z
z

0::

ZO2

w
o
Z
o
o
Z

t-4

--l0'
-.
:::J

4-0
FIGURE B.2

Equilibrium hydrogen-Ion concentration verlUs Initial

sodium-ion concentration in solution

..

113

'1 ,.0
10-0
-~---o-25N

g-O

H10N

----3{)QN

ao
7-0
I
I
I

l.!.J

::J6-0
-l

<t
>

Is-a
a.
I

I
I

4-0

I .

3-0

20

_.-- _.... -8'-:---

--0--- - -

FIGURE B.3

---

" "

....

....

/~
.,,.

../-

.". . .." .

."".

Titration curv.s. S mlsomplos of H -form_resin

.in 100 ml Noel solutions of various conc.ntratio~

114
(

no
100

40N

-~---5DN

go
so
70

w
360

. I
I

<t
>

I
I
I

0.50

40

30

20

1-0
35'

40'.

45'

50-

55
60
O1N NaOH SOLN ADlEO{ml)
~

FIGURE 8.4

..

TItration curves. 3 ml samples of H+-form resln in

100 ml Noel solutionS of.wrious concentl'c!tiom
\ ,.

65

..:;J

61 'I--r--r---r--r--r--r--r-r---r---r~-'--r-r---r-"'--r---T----"'-

...... 60
>

o 80

- l59.1

NOj' present

-....

4 present

....Z

~ 58'
0
a..

8 57.

.'

,.

,i

o .
o

~56'
w

...
_.-.

o
,"

"

5&;&
.m

.mI

02

FIGURE 8.5

0304

05

06

-eJl

08

H+ ION CONCENTRATION (meq/ml)

09

10

Measured electrode potentials versus hydrogen-Ion concentrations

In solution

iii
.'

..

)-.

62-.--r----r-....:.-,-.-,-.....--r----r;--,------r~Jj'
~

>

--<t

'--

......

--- .

6 ()

-0-

01-

wZ
o:::w
~ I-

~
0

~~ 8

--\.

~~

0504 present

~ ~56'
,,~,

/0

54"

",b
FIGURE 8.6

.NOj

2''0

3!0
4:0
pH VALUE OF NaCl 50~N

present

~O

&0

'

Measured electrode potentials versus pH values of solution

...

0-

.'.--:-~.-.'"':~

117

140,

120uCJ

~100'
E

60'

"

W
I-

~ 80-

sa.=:
( I)
.,

!II

lL.

a: 60,
w

00-

::J

30'

'40-

20

"

40~-

...J

20'

./
20-

30

10,

4(}

50'; 60'"

70'

ROTAMETER SETTING (top of tLoat)(unit)

FIGURE B.7

Calibration curve of the rotameter For process

solution Feed

,'

80

00

9(} ,

7-0

. //0

60

......
u

./

~ 50

"E
....w
-'

'

4;.Q

. 0/0/0

~3'0
.....I

<t
-0
tt-

2.0

',0
00 -1-0

.
2{)
30
4-0
50
&0
7'0
8'0
gO
RESIN FEED ROTAMETER SETTING <TOP OF FLOAT)(unit)
FIGURE B.B

CalibratIon curve of the rotameter for resin feed

10'()
CD


200

150
u

CI
Ul

/-

/-

100

50

00 ~3.0~~4'()~~5{)~..-'.-,.J.6{)---'-="7.0~-a~{)~-9{):"'=-...L-.1~(}{)::-I
PUMP CONTROLLER SETTING (Unit)
FIGURE 8.9

Col ibrotion curve for Masterflex tubing pump's speedcontroller

.

119

/e/

'.

3,0

e-35+50mesh resin

-.....
~

E
w

.~/:/8/

~2'O

.~~::;;/

9I.L.

';:)

....I

~H)

~/O

o-30+35mash resin
,
Q-20+30 mes h resin

W
~

e/e//

:/8~

'.

~8/

~ty
2,0
3'0 .
40 5,0' .. 6'0
70
8,0
9-{)
RESIN FEED ROTAMETER READING (TOP CK FLOAT)(unit)
FIGURE 8.10

10'0

Calibration curves for resin feed

'"o

65 _
'-35+50 mesh
0-30+35 mesh

-60

() z .

~-55

250

.-

t)

<i:-45

e:w

.-'

Q-2 Ot30 mesh

(,
.~

-----0---0____0~ -0_ -0- ;Q

0 ____ 0_

.~.~
'.

-0_

_0_

_0

80

9'0

'.

Q---Q-Q-

~'40

3.
.

10

~
35
30~

20

30

4'0

5-0

6-0--- 70

. RESIN FEED ROTAMETER READ! N6( TOP OF FLOAn

'.

FIGURE B.II

Volume f"lellon of resin in feed versus resin-feed rotameter's

reading

10{)

;:,;1

<

- .. . ":.n
.'-l<

;r

g
Vl

......

e-35+50mesh resin
o-3(}t35 mesh resin

Q-20+30mesh resin

E3:01....,
0

I.JJ

lJ..

u;
I.JJ

00::

~20'
w
';i

0:::

-1

'

..... '"

lJ..

~ '0'
~

/_//

'I

~,
0 .0 -.

"'-

--" '
.- "-

-'-

"-j

,-

,-

ROTAMETER READING <TOP OF FLOATl(unit)

FIGURE B.12

Calibration curves for water flow In resin feed'

I..

'"

'
'"'"
.':.';i:.:A.4

123
~

'E 0-'

. &D-09
EO-OS

0
i .

zO07

,....
,
..

..

a
...... 006

~
'g:O05

zw

0.:

~O-04

".

5-03

0-02
0-0 1.540~''~.~50~;'~'~-:60-'-.-'-'---:70~..-'"'----8~O-.--'--ro-~-J

ElECTRODE POTENTIAUm\l)
FIGURE B.13

---

Calibrotioncurv~ for chloride-ion cOlJcentration

-meosuremenls

. ..

____--~.------~------.-e-

.~.-

'

\
/.

L;

-'

GO-

0'0

10 .

20' .

. 00.!to- ' . 5(}.

t-Ct ACID<equiv. wtJ

6Q-

7(}.

00-

.' I

7(}

,.

SO-,

./
/

t,

/6'1

.
--.--.--.--.----------~----.
.--,:----.
-.......---.-._e- -.

5'

--y.-

i .

::>

,,-1

, <ta
>
-~
:r:
0-

1-t:-

4'

2'

i'

~.

l'

!I.

, o-~

, -..

""
1~

'

2.b (ab

FIGURE B.15

4!0

5~
&0
tJIME(hour)

7b

8:0

g!O - ( 1d:o ( 1{'O

Change in pH value eX liquid effluent with time during resin washing

!,.

"

:~~-:M

40

...

FS =15 ml!~ec
. FL =740 mllsec
STR.wL. =3'38 "
WH =15"
DC = ,. 0"
OLD = 1/20
DP =-20+30 mesh
Co =0'047N

resin entered stage 1

351-

It

30 I!-

,..

.
\,

3
25
>
:J:
c.'

20

...)

1\

~ ~resinentered stage 2
\

..

-'-stage 3
'.--stage4
stageS -~.- .::.-stage 6

151-

1'0 00

5'0

.FIGUREB.16

100

.---

1&0

TIME (minute)

20'0

25,0

Change in the pH value of. liquid effluent from the ion-exchange

calumn as p-.(unctlon of 'time
.'

30{)
~

')

'"VI
.. : ;/>?l!

I
WEIR LENGTH =338

126
II

WH::t'
OLD::1/20"
~

FS

,--"-':="--I-:~:.E:!!1_op_.

0'6
=53-120'
ml!sec - ----~

3/20" OLD=1/20"

OLD=3/2(}'

<
n.
~~~E:I!'~I!:
5~.-120 FL=24'-120'

34':-120- '

UL-U-

-.- - ---

8stagn.'

<
~!L3:P'"lo~._ <
FL=S3'-106'

225

I~(;'r.t---

-- -

-- -

>10Go.flu

-20+30
mesh
resin

-- -

--------

--

> 10 &o.f.lui

OG-o.fluid >106o.flu

'

"

WEIR LEN3TH=3'97"
WH=l"
OLO=1/20"
~18'

1-'-=--1

stagn.

OLO=3/20"

.::18

E-__ _
~~~~.=~ 3~'-=-_9~' __
~26:.p..:'!l~b.:

~ 2~'

06
ml/Sec >94 O.llUlo.l

WH=2'"

1/20"

DLD=3/id'
~
stagn.
~2~.:.P.
rt)ob.
5_2~' .e.'!l~
24-73
FL=24'':'' G3'
---.--> 73CYC.,

2.25 c:. ~:J~mob._

ml/sec FL=lB.... 94'

------- -----

>94 o.fluid. >94 '

,;.:O'.E il.l

4'4'- 83

----- -----

>83, o.fluld. > 83 o.fluid.

Operational behaviour of the ion-exchange column'at different liquid flawrates (-20+30) mesh resin used). In eoch , tray letup, 'the range of liquid
flaw-rate which gave uperfectU operational behaviour was determined., FL
is in ml/sec, stagn. = stagnant resin bed, p. mob. =partial mobilization,
d. flood. = downcomer flooding, eye. = cyclic resin flow, o. fluid.'" overfluidization
'

r------------...------__ :'
127

WEIR LENGTH =33S"

WH=1-"

WH=2'"

OLD =1120" OLD =3120" OLO =1120"

~
stagn. ~10stagn: ~ 10stagn.

FS

E;

10stagn.

< 34' p.mob. <34' p.mob. <25CYc.

2SPmob.
1----1-- -------+---------- --------- <
----. ----(} 6
FL=34' - 73 34-- 73
FL=25-63' 25-63. -..lPEN
mllsec

>73:O:fini

of

<
mob.
-----mllsec FL=26-63

225

>(r3' o~firie
~73

o.fluid.
J

-35+50
mesh
fC?sin

.,.....

WEIR LENGTH =3'97"

WH=F'

WH=2'"

OLD =1120" DLO=~20"

~
stagn.

FS

OLO=1!20" DLD=3/20"
~5' stagn.

_<JQ~.lJ:lob.:.

I-----I-<J B:.P.:. m.Q.b.

0-6
FL=18'-73
mllsec >j3:-o~tfn-p

10- 53'
-

;.

2.25

-: _8.: ~.~o.!>:

mllsec FL=18'-53'

~J8::~.!T'~~ ~ 1~: P:!l'2P'--.--_ _ _ _ _ J.,.. _

>

, tl

FL=18'-53'
_

_ _

_ _

18-53'
_

__ -

--

.o.fine
o.fluld >530-flUI-d >.53. o flUI

c' E 4.2 Operational behaviour cl the ion,..exchange column at different liquid Flowrates (-35+50 mesh resin used). In each tray setup, tho range clliquid
f1ow-rate which gave "perfect" operatioml behaviour was determined. FL
is in ml/sec, sta:!n. =sta,9ncmt resin bed, p. mob. =partial mobllt"zatian,
d. Flood. -= downcomer flooding, cyc . = cyclic resin Flow,.a. fine =
overFluidizatioO of fines', o. fluid. =overfluldlzatlon

WHOLO
t" 3120"

FSCml/sec)

O 6

2,,1/20"

3/20"

~25 {

1"'"

1"

2"

-20+30
m(?sh

1"
0'6{

2"

22S{

2"

O~{

1"
2 II

til

22S{
0

- 35+50
m(?sh

1"

..

1120"
3120','
1120"
3/20"

etc:

2"
1"

ooS{ .

2"
1"

22S{
0

2"

20, .
80
100
LIaUID FLOW RATE (ml/sec)
. FIGURE B.17

Ranges of liquid flow-rate which gave perfect operational

behaviour in the ion-exchange column

1/20

etc.

.----..
1'0
,.., -

I In. . .

PI.

1-5

. RESIN fEED RATE(

'~-"""-"'-";'---"""--_IIiIIiiIii~lIIIiiIiIliiliiililiiiiil_iIilii

130

10 r---..,---,--,----,--.;....,..--r--~--,.---.....,.....-Fs=
84ml/sec .

FL= 79'2ml!sec

WL<STR) =3 3 8"

LJ

',23

WH= 20"
DC ;: "0"
DLD=1/20"
DP= - 30+35mesh
Co= Q'06N

"RlJu,6
185
2,08

(f).

<l
T
0...

2-42

FI GURE B.19

Effects of relin
, feed-rate on the operating line

~.'
.

f.~

r~.'
~

, '..

,. ,

I'P
;.
\1"

r-35'

~3(}

W25'

.--

131

-.

7
ill6
>-

-" .~

-.---------.

8.

C>

-.

----_
.
.
.--

"

~.----~

".-

.-.-------40-

50-

FS=1-6ml/sec
Co=Q'06N
WH=2-0"
DC=l'O"
DlD=V20
. DP=-20+30
mesh
TR-WL=3'38"

--------.---.
_e-

07

6{).

Effects of process solution f1OW-1'ate

-e"---~'---.

70- ,
80'
ec
PROCE~ SOLUTION FLOW-RATE(ml/s )
,
FIGURE 8.20

f
00-

132

FS =16mllsec

WH =20"

WUSTR.) =338"
DC=1~0"

II

DLD=1120
DP=-20+30 mesh
Co=006N
r--

l.!.J

(j)

<t

CL

Z
0-<

(j)

a:

'-"

00 ~--L._.A..._..l.--L..1":'--L.._..L--JL....---'-_-'--""""
00
x(UQ PHASE)
,0
E~ts of process solution flo~-rate on the operating line

FIGURE B.21

133

30
,....

FL =792 ml /sec
FS =16 ml/sec

~
o
"'-"

DC =1'0"
[U)=01"
DP=-20+30mesh .

l!J

CO~e

(1)2
o O

10 _e

7
5

STR.WL=33S"

_ _- e -

,~e-

_e~e
-e

'

~e~

------

e """"'"---.:
.

e~
/

e-

'd

':_~~
2

-e

7
<t 6

tf>'5
4

__ e ~

e-_----e~

-e---~
,

~~~------~------~------~~
__- - - e -

'7

f
~.6
e~"
-e~
5.~~~----~--~~~----~~
0'5

FIGURE B.22

,,0
',5
WEIR HEIGHT< inch)
",0'

Effects d ~ir height

20

.'

'-

134

,--...

Fl =792 ml/sec
FS =16 ml/sec
WL<STR.l= 338"
DC =1'0"
OLD =01"
DP =- 20+30 mesh
Co=0'06N

w
en

<[

I,

0....

Z
..-

en

a::

.
FIGURE B;23

Effecls of WE-I,. heigf1t on the op(;rati~ line

'

......

>t) 40

w
.....
t)
.....
u.. 3utti
w

-.----

-~

--.-

'"

A.

'.

4.
Q

"

'.

w1

"

..J
..J

>
0

Q-

~2O-

;i

Co=O'06N
W H=20"
, 0 C=H)"
DLD=1120"
DP=-30+35
mesh

.-3'38"
Q'""""':"3'78"
0-3'97"

~'
0

FL =79,2 ml;S;c

straitt
weir engths

50,

<

. 0. 1

"'

.,

10
','

'.

FIGURE 8.24a

1'5

20

25

RESIN FEED RATE (mllsec)

Effect. of rll.in reed-rate on the avorall .tage efficiency

at different weir length.
{
,\

','.\

~"

((

I.n

r'
I

136

'0

OPERATING CONDITIONS

same as Figure B'24'a

:2I-8
0..

...z

.7

~ 6

.0::

CD

>-

0___

~-_.

~----o

-4

_ __

-----

'3~~~~w-~~--~~~--~~~~~~~~~~

10

1'5

RESIN FEED-RATE (mt/sec)

FIGURE B.24b

Effects of resin feed-rote on the composition of resin effluent ..

YB is the mole fraction in sodium-ion basis

-<J..7
I

0..

OPE~TING CONDITIONS

se~ Figure B24a

(f)

.6

9
w

.'tO

"

STR. W _"

~_______

L- 3'97"

. 3'7;8.:-"_--.:::.o~.~-__ o_._._ _

.. ~.~~E:3~8r~'::---(,i)~_____
~

(,i)

-------------,--------4~
I

-2~~~10---~~~~-1~5--~~~~-2~O~~~~~

.
FIGURE B.i!c

RESIN FEED-RATE(ml/sec)
Effeds::If resin feed-rClle on the composition of liquid effluent.
Xl is the mole fraction 'in sodium-ion belsis

"

50'

('

. straight
wei r -lengths

". e- 338"

o
.......

3'78"
,0-3'97"

WH=2'0" .

. . DC=1'O"
DL 0= 1120"
DP=-20+30

Q-

b40'
z
w,
.......

(.)

Fs=16ml/sec

Co =Q'06N

.'

mes/:!o

---:--;-~~-- e

G: 30'

.I.L.
,W
W

--Q

2"

"

(,!)

~20'

_e

Q.~

o)------~o------~~
o __

00"

...J
...J '

--------

<t.
0: 10.
w,

~.

0."
.

ItO'

50

60'

70'

'

,,-

80'

9(PO

100' .

PROCESS SOLUTION FLOW-RATE

( ml/sec)
\..:
FIGURE B.250
~

Effecl1 d procesl solution flow-rote on the overall stage

efficiency at dlfferel'\t weir lengths
.&

.'

-~.- . . .

'-.j

-.

--:-....,.."

"

,0
~

138

OPERATING CONDI nON.S

see Figure B'25a

Ul

----

(l.

SIR.WL=33S"

-:::::::::::::::.:-=-'--=Q

z 7
.....

~-===---::--~
0' 3.97"
0'

~6

, 37 a"

0'

40

50'

FIGURE B.2Sb

60'
70'
SO
PROCESS SOL'N FLOW-RATE(mllsec)

Effects of process solution flow-rote on the cotosition of resin

,ff1".,- VB h th, m,l,

'ro,"" ;""';'; .,,;, .

"

OPERAlING CONDITIONS
see Figure B25a

w7

(f)

.:

S1R~V 3.97~8"

q...

d5
:;

~'4

0'

Q-

'

0'

r6

==--------. ' '

a -Q,
.

FIGt.:RE B.25c

1--:-

II

0',
..;...- . Q~.

"

., 40

3'
3.38

50'
,
60',
' 7,0"
, SO,
PROCESS SO\!N FI:.OW-RA1E (ll1l1sec)
Effects of process solution flow-rate on 'h~ corr,position of liquid
effluent.
is the mole froction
sodium-ion '-osls ,

Xr

in

90'

139

FL=79~2ml/sec

. FS="6 mllsec
DC="O"
DLD=O""
DP=-20+30mesh
Co=Q-06N

!!.. 25'
>-

w
(320
......
lL.
lL.

w 15'

~
I-

00

.-I

~'O'

i3
0'5

,,0

"!5

WEIR HEIGHT(inch)
FIGURE B.26a

20

Effects of weir height on the ovemll

stage efficiency at different weir
lengths

',0 r-~--~r----""""---""'-'_

GJ

009

en n.S
~

CL

d'7

FL=79'2m l/see
FS="S mllsec
DC= "0"
DLD=O""

DP-::-20+30~

OOOS

~0'5 . -0Q

m.

. >-.

:::S:0
~Q

Co=O'OSN

-.

fr4

0'3 '---'-----'-----'------'--~
0'5
,,0
',5
2'0
WEIR HEIGHT(inch)
FIGURE B.26 b . Effects of weir height on the composition
of resin effluent. YB Is the mole fmctlon
In sodium-ion basis
.

140

9
..-8
w
(/)

FL =79'2 ml/sec
FS=1'6 ml/sec

STR.
__ WL=3-97"
-.
-Q_
3'78"
-0_ 3'38"

<!. 7
::I:

, a.

::Q___

a.... 6

::5
I-

DC=l'O'"
DLD=O'1"
DP =-20+30mesh
Co=0'06N

o~~Q _______ _

x4

._

0--_ __ -Q0-

'0

0'5

WEIR HEIGH T(inch)

Effech of weir height on the composition

of I iquid effluent. Xl is the mole fraction
in sodium-ion basis

FIGURE 8.26 c

0
-4"

2'0

1'5

FS= ',6 ml/sec

WL= 3'78" (STR.)

WH=2'0"

~.

...,11)

>-C')

.....
w

DC= ,,0"
DLD=O""
DP =-30+ 35 mesh

o,rl/ /

00
.... C')

u..
u..

l!.J

~&O

.I- N

(/)

..-1

::;l.0

/0

cr:.

wN
>
0

II)

.-

'/

<

-0/
-02

04

06

08

INITIAL SODIUM CONC(N)

ION

FIGURE B.27a

'0

EffecIJof Inltlalsodlum-lon concentro'tlon

on the cwetoll staQe sffldencr. .

,
141

. FL =400 ml/sec
Fs =1'6 ml/sec

WL(STR.)= 3'78"
WH= 20"
DC= 10"
OLD =01"
DP= -3Q+35mesh

r--.

(f)

<r
I

0...

.-.

(f)

W'

a:

.O'2.0~

00 ~..J....::~::z-.~~--1._.l..---L-~_.l---l

GO

X(UQ PHASE)

10

.
FIGUR'E B.27b

Effect of initial sodium-ion concentration on the operating

line with solution flaw-rate at 40.0 ml/sec .

142

,.......

FL= 792 ml/sec

Fs =,6 ml/sec
WUSTR.):: 378"
WH= 20"
DC= ',0" .
OLD =0'"
DP =-- 30+ 35 mesh

l.!J
(f)

<l
I

0..

Z
1--4

(f)

0:::

'--'

00 '::-::----I..._-'--_L...-~_.LJ_.l.._..Lo__L_

00
FIGURE B.27c

_J__.L.._~_

X(LIQ PHASE)

,.0

Effects of in,itiol sodium-ion concentration on the operating

line with solution flow-rate at 79.2 ml/sec.

"" 143

"0
'9

W 8

CJ)

I
0..

FS =16 ml/sec

STR. WL= 3'78"

WH=2'0"
DC= 1'0"
DLD=O"'"
DP=-30+35mesh

z 6
......
CJ)
w .5
0::

m '4
.>-

'3

2
1 ~~--'-~_.J.-~--L~L-.....L-~~
02

04

06

08

01

INITIAL Net CONc(meq/ml)

FIGURE B.27d' Effects of Initial sodium-Ion concenlrotion

on the composition of resin effluent. YB
is the mole fraction in sodium basis

'7

FS=1'6mt/sec'

STR.WL.= 378"
.
_\/_"G
WH= 20"
. n~ :::..5 DC =10"
r\.~~o

DLD=O""
~o
5:'4 DP=-30+~C\...---0
.6

~.

~JS"

l.,frO ....

::l'3

,,0

~.2

-.----.-----.,-------.~

1
0 L.....l..-...l--L---L--1.--I~L-.L-~~
02

04
06
-08
0"
INITIAL Na+ CONc(meqjml)

FIGURE B.27 e

Effects of initial sodium-Ion concenlrotion

on the composition of liquid efOuents. XT
Is the mole fraction in sodium basis
'1 "

...

"

.'"

. .m6'

---'-'-~_ .... _ _. . .=

w OW ;')8;;-0-';" 'on

t' ,

144

FS=l'S ml/sec

-g 35

STR,Wb=3-7S"
WH=2- ..
DC=1-0
DLD::0'1

DP =- 30+ 35mesh

530
Q

t!J

~2'5

()

><
w

520
l.L.

w
~1'5

0:::-

.-oJ .
.-oJ

ffi1-0
2)
0'5

(}O

-02

FIGURE B.27f

-04

06

-08

. -10 .

INITIAL Nat' CONC(meq,tml)'

Effech of inltiallodium-ion concentration

on the overall rate of ion exchanged

12

145

FS =1'6 mt/sec
STR.WL= 3'78
WH=2'O"
OC=l'O"
OLD =0,1"
DP=-30+35 mesh
-e

~1-0

-e_

FL::~

~ -9

+1lI

OO_rn.l_/s_e_c_

.--- ---.

.......

~ -7
2

o 6
.....
I~ 5
0:

l1..

,4

0'0: ";..,

-02

04
06
08
10
INITIAL Na+ CONC(meq/ml)

FIGURE 8.279

. 9

b' 8

g:'7
X

w6

12

Effects of initial sodium-ion concentration on

. the fraction of input sodium-ion adsorbed by
the resin phose

FS=1'6 ml/sec
STRWL=3'78"
WH=2'0"
DC =',0"
DLD =0-'"
OP=-30+35mesh

\~~c./o-

n~'\. /

~Oj"

~~
0

'.

IA~'~

,~~....,
~\./
e

a)/'.

... o/e

~'5

24
o
~3

0:2

,e

l1..

-1 L-....I--.L.--1._.L-....J-~--1.~L:-..L...-~--'---:I

OU

02

04
,06
08
10
. INITIAL Na+CONC(meq/sec)

12

Effects of ini tial sodium-ion concentration on

the fraction of input hydrogen-ion extracted
by the liquid phose
.

--"

60'ir~--~-r~--~~~--~~~--~-r~--~-r~--~~

FL =79'2 ml/s(?c

~
o

O---CONICAUW D=1'26")
. -STRAIGHT< WL=3'36")
Q--CYLINDRICAUWD=169")

?)5(}

zw

.....
u

~ 4(}
lJ...

i----- --------e._
....... .......

t!)

~3o-

----0

I \

mesh

.- - - - - - - .

- ----
_
--- ----Q--.-.-.-. -~---------e- -------0 ------- ___!
-.-._--._.-.

-- .

=i

ffi2(}

-- -"""----

E5

...

-.- ----

."<"

10'
.

Co=O'06N
WH=20"
DC=1Q"
D LD=1J2Q"
DP=-30+35

1, 0
FIGURE B.28

1, 5
'
2 0
RESIN FEED RATE (ml /sec)

'"

<

Effects of the resin feed-rote on the overall stage efficiency for

different types of weir

t;.

1000'[

IYI

1M
~

,,~

ao

PRESENT
WORK
CO='06

,&;

---

:::: 100'
......

.OO
00

00

......

'i)

'i)

J. MULTISTAGE

0000
o

:J

meq/ml

>

ooc9::>

o cPCDOO<X>o

--...

:;,

1k'J

'i)
'i)

01

ml

'i)
'i)

HTU or HETS (feet)

FLUIDIZED-BEDS
TURNER(29)
eo='05meq/

10

la-O

FIGURE B.29 'A log-log plot of total flow velocity (u + v) versus HTU or HETS

.-

'I

148

APPENDIX C

Operational Performance Diagrams

,,:'

149

eXcessive fluidization
overfluidization
of fines
C

_.,,
.~ '5 liquid
,
o
bypassing I

()

~-

.!l1

.~

,,0

CJ

DL 0=1120'

I
I
I

:;

DC=1-0

I.Il

05

e::;::::::>

partial mObTlization

D~~

-.

..STRWL=3'3Er'
WH=2-Q'

operation

' --

0-

.~

"perfect"

-C-

flooding

o 0 '=":::----=-0..I;::-----:1:-1=---1~.5:---.---::-':---2~.5~-~~
resin feed rate(ml/sec)
FIGURE C.1.

Straight weir-effects of resin and liquid flaw.rates

~eftluldization of fines

<.)

>

',5

liquid
bypassing

,,

'--

CJ

g. 10
o

o~n

.~

-.--- .... '" '"

~

WH=20"
DLD=lI2Q"
DC=l'O

DP=-2(}+3)
mesh

.1"

'"

II

STR.WL=a38 .

/per~

rtI

I.Il
C

..

partial
_____
~
mobilization
~
flooding

---------

00 ~OO::----=-05~-.....j"-'---':1~'5--~2~'O:---~2~'--:-.nal;;oo..J
resin feed rate (ml/sec)
FIGURE C.2

-Straight weir-effects of resin pot"tide-size distribution.

/r--~---'---.----------'---r---..........--t
.-,20
--.~essive. fluidization
Ll,

''~",1
,I
I

150.

~I

"'E"

. overfluidization
_...

:>-

=:L> ,5
o

of fines

~.,('>;o

-'
Ql

::>

I1l

,.
\

liqu,id

'-'

.----

\'perfect"
~peration

bypassing

~ '0

I-

./ .
/

C!J

a.

::J
If)

o "

~O'5

- . __ ._-:---_ _.-:

'0
::J

. /.

srRWL=3.38"
WH=2.()"
DL0=3120"
DC=1'O"
DP=-30+35

me~
. -

_ _ _LI_.

partial mobili(:~tiori
_---.-.-----'.
flooding

-'

'<-

--..-

00 O!::-:'O: "-'---:O~5=----J.1'O~--1:-L:.5:---f-~2~'0--~---'2'5---3..1..'0--1
resin feed rate(ml/sec) ,

FIGURE C.3.
,
~

Straight weir - I'lffects of fluid-diode laterol displacement

overfluidization
---.:..:...:~___ of fi nes

ClJ

. .!Q20
E .

....
u

"perfect"

.2
~ ',5

operation

'-

10

'....

"-

<U

:QO5

partial mobi lizat ion..

.2-

.~

STRWL=33S"

,-.

Q)

::J

'"r

.'

~',

-'ru
a.

".------.
..

',~

u
::>-

tfl

. . excessive fluidization

_._

WH=l~"

, OLD=1 20"
DC=l'O"
DP=-3D+35
mesh

.-

O'OI"l~O~---::~-~~O--J~-5~'=~2L;:;O---;2~;5--~3{J~
'

U'

0-5
FIGURE: C.4

\esin'fee

flooding

rate(ml/~)

. ~., --,1

-___._ov.....;e~vf:...:.lu=i.:d~i-~~-. --__excessive
~
fl uidi Za t ion
, ,zation of
-e

'ii 1.5

--~

.'

e.

ope~ation
.

STRWL"3'38"
WH=Hl"
/ . OLO=3120"
i
=10" .
I

P =-30+35

'1::10
0,

BC

n.
::1

III

.'

~O'5
'

mash
e
,e_,~,--....:..-.,;~__,_

intermit-" "...

tent resjn ... .........

partial mobilization
..................
flow '::' ... ....... .......

.
r~--~~.~::--:-:-:~~-:~~-~~~;:::::::~~~~~~:-=-=-:-::~~~~
--.............

:J

..

-~
~

I.SI

"Perfect"

.0
,

.-

._----;

...

......

I
O,O~

FIGUR[ C.5.
t

Straight weir - orf~ell of wolr holght at 3/20 Inch fluld-dlodo

,Iatoral dis laeamont
'
,

,,-,2.0

,or

, E '
~

:-....

-, .-::: 1,5
'0
()

C'J

CONICAL WO=1,28"
'.
excessive'"
WH=2{)"
, .................
fluidization
OLO=1I20"
. over-...............
'.
DC=1'O" ..
DP=-30+35
~luidizat~_ .
mesh
, e~offlnes----_____
liquid

>

l .

(OJ

Ll.
'J

1\1

~I

I .

,~pet.fect'

.-.- - - . -...........

.-e-

operation

"'-.:e

~~

'ill
D

~._

bypass- I
ing/

0,5

e. " - -

------------~~~'.~-------.-------.-----

partial mobilization
e __----..~--e- e - - e - - - - - e - - - - - : - floo~ing

--:resin feed rate(ml;sec)

(}OOO~-~0~,5-,.--'~'O-'-~':5~-' 2I:::-O~~'.~.-:2~~--~30~
.

. F.IGURE C.6

.""

Ca~'eal wolr - effect. of re.ln and IIqL.ld flow-rola .

.- -

--

.-.-

...

",-"

.,.".

152

excC?ssive
. fluidization

CONICAL WD=1-0Q"
WH=2'Q"
finas
DLD=2I20"
oVC?rDC=1'0"
r.
fluidized ______
DP=-30+35
liquid r
.
._
~~ mesh
-~

bypass-/
ing/
........ ,1

.....

.1'--...

L-

(!J

0-

:J

V!

.205
....I1l

"perfect"operation

-----.-

- - - -_____

o~ OD

05

10

,.5

eovertluidization
of fines
e

G
15
o
(!J

"perfect" operation

',0

0-

:J.

V!

05

'.

excessive
_. fluidization

>-

\J
:J

25

Conlcol weir - cffes:ts of weir dlometer

N2.Q
E
g
>

20

. rC?sin feed rate(ml/SC?Cl

FIGURE C.7.

',;l
N

____

.....- - f';;l-=o-=od:;;::in:':g:-"-, - - - -. - - -e - e -

.....

~
(!J

---~L

partial mobilization

\J
:J

---.-

-.......;,

---.-

e---=

part ia l mobi li
.

CONICAL WD=1'28"
WH=2'0"
OLD=1120"
DC=1'0"

DP::-2O+~O

mesil

zati~on':----'~---'--=--------~--'.-.

,\:

~_ _ _J~.--~---tloOd ing

_.-

00 ~0.0~-O~.5=----1.,.J..,:.0.----1.l:"..5-.. --2-="O=----"":2~5--~3;.:.I
resin feed rate(ml/sec)
FIGURE C.8

Conicol weir - effects of resin portlcles-size distribution

~

..
i

...

s:

2.0

r-----~------_r------~------~------~----~~ 153

'.

excessive..

;'.5 -."""'---,
,~.CJ

liquid

>

bypass" II
Ing I

(!J

III

"'. -.

0P=-30+35
mesh

-e---___

.~'. _ .

'

.-~

"pertect"operation

D.
:J
III

overfluidlzation ot fines~.
,

.,"

.~lO
-.-

,205
~

.----.-------.--.-

.~-----------~-r~t~ia~l-m-o~b~il~iz-a~t~io-n--

III
N

'U

:3

-.-

-.--.----- flooding
,
05 '.
1-0
1'5
20
resin teed rateO'nl/sec)
fiGURE C. 9.

20

"-

...u>-15
o

>

Co.kol Ie -

3D

2'5

~
~''''1

.rr". " fI,id-dlod. lo"~ I dI.plo"m,

........-..........". excessiv~.
.
CONICAL WD=1Q8"
. ~flUld,lza,tlon
WH=1-O"
_
OL
0=1120" .
..

overe.
OC=1'0"
I
'
fluidizati0'if:----."
DP=-30+35
of
fines
mesh
I
I

liquid

(!J

.____

I
I

.........

.'

-e-.,:,

. _ .

"perfect"operation

.fT-'- -_ _~.~_ _ _~._ _ _ _.~_~

bypaSST
. Ing I

L.,

(!J

. ll.
:J
III

,
I

c
~O5
N

,
.. ,

1J

1'

III

~-

e~'fluldlzatlon

. CONICAL WD=1'28"
W H=20"
DLD=3120"
OC=1'0"

partial mobilization

..

:~.-------,
e_1
.
flood in
- _ _ _ .-------...t...--.-

00 O~.O~---:0:::-l:.5=----1:-J,.0::---:1;.1;:.5--""'::2;?O.---"ij2~.5----";'3{)~
resin feed rate(n:ll/sec)
FIGURE C.IO

Conicol w~'r - effects of weir h'!llghl 011/20 Inch f1uld-diodc

T'
:' .
'.

r------r------,-----~.-----~------~------~ 154

~
~

20

'.

.~ '5
o

'f-..
I

liquid'
I

el

......... ~

.::
1\1

:Q ,,0

CONICAL WD=1'28"
WH=l'O"
DLD=3I2Q"

DC=l'Q"
overfluidization - __
DP=-3Q+35
of fines
----_=--..!m~esh
- - .. .____

-e-

'............
excessive
. "'-.
fluidization

bypass-l

mg'

. . .

"perfect"operation
___

~,
~
I
::J
til

.(;:

.~.
.... OS
~

u'

------.-

~'.'

.____

-----~.----~~

part~ mol;>ilization

--'.'
--,

inter-.,
mltt-,
(mt

...

._---

~
resin'
~
. - - - - -__
- OD flow'---e-' flooding
.-----'.
0-0
1'5 .
2
25
resin feed rate(mllsec
FIGURE C.II.

~20
o
til

"E

.....

.8
>-

".o: : "S
o

(J)

>

I iqlJ'ld

.!]!

, CYUNDRICAL Wo=1'6g'
.
WH=2'0"
DLD=1120"
DC=1.0"
DP=-3Qt35
mesh
of fines

..

..-

" perfect"operation

'I

~ 10

,-------=.~---'--

-.-'-

-e-E,verfluidization
I
I
I
I
I

Conical weir - effecIsof weir height at 3/20 'Inch fluld-diode

. lalerol displacement

excessive
~
fluidization

<lJ

.
-----.--

bypass-J

ing:
:

~OS
N
u

--=------,-.-...;...--.,....-.floodi'ng

~_ _ _

,../'.~

:J

....

.~

0'00'0

partial mobilization

05

F.IGURE C .12

'lD

1-5

20

resin feed rat e ( m lIsec)

25 "

Cylindrical weir - effects ohesin and liquid flow-rates

30

I
\

--

S{
o

>-

.~

.--_.-

.. perfect"
operation

"'-.

>

"

'overfluidization
----.- . ___
of fines

~15
OJ

155

. fl uid izat ion

------.:xcesslve

(}2'0

"E

III

:;:: 1"0
L-

.YLINDRICAL WD:1-69"
WH:QOA
DLD=1ILO"
DC=lO"

.
~
._

D~~~

-r,~h

------.

CJ
0.

:J

U1

partial mobilizat ion

305
III
N

"0

.--..---.----.---...-----flooding

,/

O'OGO

05
FIGURE C.13

,fit
,.

1'.

1'5

20

resin feed rate (ml/sec)

25

Cylindrical weir - effects of resin particle-size distribution

30

.'

APP~DIX 0

Miscellaneous Figures

c
. ,1

"

156

157

CUSP C
o.

20-

~15-

o
0...

~..c

01

~10<l

o DLD=0'S35cm
OLD =1270 em

5-

-I---J...~""----J.._...L.--:-L_-'--_"---'-_'-~

5{)x103
REYNOLDS ~UM8ER, Re ')

FIGURE D.I

10{)x103

A plot of pressure-drop'TCtios (high/low) versus

Reynold numbers with data from the work of
Egan (7)

1,0',

RN+l' QN+l',YN+l L~.

eNx N
J,.

,
,.
~

STAGEN
Rn~

qn+t' Ln'Cn
, rYn+t lXn

(.)

z
a
u

~z

STAGE n

':...I

001
(I)'

,.

Ln-l'
Cn-l'
Xn-l

Rn'qn'
Yn

I'd

Z
-I

... .....

.....
Z
.....

I-

r-

STAGE 1
j

,
Rl' qt' Yl
.0"'4

"5

',6

EQUILlI3RIUM CONSTANT. K
FI (JIIRE

~"'-

..

:',,:.

...

[).?

A plot of initiol sodium chlorirlc

concentration in solution versus
the~quilibdum constant fornydrogen/
sodium exchange on 00wex-50-x6
.'
resin. 'i,Oata is from the work of
, Gill iland and Bn..Jdour (37)
',."

.'-

Lo 'Co'Xo

FIGURE 0.3.

Continuous COLJntp.rcurrcnt
1
mullisl"!),, operotioll
l~

co

"

1' l"j

i'

';jI

"}k

159

ION-EXCHANGE SYSTEM:
H~ Dowex-50W-x 8--Na+

k-_ _ _.... XB ,

VB

(f)

<[

0....

~O-5
(f)

0:::

>-

NTS=329

0-0 ~L-...I..._""":"";""';'....L-_...I...:-","""=,~-..L_--L._...J....._...1--:-I_
O{)
OS

X(UQ PHASE)
FIGURE D.4

Equilibrium diagram fOf theion-exchonge system'

showing the operating line and the number of
'
theoretical stages for a typ:cal continuous
countercurrent operation

160

APPENDIX E

Factorial Experimental Design

.o:.t.
.'

I,

"

..

161

Treotment
combination
I
a
b
ob
c
ac
bc
abe
TABLE E.I

Effect matrix

Mean

+
+
-

+
+
+
+
+
+
+
+

+
+

+
+

+
+
+
+

A'

- -

AB

AC

BC

+
+

- - - + - + + +
- - + +
+ - +
+
- +
+

+
+
-

ABC

+
+

+
+

Effect matrix generated for a 23 factorial design with factors: A, Band

C. In any treatment combination, the lOY( levels are denoted by "-"
signs and the high levels by "+,, signs. 'An interaction effect column is
obtained by multiplying the columns of those factors appearing in the
interacti~. To compute any effect, the particular effect column is
ossocia~d with the treatments and the algebraic sum mul tiplied by
(1/2)3-1.

Treatment
combination
I
a
b
ob
c
ac
be
abc

Rep'resentation
by signs

(I)
l+a
b+ab
c+ac
bc+abe
a-I
ab-b
oc-c
abc-be

(2)
l+a+b+ab
c+cic+bc+abc
a-l+ab-b
ac-c+abc-bc
b+ab-I-a
bc+abc-c-ac
ob-b-a+1
obc-bc-ac+c ..

(3)

l+a+b+ab+c+ac+bc+abe
a-l+ab-b+ac-c+abe-bc
b+ab-,-a+bc+aE
ab-b-a+I+abc- -ac+c
c+ac+bc,+abc-I -b-ab
ac-c+abc-bc-a+I-ab+b
bc+abe-c-ac-b-ab+l+a
obc-bc-ac+c-ab+b+a-I

=Total
=4A
=4B
=4AB
=4C
=4AC
=4BC
=4ABC

TABLE E.2 Yates' technique for generating the total effects for a 2 3 factorial
design with factors: A, Band C. The order af treatment
combinations in the column must'be maintained. The first half of
column (I) is obta ined by adding the responses of the treatment
combinations in adiacent pairs and the second harr by subtracting
the responses in pairs (first from second of the pair). Column (2)
is obtained in the same manner. The process is carried out 3 times
for a 23 experiment.

"

162

Tr:tent
combi tion
a
b
c
abc
TABLE E ,3

Effect matrix
Mean

+
+
+
'+

+'

AB

+
+

+
+

AC

BC

ABC

+
+
+
+

-+

3
Effect matrix generated far a 1/2 x 2 factorial design with
A, Bond C as the factars and ABC the defining contrast. The
estimate of ABC interaction effect is identical to that of the
general mean. The estimates of. the pairs A and BC, Bond AC,
and AB and C are also identical.

..
163

------

;-- -

----

Defining controsts b.1g1nning with rnain effects

167'
165

I;i,

156'
154

145*
143

134'
132

123'
121

112'

224*
222

213'

325'
323'
321'

314*
312'

426'
424'
:422'

415*
,413'

5i7*
525*
523'

516'
514'

626*
624*

'617'
615'

725'

716*

beginning with 2-factor interaction effects

266
264

:)/j

257"
255
253

246*
244
242

235*
233
231*

beginning with 3-factor interaction eFFects

j/.!

365
363

356*
.354
352'

347*
345*
343
.341'

336*
334*
332*

. beginning with 4-factor interaction eFFects

A ..... "

'1/ ,)

464
462'

455
453'
451'

446*
444
442*

43Z*
435*
433
431'

bog inning with 5-Foctor interaction effects

..

," "I') .

.)

563'
56 I'

554*
552'

545'
543*
541*

536*
534* .
532*

<.

"

!
I
I

heginning with 6-factor interaction effects

~.\

.i.

662*

L.

653*
651*

644*
642*

635*
633*

beginning with 7-factor interaction eFfects

76,1*

752*

743*

734*

r\:<LC E.4 All the possible sets of defining contrasts for splitting the 27 facttal
design (with Factors A, B, C, D, E, FI and G) into four equal blocks.
The number 176 represents the set of defining contrasts: A, ABCDEFG.
and (BCDEFG), and 167 represents A, BCDEFG and (ABCDEFG). The
"*" indicates that the a5sociated numbers appeared before in the list.

164

C?
0
I,
0
I

0
I
0
I
0
I
0
I
0
I
0
I
0
I
I
0
I
0
I
0
1
0
I
0
I
0
I
0
I
0

TABLE E.Sa

0
I
0
I
0
I
0
I
0
I
0
I
0
I
0
I
0
I
0
I
0
I

I
0
I
0
I

0
0
0
0
0
0
0
0
I
I
I
I
-I
I
I
I

0
0
I
I
0
0
I
I
I
I
0
0
I
I
0
0
I
I
0
0
I
I
0

0
0
0
0
I
0
I
0
0
I
0
I
I
I.
I
I
0
I.
0
I
I
I
I
I
0
0
0
0
I
0
I
0
0
I
0
I
I
I
I
I
0
0
0
0
0
I
0
I
0
0
0 I ' '0
I .0
0
I
0
I
0
I
I
I
I
I

O.
<,,0
0
0
0
0
0
0
I
'1
I
I
I
II
I'

O.

0
0
I
I
0
0
I
I

"
(I)
ab
de
abde
fg
abfg
def~
abd 9
cdf
abedf
cef
abeef
cdg
abedg
ceg
abeeg
adf

O
O
0
I
I
I
I

O
0
0
0
I
I
I
I

O
0
0
0
I
I
I
I
0
0
'0
'0
I
I
I
I

bdf

'-,/ '

"

aef
hef
adg
bdg
aeg
,peg
ac
be
acde
bede
acfg
befg
acdefg
bedefg

,.Treatment combiriations of the principal block for the 1/4replicate of a 27 factorial design. The factm are A(B,C,D,E,F,
and G andthe defining contrasts are ABC DE, ABCFG and
(DEFG). The treatment combinations are represented by the
numbers 0 and.! (mod 2) and also equivalently by the lower case
letters. "I" represents the leiter which exists in the treatment
combinations and '''0'' otherwise. The above treatment combinations satisfy the following modulo equations (mod 2) for the
principal block: xl~+X3+x4+x5 = 0 a!)d xl~+x:J+X6~ = 0

165

slock I
, (principal)

Block 2

Block-3

"

Block 4

F D A
AB
ABF
ABD
B
DE
DEF,
E
. ADE
AB[)E
ABDEF
ABE
BDE
FG
G
DFG
\ . AFG
ABFG
ABG
ABDFG
BFG
OEFG
DEG' :
EFG
ADFG
ABOEfG
ABDEG
ABEFG
bDEFG
, COF'
CD
, CF
ACDF
ABCDF
"ABCD
ABCF
BCDF
CEF
CE
CDEF.
ACEF
ABCEF "
ABCE
ABCDEF
BeEF
COG
CDFG
,CG
ACDG
,BCDG
ABCOG
ABCDFG
'ABCG
CEG
' CEFG, "
CpEG
ACEG
ABCEG '
ABCEFG
ABCPEG
BCEG
ADF
AD
AF
DF
BDF
/"'
BDBF
ABDF
AEF
AE,
ADEF
EF',
BEF
BE
BDE!'
' ABEF '" '
ADG
ADFG
AG
DG
BDG
' BDFG
'\
BGABDG
AEG
' AEFG
'ADEG
EG
, BEG
BEFG
'lWEG
ABEG '
,
AC
ACF
ACD
C
BC,
BCF
i
BCD
"
ABC,
ACDE
ACDEF
\
ACE
' CDE
BCDE
BCDE F
\\
BCE
ABCDE
ACFG
"ACG
ACDFG
: CFG
BCFG
'BCG
'~BCDFG
ABC,FG
ACDEFG
ACDEG
ACEFG
CDEFG
BCDEFG '
; ' , BCDEG." "
,,' BtEFG
,
'
A,BCDEFG

t-

TAllLE

~ ,5 b

The four blocks of effects for the 1/4 x 27 fac;orial'c!esign with factors' )
,A, B,C, D; E,r and G. and deflnln contrasts ABCDE, ABCFG and ,
(0 EfG) .' Treatment combinations f blocks 2" 3 and 4/satisfy the
three pairs of modulo equations:
'
"
" " xl+~+x3+X4+x5;' 0 ,(mod 2)
and'~I+x2+x3+X6:X7'=,1 (mood 2);
,
xl+X2+~+x4+X5 = ,I, (m od,2)
and xl+x2+x +x +x c 0' (mod 2" ond ,
3 6 7
xl+X.2+x3+x4'?<5 = I, (IIlod"'2)' "

.'~

~J

"

"""'~~,<',~ ~ I (m~2), ~;;"'Y'

~'

~,

,I.
166

J
Effect

I
A

C
D

F G
AB
AC
AD
AE
AF
AG
BC
BD
BE
BF
BG
CD
CE
CF
CG
DE
DF
DG
ADF .
AqG,
BDF

BDp
C~E

cqG

Aliases
ABCDE
BCDE
ACDE
ABDE
ABCE
ABeD
ABCDEF
ABCDEG
-CDE
BDE
BCE
.BCD
BCDEF
BCDEG
ADE
ACE
AC
A EF
CDEG
ABE
ABD
ABDEF
ABDEG
ABC
ABCEF
ABCEG
BCEF
.BCEG
ACEF
ACEG
ABEf.

ABCFG
BCFG
ACFG
ABFG
ABCDFG
ABCEFG'
ABCG
ABCF
CFG
BFG
BCDEG
BCEFG '
BCG
BCF
AFG
ACDFG
, ACEFG
ACG
ACF
. ABDF

I~:E

DEFG
ADEFG
BDEFG
CDEFG
EFG
DFG
DEG
DEF
ABDEFG
/ACDEFG
AEFG
ADFG
ADEG
ADEF
BCDEFG
BEFG
BDFG
BDEG
BDEF .
CEFG
CDFG
CDEG
CD'i{.
FG _.
EG
EF
AEG
AEF
BEG
BEF
CEG
CEF

(/

BcDE,FG
ABCDG
ABCDF
BCDG
BCDF
ACDG
ACDF
ABDG
ABDF",

TABLE EJ6 Alias gr

s-I" the 1/4 x 27 factorial design with factorS A,B,C,D,E,F
I /..-oM-'1G'ond defining contrasts ABCDE, ABCFG and (DEFG), The aliases
~re obtained by multiplying (mod 2) each effed by the defining
contrasts,

...

. ,

'

...

.
..

"

167

B 0

0
.'

F 0

G 0

./

",,'J"

-r--.

;-

E.I

Confounding pattern for tho 2-factorintor<Jctions

.
'.

. .

"

-168

- ..,'\

'. -'

.'
~

i.'

'

169
~llY,TS lr V:,~TA~1CE

TlI'lLE(ANOVI

OF
51UARE

~UH

OEG. OC

FREEno~

MEA"l SUM
OF SQIJA~E

SQUA~E
~ATtO

I"EAN

1.~

5.R31

~.(i

100.51t3
5.1131

9.804
, .56'l

7..636

1.~

1,lll(~I::

2'l.231t
.051

1.G
1.0

2.11311
2'l.234
.051

(A I ~ (nl F

2.3'l3

1.e

2.J'lJ

'.277
2.1151
.005
.23.3

ir:(nH

17.ItItQ

1.e

(,II (Olc

1.0
1.u

1.701
.31)0 '.

lIIlI r

3.6'l6
2R.6H,

17.1t1t0
3.6'l6
26.876

2.616

~(~IF

.013

1.e

.013

.001

133.353

1.e

R3.728

1.0

2?.1t311

1.0

lCO.543

I'll c

(,llnr,('11 F
(~I

rr

138.353
83.728'

13.491.
8 .164

HFL

"(FS)X(WL>
'(WH)x(WU

1.0

~FF

(A I 'Iel.r

(j, ) .1 C ( 1) (.

'l.777

1.0

1026.2'l4

1.e

102&.2'l1t

1'lIl.025

1. c

7.310

1 . 0

l'l8.G.25
7.310

2 1,.416

1.~

21t.1t16

2.381

12.5<;0

1.t

12.560

1.225

1.658

l~C

17.003
(.315

2.741
lQ2.405

1.e

2 . 741

1.C

Ui2.ItOS

127f" ~'l3

1.e
1.u

1276.S'l3
.656

3.200

1.0

3.200

.312

.1I'l0

1.C

.1I'lC

..087

lQO.072 . .WL
1'l.30-'l . . .(Wl)x(
.713

6.2'l2

( r,) '1

r ,-,

1"1 c"',

1.043

1.0

1'l.45'l

1.0

17.00.3
1.315

1.e

.128
267
17 .7.86 "F'S
124'.508 ..WH
.064

. Cont'd p.170

170

l.C
'IP"\L '1.\LIIFSI
;(t,:l,."j,!.lJ):::

F 11,&, a .<)9") =13.75

].71\

TMLE E. B An:!1 ysls of variance table

<

"

..
."

.1.'

"

J.

r---.f
v

'171
31~~TM~~~~~~~~~~~

0-12634
. G -11326

30

29
8]28
~26
~ 24
8]
o 20

,.

8G
A
AG

"
;'

8
DG

.:'
I:.'
0"

0::16
0

./

12
8
4 "
00"

o/'
.:"

3,

4'

5'

6'

8'

ABSOLUTE AVERM:. EFFECT

FIGURE E.2 . Half-normal plot for the absolute overoge .

. effects

30

29
8]28

SG
0AG A

al

~26

B
DG

~24
0::

W20
o

0/

g 16

",

,/

0,

(t/

12
6

\,

~\

p:

"

~I

- 00' . 0-5

1,0

1-5

,\

2,0

25

3'0

3'5. 40

STANDARDIZED ASS, AVG.

EFFECT
,

\\

G 6483

.~"

-.

......

FIGURE E.3 . Stor,dordlzed ha!f-normol plot for the .

',-ab,ol ute ",,"rag" effec tt
IZ' )

172

..

APPENDIX F

Half-normal Plot

/.

'.

"

,.
'/
./
,
/

Ij

.'

."

'173

The following discussion on half-normal plot Is adapted from the work by

Doniel. (49) ond also From the text by Johnson alid Leone (48).
The Jesul tsFrom a complete Factorlal (2 P) or a Fractional Factorial (2P - q )

..

experimental design can be studied using the analysis of variance (ANOVA) table.
Sometir""s there is a question as to which effects can reasonably be pooled together
into the residual sum of squar/ls, I.e. assumed to be zero. The half-normal plat
proposed by Daniel (49) is helpful in solving this problem. The plotting can also
be used to detect such deFects as the existence of wild valuei In a set of observations
(effects), the error variance being nat the same for all observations, or the planned
randomizat.ion be ing inadvertently modified (e:g.
.

~r plot splltting).-

'.
\f
..
.
The half-:normal plottin ,is based on the fact that the theoreti.cal cumulative
.

distribution of the absolute val

a normal distribution) is a

s of the average effects (u

lui, where u has

aig t line throuph the origin on a half-normal probabi-

lity poper. The slope of the s ra

ht.line

isfdete~ined by the poPulation standard

/1

deviation.
/

In the halF-normal plot, the averag eFfects are arranged in the order of
eltimated absolute magnitude. Th effec
r,obabil ity paper. ,T~ half-normal

ra

ar.e then plotted on half'-n"orma'. ~

bility F/o,per is obtained by modifying the
.

V", tical

scale of the normal. probabili

., The. range of P less than 50% is

deleted and the range of P greater than

r

2P - 100

where P is the

cumulativ~

prooo'bility of t e normal

distributio~ dnd p'

the cumuli::sti;e

174
"

distributio~ used in the half-normal paper. Instead of 0 cumulative probability

far Ihe vertical scole, an ordered rank of i = I, 2, 3, , n for the emperical

dJla is u,sed. A general conversion formula for the vertical scale is

p' = (i- ';2)/n

=1,2, ... n)

where n is the number of orthogonal contrasts and n + I is then the number of

"

observations (effects) in the experimental design. The.abscissae are the absolute

v~lues o( the average effects.

Iri a standardized half-n0"1al plot, the overage effects are stondardized and

"

significance limits can be placed on the chart. The standardization Is performed

by using ua as an approximation to the stand.ard deviation (a). u a is the
absol

10

(0.

value of the rank QI'CIer statistic from the n average effects that is nearest

13 n + 0.5). ~s of u~

for different values of n are given below

d
n

15 '
31
63
127
where

11

2.09
2.42
2.6?
2.88

rank of u
a

a =0.05

0.20

0.40

II

0.94
0.94
0.66
0.66

0.26
0.33
0.23
0.25

-0.14
-0.05
-0.04
0.00

22
44

88

n = rumber of orthogo~al contrasts,

~ ~XpeC!ed. volue of the rotio of the~rgest absolute value of the n

, . observed values, to the corresponding a-estimatOr, lfx

/
.
;

'\

fre,quency of false positives per experiment

~.

deviations to right of centre line.

175

i
I

The "centre line" and "significance limits" for the-standard plot are determined

using the constants in the above tobie,

Centre line:

Through poi nts- (0, O) and (

,n)

Significance limit: Through points (T + d,n) and (I, rank (u )}.

,.

176

APPENDIX G

Comp.uter Program for Mass-Tr.ansfer Experiments

,
", .~

<--',

.I~ 1

(START)

,
~

..

/READ OPERATING CONDITIONS?

ICALC. OTHER EXPT CONDITION S I

-

/ READ BED-HEIGHT RESULTS 7

I CALC. RESIN HOLDUP & BED VarDAGEJ

~

.J

/ READ ANALYTICAL RESULTS?

~

-.

ICALC. CHEMICAL COMPOSITIONS I

CALC. NTS,HETS, PLA1E

EFF. 8. VOLUMETRIC EFF.

I MASS BAL ANCES I /READ COMMENTS /

~

/PRINT RESULTS/

.. IBENSON-LEHNER PLOT. I
( STc P)

,
I

1"!C',lRE 0.1

"

.. .., .

Flow thart of the '_omputer protJram for analJ si ;"

' . ments
the resul Is of mou-trans f er C)(PUI

'~.;

-~~:,:

-'

i.

1 .

',- .

COMPUTER PROGRAM USTING

',l\.)

" 178

CO~IPUTER PROGRAM TO CAI.CUI.AJE niE NUMBER OF THEORET ICAI. STAGES

INTSI. HEIGl-iT EQUIVAI.ENT,OF A THEORETICAl. STAGE IHETSI. AND PI.ATE
AND VOLUMETRIC EFFICIENCIESUF A cUNTINUUUS CUUN1ERCUKKEhT MUI.TISTAGE IONEXCHANGE c0l.uMN WITH FI.UID DIODE DUWNCUMER
TtlE PROGRAM ALSO PLOTS THE EaUILI~RluM DIAGRAM. OPERATING I.INES
AND THE THEURE TJ CAL STAGES US l,iG THE BENSU'~-L.EH''IEK PL.UTT I"G .'(S TEM
\,;HITTEN BY C.Y.F. L A W '
'
JULY. 197'1
(JP!ENSTON XI201 .Y1201 .COMI801.FBEDI12 1 .SBEDIIZI.
~I~:FNSION RVI 141. R VF141 .TVI14 1 .TVFI41.CLMVI14 1 .CLC"'VFI4 1 P HI 141.'
I
PHFl41
'
DIMfN510N Rl IItI.RFI41 .CL1I41 .CLFI41 .H1141 .HF14 1
DIMFNSIUN XEU~4001.YEYI4001.xPIZOI.yPIZOI
REAL INTER.LABXTP.LABYTP.L~BXBP.I.ABYBP

wH I IE Tl TLE
I,:R IT E 1(, 101
rORMATI IHI. * CONTINUOUS COUNTERCURRENT MULTISTAGE ION-EXCHANGE W
IIIH* II * FLUID DIUDE D~WNCOMER* III * EXPENIME"TS WE"E PENFURMED
l~IT~ A COLUMN OF FOUR INCHES INSIDE* II * DIAMETER.
THE EXCHANG~R
1 HAS SIX STAGES SPACED 4.5 INCHES APART*
II * THE ION-EXCHAN
1,(,( SYSTEo-I IS SUD I ur', IUN:DUWEX-50WX8 KESII'.-HYDKUGEI'I IU"*I
'
(OllNTER"l.O
(aNT I NilE

t..

"

.0

BEAD AND WR)TE RUN NUMAER

REIID 15.70 I RUN
FOR~IITIAIOI

IrIFOFIC;119~9.30

,."

."

(ONT INliE
Ifl(UlINTER.EU.l.OlWRITEI6.401 RUN
ruiIMIIT(111 * RUN NUMBER
*. Al0. I. 15H ~*************l
IF(C UUNTER.GT.l.OIWRITEI6.35I RUN
r(lR~IITllHl' * RUN NUMBE~ *. ~lU. /. 17H ***************1
11[~I)

wEIR TyPE NUMBER

IYI'[=1.0 FOR STRAIGHT WEIR
1Y1'1>?~0 FOR cyLINDRICAL ~'EIR
TY"t'=1.0 FORCONICAL WEIR
'~f",\D 1c, ,RO I TYPE
[011':11111'10.01

Pi ,\11 IINIl WR ITE- OPERA Tl NG PARAME TERS

cuLlI,.''''' DI(01E,'1S,IU,, IN I,,,CHES. SUD IU'" CU''iCE'jHKATlu"
i'~kIICLE SIZE IN MESH NUMBERS
i'l I, I) 15.50 IF L F S. SIC ~iL WH. Dt OLD. DP

p" '~E\J/'~L. AI'D

rl)I;~IITI7FIO.O.AIOI

h111[lh.60IFL.FS.SIC
~Gi;~IITIII. * OPERATING PARAMETERS* III.
IS\, 'LIOUID FLOW RATEIML/SECI~*.F7.3. II.
",X. 'SOLID FLOW RATEIML,SECI" F6.3. //':'
",j,. 'INIHAL 'SUDIUM CUNc.I"IE\.i/,'ILl=*. F6.31
II (TYPF.EU.l.OIWRI TE'16 .,95 )\~L
f,,:CIITI I . 5X. *STRAIGHT~:EIN LEHGTHII,'1CHI=*. 'F6.3)
III ;vi'r.EU.2.0JWRITEI6.10511;,L
'
f,','.,'IITI I . 'iX. *CYLINDRICAL WE'IR LJIAr-'.[:TERllNCHl=*. F6.3 '
I' I 'Yf'f.EQ.l.0ll~RITEI6.20o;l\'L
C\";,"\JI I. >;X. *CONICAL WEIR DIAMETEI{IINCHI=*. F6.3 1

..

'

..

~-;"-{~' ..l:'l:.H,I')C.DLr,'t')U

/~

.".".i(

~;(.

,,~f:ru-'~
,

.) / :>

~."

L;.TF'\t.i- c.'.:~,r-L:'CE"E';l(:/~U !I"CH)='",

~J f. P. TIC L F S I.Z =- ( ;.. ~ c .:, t-d = .:.,..,
,:\ 1 .~ i
/

<J. :- _~ t i.

~6.3,

_r_~:.~c (";,, ..,;'<lTr

: 'T :- { t-... :-: \ )
=- _ " ' : . : { : I /,
-( .vTHFI'
~

*-};'r.:!I\ I~EIGHT(I'''CH'I='''', F6.;, I I ,

CLEt\,-(',.C~.t..l"iCr.)="'"
F6.3.,. iI,

;;~':".,;jCuj:::r\

'"

\79

,,z.;:.?,_,Y
~

~.T,ili( U"t::-ATI:.G

II.,

(.:"JI1I'),,:,

;- X~""r.i\ I,F.~r .\L.: C...;,~:-.. I i I """Sri)

:~

:'llr.-'=:.GT.l.:,)GU TLJ 215

~ = "'.l.~ ? c., If
., =,:.,; ~~ T ( r~ If ~~ ? ...) - ( 'I! L / 2 .J ) .... 2 0 )

: c I " C':l . ) ;, ) T H F T A 1 "> -, 0 i:l

"

:~I~.Gl.~.u){HFTA=ATAM(~L/(?,.OH))

.=:IO~=(~*?J)THfTA

.cG,F:n=:'::CT..,K-('::L*H/2.Gl
c~

)S IH~ CATCHING AREA OF ~EIR

:.~

'IS. THE

~OTTU~l

AREA 0,. \';EII-<

:. ~ = :,')= SE G:';H! T
',-' TO 7?">

:. ~:~: T I :'-!I.J~
1=ITr~F.GT.2.0)GO TO 285
:r:~D=0.25*3.1~16"I~L"2.~41"*2.J

n5

.:' TO

~v'.lINUE

.'/ : :,.7 'i"" . 1 4 1 6" ( 'tiL" 2 5 4 ) "* 2 0

',"'''''::. ?C,'\f-":l..1

'~16*

.',

''',

(U.6P7t:..*?:;4 )**2.-,

" '. T I 'l'.l::'. ' - , ) " 1 "

.~. J
. : :-:, r It.. T
':,\:JI:T

c:- :c :,. r -

* I R r, " J 0')
.

o:,F/;,[,'

: T t ( h 1 .. " ) AT

,AF A C AR 1 ,AK 2 ,Ior<

. "Ail". ~X, *t~LJSS-SECTlv~AL A~E~(Su.C~.)~~, F7.3, II,

'. 'FL~lrlz~rlaN .R~~I~U.CM.I=*, F7.3, II,
" *r;a ..,t',t'-":FK CATCHII,C;' Al<E~,(SI.1.C,I.)=". F7.3,11,
' . > F R ACT I t.l~.
F FLu I D IZ A T 10/, :. K ~ A=". F6. 3, /I,

.;

FR~CTic~ OF COw~COMEI-< AREA=, F&.3, I I .

Jt~LI_.iDIl:\Tlui . . Ti.J Du,\I"C V1 l t l'"AI\E.A I"\A1I\.1=*. F6.3J

'"
'~

.::~L/I~7

,.ci'L/AF
: 1r (

11

,.,fl//.

. r. 5 V .r S V

<,X . *IJVF~ALL

.5U"E"FICIAL VEL.I'''''/S~.CI''-SEC)=*,
5X. *FLUIDIlATI""SU"'E"FICIAL VEL('I'L/S~.(t'I-SEC)=A, F6,31
(vLlJ"'"

.'. II.
cl,"-fl()*7.">4
:i,cl~.l?'u)DS

~il

II.

.,

',;;'(Ir,(,

-: .

I.~"l.l,

- i""

-,:)(}

')X, *0",;:-.(01'EI{ SEALI(o'tl=", F6.3)

IS PS 11. Ir--.CHt::S

.?r,~

::' I',. 1 ;>:' ) T DL

';1/.

sx. "TOTAL DIUDE LFNGTHICM)=*, F6.3)

ISO
~'i;.?70IVO:).WATER

.,'.:

\TPFl,).'~1

."-,.0

.\NCl '::i~ITf' FXPERI:;r .. U,L HESULIS

I"'''uT DI.J{, IN (NCH!:.S '

,'Ei~f1T

I.~

13ft> HEI(,HT

";~I",j'<;1

FXPEHI;J,Efo:TAL RESuLT::,*)
'i) IF' F [) I I h lor I ,I? I
:.' i c, , 1 "5 ) IS" ': 9 I I ) .. I =I 1 2 I
. _ ,",'.\ T.( ~r 1 J . ,)
~_'
lC,C,
1=1.,17
CC'~I I )=FnE"1 1)*7.<;4
"""1 II=sn.E[)1 I 1*2.'i4
','TI'WE
',,",\11///,.<
:, ' I ; )4

,~LL

~TATIFBED,12.FMEA~.FSTD)

. !~L ~TATISDED.12.~M~AN.~~ID'
j " : vvLlJ"'EuCCU"'IED BY THE DluDE
IS ACCJU"T!:.L> Fvt< I" CALCULATI"G THE:.
. 'c,l VJLU:Il::S
",!ICK=il.R281*2.54
r,o,: DiaD;: \.IDTH I-IITH lEKU LAiERAL UIS~LACEMENT Is 1.2562 I"CHES
'~!rTII=(1.2'i&?+O.U'i*DL)I*2.54

'-:,::".\ = f)TH I CK"['1" 10TH

Fn,=F~rA~+TDL-PS*2.54
i I I , .' :,. L E li. v ) [j E L T II F V= (j.

:rl=)S.~T.O.J)DELTAFV=FDS*OAI~EA

" "., =:.,';EMH T)L -P S * 2.54

1=( ~,O s. L F ~J. ,) ) f)E L T t, S V= u. ()

...

:'ISDS.GT.G.;\)DELTASV=SDS~OAkFA
'""V=F~FnN*~F-DELTAFV

'. '''V=S,\F.,\N*IIF-DfL "r'ASV'

,JIP= I r!li'V-S~~W"1 l.a-Vv[DI I/FE/V
,,": i, 1 ( Ito? IF':FM"Si/,EAN.FSI,).:'STD,FeMV.SerW
.,[(1, 75X, "'FLUIDIlEO
SETIL..EO*. I. 28X, *BEO", lOX, "BED*,
"~, '~En~ HEIG~T(CI~I*. FI3.3. FI3.3,
.' ~\, HFIG~T STD [)VN*, F14.3, FI3.3,
I f . ''x, *r'\FA,'. vOLle.C.I*, F14.3, F13.3)
.;.: j f (to. 70) F~"!V.) I [)
' . - - t I I I I , 5X. *;lEAN Fl-UIllIlI:U 9LU VO!uAl>I:=*, F6.31
r.,.111?~)4 :
c,

~::-"

(6 ~ 1 7~) ~.:~

.,111,

'iX,

"1"U'TE9 :~EI" HEIGHTIC"I)=*. F6.3, *1"

LALCULA.Tt: ANU .... RITt: THi:. INITIAL 'AALJ ~INAL

e""CFr..J!<ATI""S A"O HYDK... GE" C.J"CE''TKATI'-'''S
'.'. ':U!"SfR OF SF"TS OF SA'-1PLE DATA (MAX!KUM OF FOURI
.
:" .. LYTIC':'L RESLiLb.

" ; i " (hf-'I(ITlr-$,

"~".

~UDIU'"

IS THL SvUIU", HYu,",X'iDt T-IT.<A"T "vK"AL..ITYI=,,t:,"'/"'L..I

. = ,:. J
fir ''.DIi~r,~,

. '; {A.? P (1

I"

"RESI~ CAf-'.('~fU/MLI*, 5X. "CHLvrtIOEIME"'/~'LI*, 5X,

.'Cr"!!i:r"/::LIlt. /1'luX, *INITIAL*. 4J1.; "FINAL*. 7J1.. *INITIAL*
... "I",\L<, ~X, *INITIAL'"
4)(. "FINAL*,/I.
, \v'-v',r I.' "lL '-'F r'AC"ED 8'1:0' II Tt<A"T-- VVL:')"'E I" "''-, CHLvt<It>E
1>- .' ... ,.,,\, Iv'. I .. ,,V, A"I) tiYOf\v(jt.,. Cv .. Ct"TKATIv"
I .. I"H VAL..Uc.

"

"I/I'I'~'

,
\

"

181
,C,,'{',.:,~I

((UI,Vlll) .I=l ,,',I

'.'~I c, ~!,O

I (PI'.! (I) ; 1=,: ',r'!)

.' .. ,~ " ! :; ~ !. c; ) ("( L .\~v::- (. I ) , I = 1 ,\ J
PHF( i) ,I=I,N)
;' V I ( I ) , I = 1 , " )
VI (I) 01=1,:-1)
-......~_'-"......,_..., (KVF( 1101=).''0)
, __ ~ I 'i :, ~ ) ( TV F ( j ) , I =1 , ,l )
: ;}',i( LF l v .,;)
/
,~
, " ,_ I = 1 ,,":
, , I 1 I = T\ ()" :'P'l' V I ( I ) I ( R V I ; I ) ", ( 1 0 - V0 I D) )
" I 1 I = T~;Q;":':'TVF (I) I (RVF (1)* (l.u-VOID) I
:'L :;r[ = -,-'. ;::'1 A A05 '
::, T~ .-;; = - 0 09547 5 4..?
.1 =CL;':'JII I I *SLUI-';:+i'i'-lE;'
.i=CL"vF'{1 I'SLOPE+llqEI<,
"IYI.LF.L,.UICLI (I l=l.'0Ill;.ij**(AUS(Y! I I
:={Yl ,(,T.C.O)(LI (I 1=10.lJ**YI
:' Iyc.Lf.CdICLF( 11=1.0/1lJ.,j**(AhS(y;=1 I
IcIYF.GT.O.OICLF(ll=llJ.lJ**YF
.. I I I I = ) J f ( J c,. 0 * * ( PHI ( I I I I
""I II =] .~/(1~.C**(PHF( I I I I
, ~ ! T-'- I " ; 5 c, I K I ( I ) , K;= ( I I , eLI ( I I , C L F ( r I , H I ( I I , HF ( I I
'

.-

"

"

'C~"ATIFl~.3,Fl1.3,Flj~4,FiC.4,FlO.j.F11.4j
:~.'H 1':UC
' '

;~LL ~TAT

TO CALc0LAT~ MEANS AND STANDARD DEVIATIONS OF CON(ENTRA~Ib~

'I' A:,8 PRINT RESULTS

i...

~TATI~I,N,kIMEAN,kISTDI

.' _....

~T~r(K=,N,I'FMEA~'~FSTOl

:.

"

'

, : l L J T ,~ T ( C L I , N , C L I ~: E AIII , C L I ;:, TD I

:. 1 .\ r 1<: L F ,
";\

,~

.'

, C ~ F 1'0 E: A,' , C L. F S Tl> I

~T,~r(HI,N,'I\lEAt;''''ISTCI

',~ .. ST,q (HF ,N'~~{'~EA'\;~HFSTDI .

,'," "
, ' '
, ; , cit, ." 'i III I ME,,!' , RF I/'r.!, " , C L I e,E ,~I'I ,C LF .'.EA,' , H IhEA!', HF"If-A"
'~I( I . 5X, .~EAN*, F6.3,Fll.3,F~3.4,F10.4,FllJ.3,Fll.4j
:'::FI~.7~IHISTO'~FSTD,CLISTD,CLFSTD,HISTD'HFSTD

,\i

I I,

" i(=CL

)X, "STr: D'/"~'

F?3,FU'.::>"Fl:;.4,F1o.<+,FlO.j',FJ.1.41

I;':E:~,\J

._:; =::Lc... Ft\~J

. : ~::,J

lOOt:'

li~l

"'-=;:r"'T~\i'l

.. ,:,,'''I .. G ::'TATEioIE"T .... HE,' THEr<t. IS Ai, II~::,uFFICIE!\T ~IASH ,,1" KESI"
RAT ION
1:, IHE EXCH/I,\GEf,t\Lt ~lSII' CP'ACITY utlTAII,t.O t:XI-'EI\II"C'~TALL.Y
;'~f' (;r~~f

'co.f,SJ

;:,=:xl<C-RIC
LT~.GF.0.{j)GC

,,~ (;,.'?~)

.TO 115

rXRC

,.~(//, ~x, *EXCHA~Gt.AbL[

( f, ~

1 ,Q

:", )

.ill" ;'X,,*~WTt:') INITIAL RtSI" CAPAliTY llfH:."Tl'-R TtlnN t.}\Cn"NlJo::."

.~l;. CAPACI TY*, 1/,' llX, *D0E TO INSuFFICIEIH "ASHII~G AFTEI{ ~
~CFt"FI<I\TJ()N~
Y( 110*, I I , 11.1., *S'ET E",UAL TO LEf(U*1


182

;:-<.LIC
: ' ,; -0. "!<.I ,C),.p; L " :, 1 C )

: 16ol4C)RIF
-~.
'r I I I . sX, "lATI ' uF HY[).huGE" Tv SvDliJ," Iv" FLuw=*, F&.3)
1 S THe SLOPE 0 . TH l FLOI\ OP[,<II T 1,',,(, L I,',f
-1.c'/RIF

_;

..

~(If,,?C,lFOLS

.-,,,,~rll,

5;<, *:,LuPE vF FL..V"

V .... EI<ATI'GL..I"E(L..C/""')=",

'

F&.3)

_.LlJLIIT~

TH~ INITIAL k~U ~INAL Iu~ FRAlTIUNS

, "",l?E lUN,FRIICTIO,', 11. THE LluUIO "HAS"

: I', HIE 101, FRACT lor. 1" THE I{E!>II\ PHASE
:=C=SIC-FS*(RIC-RFC)/FL
. r ',Fe I I S I C+H 1()
" . I C I { S I (+H I ( )

';=~rLTA/EXRC

:={rFLTA.LT.C.OtYT=O.O
., ([,cLTA.Gl.u.u)YB={EXRC-RFC)/EXRC
:'[orLTII.LT.G.U)YB~(RIC-~FC)./EXRC

-,\c(tlLATE A~D ~,RITE THF SLOPE OF THE ANALYSI:, OPEHATli-iG LINe;

15 T~F SLOPf OF THE ANALYSIS O"ERATING LINE

',L~

\~L~.{YR~YT)/{X~-XT)

..,:~ITF{6ol5)IIOLS
,

5X, "'SLui-'E vF A"AL.YSIS vr-E~ATI"G L.I"E{lYB-YTl/(Xo-XTl)=*

:v"',.\III.

-:,L'JUTF THF "'u,\1E~ER UF THEOkETICAL STAGES

,- is THi' INITIAL SODIU'," CONC,,;nl<,qICJ.~ 1"
c(-ftLC:GI0(SIC)+17.75)/12.5
. ; : l

i~ni{I!t,L(=~lEU/MLl

~ y~

!~)

I,:: 1 9'?\1
': I.Y{I I/{FK-IEK-l.~)*){
;,(11-XTI16J.l&U,11G
1:+11=YR-{XR-XII1)*AOL:.,

Ill~

.:.1'

-- ' ,; I': LJ F.
: ) I 1 R lO .1 8 ;) , j1

-J+

,i .

i"

.:

Q l;

(XT-X{I-lll/{Xlll-1(I-l)

195

,=(XT~>:",l/{X(I)-X'3)

Tl'~uE

_LATt rllTS, AMU PLATt IINU VOLUMETRIC

, ~) I>

EF~ICIEN(I~S

? S (l

,. ""'liS' Ill.. :'/P;'j

. '='lhvU. )"lrL+FS1/(HET:,'(q)
:-, {(,,2Gl,l
.',TUI, '>X, .~, IS THE Sui>1U'" Iv" FKACTlv" I" ,LI",UIl) .... 11A!>E*' I I ,
'Y IS TH" ~,vDIur. 101 . FrtACTIU" Il~ rtESI" t"HASE*)

';"{6,21 :lXP.XT,YT.YP,,<i\T$,HETS,t'EFF,VEFF.

183
=_,~rl
. :

-,

.:

.'.

~x,

;.

*IIIN1=*, F6.3.

luA. *110UTI=*, F6.3, I I

, (d"I=. F,,~3, l:';X, *YluUT1=*, F6.3. II

, >,.1 ~:-" F6.3, 111, "HET5IC",I=", F7.3, I I .

'~L'i~
'Vv~

. ,,
~

..

EFFICIENCYIMEKCENT1~".

~F~ICI~NCYII/H~I=*,

C.,C,,,: ,,:.55 ",-"LANCE FOR CHLORll>1::

F7.3. II,

F7.~1
AN~

HYDROGEN

IO~5

, ". ".: ! ;, r ;; :. L "N (E

, :,;:Cl.i;:. .. r=L

....

... _ JT=CI....FC .... (FL+':::\TEK)

.c"T.\(L=Cc.I,,-CLUUT '
; '-, .;,~a = I :," 5 IDE L T ACL II CL IN)" 1 Ou. J '
-Y~~UGE~ PAL~~CE
'
-!'.=,l!C*"S+Hlt*F
-CJT=RFC'FS+HFC* FL+WATER)
.=~ T ,'t1=H I /i-HOUT

:1'

~";::0!~j~= {.:~~::,

(D

, ,-, i TEl 6 22 G 1C

" , H I'~ CLIJU T HVU T ,DEl. T AC'- , DE L TAH E""I."{C '- , I:: I<I<I."'H
tv",,:.\!I/!/, " IUNIC MA::';:' BALA/-ICES*. 1',.25)(,*CHLvHDE*, 51.
:,,,y'),<UGP.*, I
5
*IUo~ INIMEW/5EC)*.- FIl.3. F13.3. I I . 5X.
"lvi, uUT~"E""5E )"" FI0.3, F'\3.3, 1/. 5X~ *1,,-uUTlo'E""5EC)*.
'='11.3. F13.3, I I . 51, *AR5(~-UUT)/IU~ IN*, F8.3,
FI3.3, /,
.
,.'-x. "IPERCE.'lll")

~~n

"

~~D

~~ITE

COMMENTS

. , ivH r c;,<uS. AKE f{E\.IUIKt.~.

"F A D I ~ "<!..1 C ) I CU~H 1 I , I = 1 ,4 0')
;:(;ld:AT (Ij~::)
:.: : TEl f, , ? l.:- )

ThE F IK:;T

~O

COLUo1'i5 uF EACH ,-AK{) ,IIKt. U5t:.IJ

=_:<;',ATIIII. cOMMENTS*)
, ! r F If:,. 2 S ~ ) I (0:\ I I ) , I = 1 , '1 u )
=_~':~T(II.

"F~UIDIZATIUN9

5x.

*, 5Alo. I, 20X, 5AIO)

")WRITEI6,235)

::lru'll"I.~'F.IU'

*)

IF., 2 ~ ~ I ( CV,1 ( 1 ) I =11 , 20 )

c
"v,,AT(5A. '~E~l~ FL0~9
*, 5AI~. I. l8X, SAlOl
!CI(O>:(161.,'IF.l'lH
)WKI'TEI6.2351
" : T~

fl

'TFIA,?4~I(CO~(ll.I=?1.30)

)".\T(~)(. 'OVF'l/lLL OPERATION9'" *. 5AlO. I. 25X, ~AH)I

C,~\~ 12'6 I \;r-. lOH
) iR 1Tc::t 6. 2"1~)

;--1

~T~16,2A~:ICO~III.I=31.401

'''.\TI~~,

':JPER/lTIOrl TliloE9

. *, 5AIU; I. 22x. SAlOl

FvLL. ......,I',G 5ECT1\.1.; P'-vTS THE E",UILIBI<IUI"l DIAGKAoo.

iH" TH~0i;fFT ICAL STAGES
...
'C'iU',F".CT.l.O)GO TO 400
~~ ~L~TT~H UNIT AS CENTI"'ETEk
~L0Tlv.~,C.u,-20)

.L

. - t

, ,,
I

UKIGli,

~L~TTfR

~ L .....

T ( =- '. ,

I~

,~)

- ,

,'.;:. .;~ ,'\i.; DA Tf

T f I 1 L..I t:T) /1, T F )

r).\

L:. T Tf t\ ( G., l'. 6 ,90. \) , 1. 0, 3.0, 9rlC. Y f LA ...

L~rTE~ll~,0.5.~C.O.2.l.!.O.THECATEI

f
184
- 1":

t"JLlITTEI'\ u"IiJ! .,

-.

.,... r-

''':':'1''

eJ L L: T ( 'i ( , J. 0

-----... ' . __ l.

't -

'1 ]'

\ '"\

r "j: J:-

~L()T(~l.~.C.u.-~)
.T!.\~J::
. ~ :,.:_

r'~ ......

rTI':.J SC,\Ll:. '-:ACTv:'( :.. .. ..; .,JvU .... Li'\'''':LS ruf'\'X

Alit)

_ _ v- ~!\L~=l.'~/t'i.:)

"(C"lLF

1."/l,.i:

'-,.~'XSC!\LF
,c-l

,.'

.yc,C\LF

,"I:!=-,.S*XS(ALE
,~~X=I~.~XSCALE

\ , 1;. = - 1 "

~'y SC ALE

Y'~\=l~.JYSCALE

::';,LL ,-JLT
~'I~

1"(XSCJ\LE',(SC~\""~'V'~oJ'X.~I,."x'';rx:YIa1j~.Yf'':''X)

co~NFRS

OF EACH PLOT
~o 410

:'t(nUNT[R.GT.l.O)GO
.

(:.cL

PLCT~1.v.u,.~,.,1)

C.LL

PLJT(,~.D.O.~.7)

'~:'LL

C;,LL

PLOTt v.v,2-0,21
PLUT( ).G.26.v.3)

:~~L

PLUT(U.~.27.5.2)

C~LL

PLOT(1.J~27.5.2)

T':; ~lO'
"\'--~T
~:~.L

It'llJE
.....
PLOT(1.i.;,Q.v,2)

:~LL

f'LO~t~.~.27.5.3)

,'_L "LOT(1.c:.;>J.').2)
'. i I .'o~.J E
' : .. L !-'LUTt2 . <;.77.5.,)
:O_L uLOTt?1.~.?7.~.~)
": L f'LOTt?l.~.2f J.;
::,LL PLOT(?1.'i.?l.."I)
',' L PL Crt 71 'i V. ,J .7 )

,.

",~~

f'LGTI7:...C, C.(;.2)

'~T

..:u .. r,udlEll

':LI,

L~TTf~t3.C.~.C.~.3.5.25.5.3HKU~)

':~L

~~TTrRI4.J.5.~.0.~.5.25.~.~Uh)

I F'P'::: OF Till: FOUrLiRRIU~'; [)(AGRM:

~)L:JT (~.c; ,?5.;';,1)
c'.

; ,

~LOTIIR.'i.25.:.7)

~LOT(lR.5,1~.:,21

I'LOTI,.'i.10.:J.;
,) Lu T ( 1 c:: ,,2 ~ u ,2 )
t. ") ';

I: 1

"C)

: -: r

""'~'.--1.C;~(I

'"I

DL:';TI1.'i.yy.~)

\ .::: "l.

( : .~'';.I)

',I

1 YX=~.9

~LOT(XX'YY'2)

.f

, .. ', i 1,',t,l:
1

f,1

'!;:"1

C;

: !:-1
,_,:=,.s+1.:"*~I

.
.'.,;" r.lL....;1 {'II,:.:; ...
, ': = 1 . ~
: :. ( i ~ ::. ;, 1 Y Y = 1 :) I.

~:.~L~

,"

~'L,-,T

(-<.x.yy,;

1 = 1,0.

/,I.l

. ~ := 1 .' ~ + 1
~.~LL

'=!l

=j

-* -::-

r'Ll)T(lA.5,YY,~)'

'\:=lQ~

: F ( I F () " ) XX= 1 fl. 1

;~Ll.

:J

!...

~L~T(XX.YY.~)

I=l.c-

':-.r!

; I =I
":=l~.'i-l.'i*Cl

:.'LL PLOTIXX,25.\",3)
\' Y::- 7 1+ H

l'II.:::Cl.5"lYY=24.6
r~LI"

PL~T(XX,YY.2)

': \.':.: T I "~l;E

.

'.

r:~~~ATE

P01NTS F0R

THfEWUILIR~IUM

ClJ~VE

\::~{11=':.).C

'i?_ 1=),3~!
, '" ,; II )
XF ~ ( I ) / ( I C+ ( 'Of. ': ~ I 1+ 1 1 = x:: l.; ( ; ) + 1 G / ,,~ a a
'; j I:, Ur

'J

="""

, , '1 [oJl! IL I f'R liP' CUR'/E

c, ',il' "L j""L (x::U,YEC, .1vl)

:,-,'i:,LY~IS .OPERATlnG

'. ,"1
,'\1,1

1 61 ~;: J

I ) )

,c

.,

Ll'lE

l~r'I

lTu(X'l,YB,XBf.',YBf.')
" :, ,- L VI', I r I u ( AT., YT , XT'r ,y TP )
',l.1. eRAF (X,!p ,YBP,u.1n'HC IRi
,.,'Ll "Lvr(XflP,y9P,3)

"Lur(.qP,nP,?) ,
'il. r,C",F(XTP,YTP,;:;.4,1HCIR)

: ',ii.

r RO'" OPE:~AT lNG'

'; c\",-Y,,/FOLS

Llr,E

" ',l u " II Tu I r X T , F YT , F X TP ,F Yr .. )

'~~~I~XTf.'-XTP).LT.~.l)GO TU 460
',I!
""';i:)I'IFlTP,FYTP'XRP,YRP,~)'

r r:.lj>:
CJT

,~.

trlf

THEORETICAL STAGES
\
..;)

"~u,IIXW),yaf.

. . . ):--:

.7,

~J of

'eSTEP

Ci ,';, ,
,

,"
,

Tu I X ( ,,) ,Y ( I ) ,X P ( I ) ;Y P ( I ) )

~:\',T

<:

I,\llE

186

I = ) '! ~ F' 0-

.'. '-'

::,:1

PLCT(XP(ll.Y?(I.I.21
r,c) TO 49 C!
,',,' ,'LCTIXi-'III.YP(I+1I.21
',,'TI:':'JE:
:' 1'< i I I LT. X Tl

. :- ; :.1 IE

'.",,~

~~~L

;-(;UILlRi~IUA

DI"G~'M:"

LcTTERI13.U.~.90.u.3.u.14.v.13nY

~~~IN ~HA~tl

C',L ~~THI3.15.14.6.0.8.9~.U.6HLPA~ENI
(-"Ll, '-,.' Til 13.15 .2J.l,.v.8. 9u. J .6H/{PAI'['~ 1
r~lL

LFT(F~ll1.u.5.0.U.d.v.9.u.ll~A

(tLL

~aTHIB.6.P~85.0.B.O.C.6HLPAR5~1

LI~

PHA~tl

(.', U',' a T H ( 1 '\ 1 A. 85.0.8. u. 0 6HI~PAKUII

~.'.:l;': rD=XTP+.'I."l."

'- .'.::)' i P =q (,

C'lL

LETTE~15.~.3.0.0.LABXTP.LARYT~.5HXT.YTI

i,:'" Xf'P =X ':lP+;;;]

I, :,'!Y'l" =Yt3" .. u. 6
LFTT[RI5.0.3.0.0.LA3XBP.LABYB".5HXB.YBI

r'~L

C~LL

LFTTER(~.~.3.u.Q.18.1.9.6.3Hl.OI

cnlL

LFTTFRll.).~.C.0.2.6.9.6.3HO.UI

( :' l.l

L F T T F R ( 3 t . . 3 O. D 2 6.24.8. '3H 1.01

.',,'IT" I'';ITIAL SUDIU'''I lUI> cv"CE.'lKAT]\J" i.JF EX"E"I"'EI'TA0ul>

r I, eLL <: TT E R ( 1-;11 ~ 5 0 a I, 5 .2 3 .. 5 1 HC i
:,LL'LETTERll.J.3.0.0.5.u.23.5.1HOI
r:,l_~_

:'~J\TH(5"1,?'3.5iJ,O . 5,v.U~~HEUUAL)

-:,

~U~RF~I~.n.23.5.0.5.a.G.SIC.6H(F5.411

.; F

r:F 5UL TS uF

"

EXPFlq,:E,'JT.\L I{U':

'!,:.l LCTT'I'12c).:;.5.0.0,3.0.7~5.29Hlu" FkACTIv"S

~ :..... L l_ r T T F!\ ( 2 '.) :, , u. 0 3: v .6 5 , 2l:i.X~~ J

VI>

SuDIW' OASIS)

.,: c,I,\THI4.0,f>.5 ,0.5 'C'O'5HEUUA~')

" : .. c :; U:< fif R I 4 ~ 1& 5 , 0 5 , 0 J X R 6 H F 5 3) 1
.' ',I lI'T TF R I Z ,j.~ (). 0 ,9. 5 .6.5 .2 H Y , )
"'cL ,';ftTH(ln.5.6.5,0.5.u.G~~HFUU
1

"'Li
.1

:"Ii""r~( 11.~1,6.5.i!.5.0.0,.YR.6H(F5.311

LCTT,RI?

... S,O.0.~.O.5.5.7HXT)

~~~T.I(~.n,5.~.r.5,c.Ot5HE~lJA~)

. -. L i~J::,(IFI, (4.',.5 .5.0.5'0.U.XT .6H( F5 .311

: :, l L LfTT[I~(2.U.5.U.O,9.5.5.5.?H,rl '
~

:,I,\THll','.5,5.5.0.5.(;.(;.5HE(.:UALI
I,

';l!~Hr~(11.J.5.5.u.5.U.U.YT.~H(F5.311
l~TrFI{(3,' .~9C.G,3.Q.4.5,3Hr~TS)

,'"

.-

','~

"

, ,,

;:,."'0;- H (' 1:) ~ fl.?, \J. 5,0. L), 5\...IEulU.o.L)

;;LJ:,'f1FI, (5. i
C;.O. 5, d. 'J .kNTS.6H (F5.1) )

.1,.

LrTTF;,(9,: .5.0.0.9.?4.5.9HPLATE EFFI

~'~THI14.U.4.5.0.5.U.O.~HEWUALI
,H I"

b 1'.;' ( 1 4 5 .4 .5 u 5 u c ." E F F .6 H ( F 6 .3 1 I

LrlTFI'(22'J.?O.J.~.v.j.?.22HA,iALY~I:' v~,

'--I"E ~'-v"'E~
::;THI14.j,3.5.Q.5.J.Q.5HEUUALI
::l;:-'nr,'( 11 5,3.5,;J.5.'J.:I./I,ULS"6H(F5.311
LCTTFIlR.O.5.0.0.3.G.Z,.5.18HFLU\~
L!J'E 5LUf'EI
':,,\TH( 1?O.2.5.0.C;,.l;.u.">,HEQUALI

v"

",- 187

~ :;.-

THE i'di',QFK l,F Ex .... r" 1I"E"TAL

;.: ( !t-J 1.,

'

.... "uC[SSED By THE P"U{jKAI"

~,)C"lrO.J,FF~-l.C

: :-1

~ !...~T (~"'.:...

~h.J,\\QfK

u~

to.G 9-1)

F6.11
.

,',.

'LOT(~.~,C.O,9991

"

,~',v,-,'TI'''f
.:\~.

If 1;'[

: :,-,; 'F ( 1 u.

~,.G:"E. A~UI'"
I, '

I,U'iPE" (X. Y .HE 1 GHT

TO E~COCE ~UMAEK~.

F"IT I

F'n. PCD I r.NUM

:~cL.SY~RUL(X~Y~HEIGHT.llC~.AN6LE,1~1

,.~UTl"F

STAT(X.~.XMEAN.XSTD)

":~JTl;'E

TO CI\LCULATE THE l'lEA,"

'''C',J1G,'';

'

..

A"D STMI/OAK!)

DEVIATlU~

uF X

X(41

,-,'

: ..' l=l.N
,.=',\.1+X (ll"X (1)
'. :: X '-~(J!:,+x ( t )
" T : .. Ut

-c',=X~U~/FLO~T(N)
" ' ; :,1;

..

,< T (

( ::, S~ -

x,..' F M', * x:: t A~. F liJ~ T ( "

CD TUT

I I / F Lli AT (I~-l I )

0549

,;>'

..

.. ,

"

188
':

SAMPLE COMPUTER OUTPUT

~1.

WIT~ A CJLJ~N OF FOU~ INCHES IN~IO~

T4E :'XC-i,HI:;E'J -i.lS,SIX 5P,:;::S SPACEO 4.5 IN~HES APA~T
.
,.
.
.
!"~~-=:~CHr:G:: SYSr::~! I'> SoflIUH IJN-'lOHEX-5~IIXI\ ~ESIN:-HYO~OGEII ION

li:"T=.-'.
T~'

I' ,~., ~:T'l::~

74

,J, #>.> . . . . . . . . .

"
~HE(HL/S::C)=

Ll1'JIJ -LOll
SJLI1

FLO\~

T'~!TIAL

.KATEIML/SO:::)=

SO'lTU'~

=r~AI;HT

wEr~

1.E>~:;

CONG.I~IEVHLI=

LENGTHIINCH)=

'c1=!~'--itIG'n:II'ICH)=
~~w~~aME~

4J.::CJ

L~T~~AL

"'~[~

PA~TIrL~

3~7'1

?,.~i:iJ

CLcA~ANCEII~CH)=

ill"~=:

.100

1.~47

I~CH)=

lISPLAC=:hENT(1/20
SIZ~I~=:S4)~

2.CQQ

-3].35
I '

~"~3-S=CTrO~AL

AREA(S~.CK.)=

7~.alC
I

~.~!l:znTION

AKE,lsa.::r.l= 55.C~7

"':,;~.~J.~C::;> CArOHI'I~ AK=:'\IS'l.~=

"'.\CTrON OF FLUIOIZATION
I":,crro'! OF OOWNCOME''!
'Lcll1!7HII)~J

ARE'A~

A'!~A=

TO.I)OW"~O~=:~

210.713

,. 9i!"

.31~

ADEA

~A

10= 2.229

""''','LL COLUHN,'SUPEI1FICLAL VEL.(~ IS!l.Ct',-SEC)=

':Jf1!7:AT-IO~1 SUPEKFI~It..V'VE'LI'1L-1S .r.~-SE'C)=

'.5Ql

.72&

-,.,"

I
,

'

"

,.

..

189

SETTLED
SED

C,,l.:UIOI7':O
'E"!J
":~~:

'lo:r';'lT (C'll

.-Jr~'iT

:;T" "lVN

(';JE: wO:lo

2.910

. 532

.429

~::!)

158.534

2Bo.4C7.

"':.1J I/QL(r::.C.)

":::': "L:Jt1FE'l

5.583

VOIOAGE=,.G31

HaGHT(CH~=

5.C!lG1

HYOROGEN(ME1/HLI
INITIAL
FINAL

C-ILJR IDE (ME:l/I1Ll

2."'''~
2.S"Q
,,2.635

I 'II TI ilL

FI~AL

G'B7
GB7
.Gllq
G67

.1)17

.119Q
. 11Q4

.1022
.1: 22

.G~'\

.Opl

FINAL

HJITIAL

-,'.'
. ':---, 2. 1',59'

C 00

. 118,.

.1189

GOO,
.COO

.OG93
.OG95
. OG97
.OG91\

.1:;20
,

.1192

.000

.OG9G

.OJ"3

.ilOO'7

.COO

, aoar{

.1al

.."-I [1/0<"

HYf'!~::JGEN

~L'0~

OF

FL~W oPE~ATrNG LIr::(LC/~l)=

~~~,::

o~ :~lLYSIS

I~ THE

SO')lU~;"

::Jpr~~TI~G

SO"lIUH ION

F~A~TICN

DN

cLO~~=

1.

O~9

. t(

LtNE((Y~-YTI/(X~-.TII=

IN LIQUID PHASE

:\~ = "-1.:" C

X (OUT 1=.231
Y( OU Tl

!l '; 1'"-,, [ ~ !

,: .. !.

r:: EFFlr:l nlCY

:'Ll,::

H~TS(~HI=

(PERr:E~Tl

::-FFT~IEN~Y(l'~~I=

.9G3

!(~' TH": S11IUH PN F~A~TICN IN ~::SIN .PHASE

TO

.GOO

7~ 1
2~.143

'" 36. 'HZ

~4.3'9

..

r
, ~.

.,
lJT (~~ClLS::r;)

-1,,'11

(1;::/~;-,";l

.... .
:. 1'; ( I .; - 1 UT)I 10 N
f

( ? :: ~ ::-i ~J T )

f'L:iJl11zn TIO'I:
JOS1N

FLOHt

J~r~~LL

'+.23&
~.')G~

-.~-,,?

I tl

21 77~

OPERAT10~1

n'rcnT10N.Tt~~t

..
flJir::

~UT

S~OOT'1

F~O~.~ATE

HftS

SLIGHT PERIODIC

r,oon-EXGELL~~T ~

24 'tIN.

_.

? 36 .

r:OMPLETE AND
~ONT1NUOUS
VA~IAT10'J

190

'.

J,I

,
.'
I

,.,

,-,

191

-I'

Co= 81.020

'

: ' i
k. '

\...

~'J

'C
,

('

>

.
,"

G.O

--.I

1.0 ' '

XT,YT

X(LIQ PHRSE)

'.

~~N)FRRCTION~ ON SODIUM BRSIS

A3-1.000

XT= , .231
.
,

'\

YB- .741
YT= .001

. P L!=l TEE F F= 3 8 .. 8 12
:j[\jqL YSIS OP LINE SLOPE= .962
!' LOW
0 P LIN E S LOP E= 11\ 9 6 3
i iTS 2 II 3 2 9

'"'\

'

'

'-

192 .

'.
. '. .

APPENI;>IX H

(
ComputorProgram for Fractlorlll Factorial
. Exporimontol DoSign

0-:

'.

"
."

...

(START)
'

LREAD YIELD/

I CALCULATE TOTAlEFFEC~
1CALC.

AVERAGE EFFECTS 1

1CALC.

S. S.,'

ARRANGE ASS. AVG. EFFECTS IN

ASCENDING ORDER OF MAGNI TU DE
r-~i
:
1 STANDARDIZE AVG.EFFECTSI
I

ml -

ARRANGE STAND. ABS. AVG. EFFECTS IN

ASCENDINGl ORDER OF MAGNITUDE

~:

.-- -.,..----------,

~t

I. ;

- I CALC. RE'SIDUAL MEAN S.SI

IN
1"

ten(/)

1CALC. MEAN S.S. 1

Ir-::C'-:-A"-UC=-.-:-M=E-:-AN~S"':::.S-:.R~A-=J::-::IO~S:-1J
,

:~

\ ... ______ ...: ____ .:- ___~------,---- ___ __ 1

I

/PRINT RESULTS/

~
( STOP)

FIGURl
H.I
,

Flow chart of the corr puter progrom for o""I.s;"9

the results of feictoricl expaimcntol desigr

.r

COMPUTER PROGRAM USTING

194

,,',"1:C,~L

PROGRM.' FOR Ar,ALY!>ING FRACTlONALFACTORIAL t.XPe.RIMENTAL

'.:_; -, _ Tv II"Y ,~(j"IBER vF FACT";":; ,11"0 "'Et<F':,"",I.'G THE A"AL..YSIS vF
"

;:, .-E(ANCV)
r,-,~,~,.'II'-' IS

DIVIDED I,qv TH"EE :'"U''CTIV"AL.."A"TS.

THE "'ASS ,'u'at.r(
,_ ;,_:.' -;;'TA DECK Cl.iNTO<L.L.. TIlE Sf:CTl"' .... vF THE' f'"uGt<A'" Tv BE uSE".
:" '~Ll_:;.'; 1:.6 RESJLTS ,\"E USIA I"ED FKu'" THE f'tlS,=> r'U~iBEO<::, .. HE"
,~. :~TE D;.TA IS P~0VIDE9 1'1 THE OIlTA UEC~,
'=:.
y.\,"S f,NALYSIS A .... O
_
~v~~~GE EFFECTS LJSTL~ IN ~SC~N~Ih~ Uk~ER
YAT.;:'S ,'\:"\LYSI~,
:
t
,FFLU!> ,l.~l.
STII"::lAO<DIZED !:.FFLCTS L..J!>TED J",ASCE".;I"G "'''DEr(
- : :='" ,~ YATES iI"ALYSIS.
AVERAGE EFFECTS ANO
ST .D,\RDIZED tFFtCTS LIStED'II'; ASCEI'.. {)J"C, ur(UEt<. AhD
OF VAkIA~CE IAHLE
, 1::-
LA'I1
~:;;"-"~R

..

, .= .'

'=,,',,:";,'1 C0DF(100).Ylf L D(10Q).SYIELD(100 l .EFFECI(10.0 l

AVGEFF (luO') .5::'( 1;)0) .SCUDE (llJu) .:'EFF (lUlJ)

,'c';:,!Oll FACTOR( 160) .(",r-1(8)

.. , -- c';:, ! Ui'< flu S J ( 5 () ) NPUS I ( 50 ) .'-: SS ( 100 ) .~SSr( ( 100 ) .0 F ( 100 )

~:"='::,!ON

C'/ALUEI,)

; ' ,:,:-(.~? 'A,~'C

" :.:....

:,SS,

~SSR

~'. THE PASS ~U~clER

-t"L)(,.740)f'ASS
,':.AT (FlO.:))

'i:.~ FeR TrlF 1I~IILYSIs ~Y YATES TECrll';IUUE

.... ) ';'J~-5rK UF

Ft,CTOkS

I;~ FAC'TGr(IAL DESIGr~

.:: : .;ut:
,,;(".IO)RK
, ,'.\; ( FlO. C )
; =c, F ( ~ ) 199 Q
. '; ~ :-.! !!:
-:.1(-~K+C..v1

c.

irlE

1 ~~

6~ACTIUN UF THe. FHACTIU~AL FACTUKIAL OESIG~ (EXPr(ESSED

, (-~)l.

I.E. REAC M

:. ~ : ') ,Ie ) ~ ~~
; T ( i-~ ~.: + C

;:.:;

~:

1)

"-,

io.14~'(FnCTOR(I).1=1.C'
\:.T{A:I}I..o)

"

'riE TY~E DF vlELU IN

. ~ ('; '.?V:) (Cv.-I ( I) .1 =1.8)
... i ( 8 ,\ I ',' )

THE uEslGN

: c

:'1:: SiZe

.:- : :' . ..) *. Rt<. ) I

.-.

UF --HIE FRACTlO:'-lAL FA,CTURIAL LJEsIG'"

(

i ( ',~ + 0 'j! )

? u *R~")
,,'

195

., . T ( .' : '

L -

J. ... "

. : ' ,. ,'.' ' 1 = Y 1 C L D ( I I

',"-\
J.

"

J~l,,":...

{,

,,' 7. I=!.P.?
. , f ( T ( 1 + 1 !? I = Y I E L [) ( 1+1 1+ y I E L D'( I I
, r F "( T ( \+ 1\ / ;: + I /2 I =':( I ELD ( I + 1 ) 1: YI EL D(
~" ~. T I t~iJF
1"1.1.[8.,\1(,(: TO II:!
.,~'o

I)

L=l,t~

V"LD(LI=EFFFCT(L)
CL,':T

1:'j~E

,,'';T I:'t.;l
[. <I'

T I :.~ LJ!:

EFFECTS
':'1=?t;*(PK-RM-l.01
:'~

I = 1 ,r\

~.J .J

~i~F(; I=E~FECT(.I

CJL~T~

'=7.

S~';5

UF

)/DIVl
~UU~K~~

'*(,,~-R:-',)

.: 1 ), = ( t, r. ~. (

F F <;C T ( I I I ) 2 CJ I [) I V2

, : I';' JE
.. '. ; !

TIT L E

:rE{fH:'C)

'-',~T(lHl ...

YATE5 TtCH,il~uE Fu~ I~E A~ALYSIS UF FKACTlu,'AL FACTu

EXi-'Ei<I"'E"TAL UESIG,'*)

::''-'. /1 . '
:.-T=! . l/( ?~~*Rr-~)

:rr(A.l?~)K.FRACT

" ':,T ( I I . " NU;':f7ER OF FACT01<S IN DESIG"~T

, F~ACT!O~ OF FACTbRIAL DESIGN=*. FS.3'
: r [ (f., .. 2 ~ 1,1 )

FACTORS9.)
!(FACTUR(I).I,=I.CI

12

':'AT(/I ....
::E(~.ll.

. ",,o,T(I.

r,~.

Ao\lCn

'TrIA.?! 1(~OM(I).I='.S)
:::T(// YIELD9 *. I I . 5X. RAID)
'r~

t\fJ\JLTS

'iTt'(6'1~r) .,
i "',;,1 ( I I . ' I.X.

_",<f.r,"-. ') ... ~

*T"EAT''E"T*. 5X. *Ylt:.'-D.... 6l<. "TuTAl."'. 6X.

*SU~: ,UF*' I . 3;<. "C""'BI"ATI~I'*' 15X. "EFFECT*. 5X.

196

.,.ecC,'''.

hX.

*SOUARE*)

3;<',

.SY[ELI~I 1) .rFFECTI 1)
Alv~FlC.~'F12.3)

:I, i'.2? .. reOO[1 1)

:.i(/,
~

: (, ~ ! = 2: ,N
: c , '" , 7 If) ev D E
~l

jX.

:/.

( [ ) S YI E L

r. I [

J ." F F E C T I

erC T~E ~~SOLUTl vAlUtS OF THE

. ,.
I : : ? , ;-~
'. :',T, ( I ) =1'."'5<1 AVG.FF I I ) )
. " ~ ,,, I 1 ) = COD F I I )
...

[ J t. vc, r F F I [ J .55 ( I )

Al~.Fl~.~.Fi~.j.Fll.~.Fi~.~J

Av~RAGt

=FF~(TS

ANu STuR~ C0u~S

~-

- :~,". r Jr!L'::

-.......r

.. , --'.',-'.

I
I

I =?

c, I."
.. \

-:

~'.:
,

I:H'I:CTS

I NG OOLJ':R

1+1

J=;,:?N
lcl.,VGFFFIII.LE.AVGEFF(J))GO TO 560
, TJ'E 1 =SCOOE ( II
. i ,,",E2=AVGEFF I II
( . ! ~ E I I ) =S COD E I J )

:" -."FF I [) =f.VGFFF (J I

~ ~ "Ol[ I J ) = S TOR E 1
c,f, .'

.'

C,:)1'; T I iJuE

- ',1 P:iJE
;. 1~~SS.GT.l.~)GO TO

i:

ASC~r,u

.~, 1

:~~"FFIJ1=STORF?

=<,

IN

66~

:.T THE .\';SOLUTL AVERAGE t.Fh:.CTS [1'1 ASCEM)ING vRui:::r<

:;;'1',.6;>--:1
ATII/II. e AVEHAGE EFFECTS IN ASCEi'DI~d Ur<~EH FUN HALF ~UHHAL
" ,]"./1/. ,X. *ORDER". 5X. "E'.FFECT*. 61-. "AVERAGE EFFECT*. I.
" 1 ,\P:,GLLlTf)* I
">.

I::::?,~!

:;f( 64~)I-l.SCOOEI[).~VG[FF(I)

i 1:.1 .IF
:IIl:
. :,' I

1'"

f, C;

O;,!lER UF

THI:. EFFECT \~HICH IS USED FuR SlA,'1DA HOIZATlOI"

v ) Il Nul'

.,

",\T IFIlJ.v)
.. I ; . T ( 1~ ,' ' ; ~ ; ',+ r :. J 1 ) + 1

'~_~It Th~ AbsOLuTt ~IANDAHUllI:U EFF~tTS

:,1:1
I =? ,,'J
ell ) = r. VG t: F F I I ) I ~V G F F F ( IlU "1 )
.,

r,,~ ,WE.'o\{'E EFrEClS 4~D THE STA'~DAKOIZED EFFECTS II~ ASC~I"IlI"G

. 0

:'.~.::;;~;;:*

AVt:.:RAGt. EFf't.CTS ANU

STANUAR\jlli:.~ t:.Ff~CTS"

ARt. ARKANGt.

.:. II ASCENDING ORUI:~.FOR HALF NORMAL ANU STANUA~ulltD HALF

. f.L'. II. " i-'l...vTS o{ESi-'ECTlVELY*1
,

,.
, :.,.

- ( f, , ;

.,~. :.;1//,

-'X,

" I' ," f, ;."

.. ,

197

P ,: )

! -;:?

if

*";"DE", 5x, "EFFECT", 7X,. "AASU~l..ilE", 71.., *ABSI."-UT

AVG F F F

*,

8 x,

~.... 'J G'" )

.*!:l T.)

,f\;

! ; c , ' " sq.) i-I ~CGDE I I ) Wr,EFF ( I ) SEFF I I )

: " :.: II, 16,'>X."1,,.Fll.3,Fl?31
".
.,c
. . J'_
~.

( .' :, [,:,. r~.:. -;; ~

1GO

T C 3 G~

'0:':'

T.,E i,lJ,',:;ER OF EFFt:( TS f'uuLt:D

'l N\!
.= .~ I: '.' :. T ( F 1 ~ ~. ))
.. .c I:, T'I R':N+r. \1} )

TIJGETHEK ,\'::> THt. HE'::>IlJUhL IIAK[Ar.CE

;, t \ .' I 'j , ., 1 G )

.,c~n r~F fJGSITIU~S OF TH0SE EFFECTS

,.,'; t. n IS. 32 \.' ) I PO'S I ( I I , 1=1 ,"N I
r J I,," f, T I 9 F 1 c: ,; I
CJ I,'J 1=1./';N
,~CSIIII=INT(POSI(II+u.U1I

,'; '~I) CRT ICAL

VALUES I" THE ORClER lit 0.9G. 0.95. AI'L> 0.99 LEIIEL OF
(.':"lnH!CE
I: \r)1 5.47" I I CV,\LUE (11,1=1,3)
C :,-:.\

T I ., F lL: ;., I

::L(UL~TE
~

'\

..,.;

~ESIDU~L

THE

UF SUUAKE,::>

MEAN SU~

f'\

1=1 ,fIr!

=lSIII)

.,

',.

'<,IJ~:+SS(KK)

.'I'\Ut:
1:'+ 1 ) = SUI:,I ,~i,:~
. ;. 1':+ 1 ) =,~Nr,

..

j~

'.(U.ATF TH" !'JF:AN SU"

OF ,::>UUARES

'h~) I=l,N
.>I!I=SSII)
... 1 I ) = 1 I.
~

:r r/"ut.

cr,JetTE THE .-HoM' SUi. uF :"~UAI<ES I<ATllJ

.::' .. l=l,~~
., I I ) = 5 r. It) I ,-:55 I Ii. 1 )

:' r

f, F S tJL T '::>

! r r (I:., I. A .; )
: ~.:~ll/III,

..
~ A"ALY-5IS uF 1Ii\I\IA"CE TABLI:\AI~ulIl*1

:TFI".4?d
.. .\TI//. <;x. *FFF!'"CT*, 7X~ *SUI~ UF*, 4X, *DEG. UF*' 4X,
.'!.'- ')lJl-.*. 4X. *,IEA.; S"'UA~E", /, laX, *S"'UAI<E*,
,c~FDO~*, 4X, *OF ~QuARE*. 6X, *RATIO*I

198
I: / , ",

,c
."l~

_:=1.,'":

-.:.,~ ..... ,.P(JSI(J))GO TLl

l'

3'BC

! ~~ E

TU <;l,~
, ," I : ) r ~ (V II LUF I 1 I ' ,In. "'S sn. I I I LT. (' VA'-U" I 2 ) J Vq T F. I 6 ,480 I
,1 J ,S':;( i I ,"'~( II ,~'15::;1 1 J ,:l;;SR( I J
'",(/, ;:" ,;1~,Fll.3,F9.1,Fl".3,fl~.1, 2n . '
, ~." 1 J .~. r (' Vt,'- L.E ( Z ) :,:"D "'5S r , ( I J L. T (' v AL",E ( 3 ) I "K I TE ( in 490 I
- II) ,SS: I J ,DFI I J ,MSSI; I "',!:,5KI I J
~
,
oTI:, 3:, AIQ,Fll.3,F9.1,F!4.3,F14.3, 3H ,,-,
" \ ,~, S" ( I l' r r (' VII LljE ( 3 I ) ..;~ I T E ( 6 , ') 00 I C.... DE ( I 1 , S 5 I I , , OF ( I I , ",55 '( I , ,
C"(!I.L.T.(,V~LUFIl))(',O

~ I : I
,_."~T(/,

T1

3A,

"
AIJ'Fll.3,F9.1,F14.3,Fl~.),

4H "".)

~'l

,T PHiF
.. '" TE (f>,3"~
;wiC~T(!,

l('-'DE( I I ,sst I) ,DF (I' ""::'S( I) ,,'15S K ( I 1

3A, Alu.Fll.3.F9.1.F14.3.F14.3 1

:, ',7 HIUE
-'" 7() ~6Q

,-,rI'\uE
':T~(A.4~'lrOnF(1 ),SS(I).DF(II
;,i'AT(!, 3A, Al""Fll.3.F9.11

:", T I c!,uE
:T~(6'41wISUM'DF(N+l).N5SIN+l)

~T(!. 3~. *RESIDUAL*. F13.3,F9.1.F14.3)

, " ; : :: I 6,44(: ) .\ CVALUE ( 1 ) ,'" CVALUE ( 2 ) .1" (VALUE I ? 1
-,'~TI/I,. r~ITICAL VA'-UE~9~. I I . 3X. *FIl., 11. *.O.901=*.,F5.2
',' . '1"11,1,:, 11. ",:;.951=*. F5.2, 1X. "FIl.*, 11. *.0.991:*. F5.2 1
T

":'()
-

~ll~

1?(jt'

CD TLIT

0263

..

_.


199

.,
SAMPLE COMPUTER OUTPUT
F>".Tnu~

,:1,;

T'-I c

""Cl~

HLlLVSIS OF "~~CTIONAL FlICTO~IAL

'Y': "p':~r ~L- '1<::<;r"~J

,,=

\1' :;)

r,\,~T'll;><;

f'\EST~N:

IN

"~CT:)o.IAL

"","nO'J 1F

nESTr.'I: .25:

;,l:rr,'SI
3=~~5IN

~lITE

"LOW

1=LI1'JIl FLOil o.ATC:

,.=~r3!'J
~=Htn

D~RTI(;LE

'-IEIGHT

~=lIJ~E

OISPL~C~~ENT

LftTERlIL

F=~OYNCQHFq
:,='i~V

SIE

CLEAo.ANCE

Lr"'GTY

Y~:L ;t

PLATE EFFICIENCY
P'::,U'FNT
T ID~~

TflT AL
EFFFCT.

VI ELD

~Ov'1 p:~

,
(

:, )

'.

SU>I OF
SQUARE

26.34:

727.57<'

1~.952

-?;.72~

-1.545

100.543

-13.r,iiJ

-.1154

5.831

~'"

) ~

EFFECT

AI/E~AGE

,--

21 131',

<J.S3e

.59&

2.838

( -; 1 :-

, 'l.335

-1').5B">

-1.912

29.234

(",) 1(')::

2"1.!tC~

-1.272

-.OBO

.051

( , I: ( '1) =

27.

hz

a.75:

.547

2.3<J3

1. It 77

17.440

B.J7~

23.&2'+'

2g.75~

10.117:'

.&80

3.&9&

5U.314

30. 39~

1.'3CO

28.87&

41.55:

-.&35

-.01t0

31,.32'.

-&&.53~

-1t.159

01
138.353

1f>.&21

-51.762

-3.235

83.728

( )

( ',) Il)
'( -I)
~

f '1)

( . ) 'I r; ( 1)

r)

c,

200

1- -

2,. ?Gil

-26.79:'

-1.;;75

22.,.311

..

22 ... 73

8.9~:'

.5~2

2.523

I'I~"~=

17.;;LI,

17.h8~

1. 1G &

9.777

~,

2').353

-1111.2?'

1G20.2"l"

-79.004

: ("') -,

lS.fl9"
? 7. ~ .. ,)

-11. 32 9
-4.975

15'.29~

.950

Q25
7.31Q

("ll:ClIG

18.37S

27.952

1.7 .. 7

2 ..... 10

10.56&
8.503"

2U.C"~

1.253

1t2.50D

1 ... 190

.887

0.292

( . '. 1 ( "')
:1 ( ~) ,..,

(~) ~

:-::::C;

.....,.. ,..
J:_ v

1,98"

-5.775

-.301

1.0,.3

(~lcCc:;

10.11
8 ... 8&

-24.95~

-1.500

1'l.45'l

17.2'la

~23.32:;

-1."58

17.003

t,

'l.2,.2

-0."8~

-.4D5

1.315

('~)S~G

q.78 ..

-g.300.

-.5115

2.7U

It. 775

182.405

( ~ ) ~ I='"

'1.91,2

7E, ... no

21',.587

2C2.1"~

12.034

1270.8'l3

of:;

14.6&5

4.582

.2116

.656

( tiS ( :ll "F"

23.129

.032

:l:U:'lF=--;

i''l.R37

1G .120
5' ..,,3'1

3.200
.8g0

1-" C -,

(llcFG

..,

(~H ([11

/'

334

IJ:'~~E EFFESTS A~O STANOARDIlEO EF~!:TS ARE ARRA~GED IN

l)~= ':1 PI:;

HALF
FOR

O~r)rQ

'10~~'AL

A'IO ST AlmARDIZED HALF NORMAL

'l~"r" ~c-~E-TrV;LY
J.J

._

(AI'1I'1IE
IAI'3('1IEFG

.G8e

A'lSOLUTE
STn AVG
.e23
.(46

.28&

.164

flr;11I;:F'";

.33 ..

c;.r;

.3&1

EFFECT

C(DI"

AOSOLUTE
AVG EFF
.~ .. C

-'"

.lg1
.207

201

~
Il) 8FG

..,.

.ftOS

.232

CAIC!'ll::

.547

.313

'.
r."""

.<;&2

.321

C~Ir.FG

.585

.335

B:;

.59,

.141

.&32

.3&2

:2

CAI r. (I'll EFG

c.,) C'lIF

.68C

.389

: 3

(AIC

.1\54

.IoS9,

! !..

!EG
C(OIG
( ~ I RCEF .'

.887

.508

.q5&

.547

1.H&

r, 33

(AIEG

1.253

.717

1~

F:,

1.458

.1135

19

BCCDIE

,1.477

.845

7-

(AIf\CEG

1. SOC

.11')3-

?1

g~"

77

C\i,nC.IOIG
A(-l'IF

,
"1

.'."

1 -

23

. 1.&75

.'l5'l

1.747

1.J Q-C

1.S88

?~

I'll::

lI'l~;( .
1.q12

?~

( 1\ I

3.235

~1I52

3.545

.2.G2'1

'4.159

2.380

EF

(AI?

')

(AlaClnlF

?~

9r.FG

4.775

2.733

?1

BIJIG

4.'1 7 5

2.1\46

(~)(!lIG

11.326

&.463

('lIFFr.

12.&34

7.232

,...

"1

.>

stii1

OF

S'lllA~E

!lEG. 0"

FREEOO~'

1. ~'l4

?7

":

!'lEAN SUIi
F S'lUARIO

!'IEA"I SQUARE
~ATIO

\
202

I '.) 1

1[0.5"3

.. ) '

5.~31

1. C

2.8311

r:'

2'l.23'+

( ,) C

051
2.393

(.:)"(11"
(.:)

( '1) ;:

1:(1)=

17 .'+4t'

('" H.1)

3."'%
28.1176

3 (.') ..
~ (J)

( n) ; ,

10C.5'+3
5.1131

'l.8e4

2."3~

1.(

. 29.234

.277
2.851

1. : .

1.e
1.a
1l
1. C

.051
2.3':13
17.4,.0

1.70 1

3.&9&

.3&u'

28.876

2.816
.001

.013

lJll.353
33.728 -.

1.e

1311.353
83.728

13.'+91

100.072

22.'+38

. 1. C

2.523

'1. S

('" 1 'II:~"

'l.777

1.(

(.:)(,)';

1J2&. 2~

1.e
1.e

lJ2&.2~}(

l.t

7.311l
2,.. ,.16 .

F'
"'"k

1 ( '1) ...,

lq~.025

::: ( '1) ,-,

7.310
24 ... 1fi

(.:ll,~!lIG

:. ) : -,

12.560

...

. '
--,
,
.):-

1'T 1. _

"'.292
1.1l 4 )

l.i:

(.:)):~-;

1'l ... 59

1 C'

::'

17.003

1 .::

~,

( :. 1 ! r-

.005
.233

1.0
l.e

....
8.16" ....

......
19.309 ......
'-..-:

19!1.0~~.

.713

12.5&0

2.a81
1.225

17.003

1.658

l.~

.-~

1.ll5

1. C

( :,) . c c,

2.7'+1

1.;:

1~2."J5

1.e

1271) .1\93

1 .:

162.,.05
127&.'193

.&5f>

1.e

.&5&

12".508
.06,.

1.2 .. (

1. :

3.200

,.312

.1I9C

1. "

.')90

.0117

.,

]"'

( "I

'

'G

.('I'(,I)=:F~

('.I'(1)::Fr,
\~

( ') .=;--;

..

....

.5&'1

1113

(.I)IS('))F

1 :

1.315
2.7"1 '

.128
2&7
1.7.78&

....

~'

:
0

.~

,.

-'~
r

202

r
( ~)1
(

.. )

..

"

HO.5'+3
5.!!31

1.~

1. C

10C.5r.3
5,'.1131
2. ,~3'\

9.8er.
50'!

?9.23'+

~:
1

29.23'+

2.851

(:)~(~)o

.051

1.:

G51

.005

(,~)~('1)::

2.393

1. ~

2.393

.233'

17.'+'+(;

1.0

17.r.1t0

1.701

3.1;'l6

1.(

3.&9&

.3&il

28.!!7&

1.e

28.87&

2.81&

.ij13

1.0

.013

.001

. 138.353

1.e

138.353

13.'91

33.728

1.G

83.7211

8.1&,+

F"

,22..38

1.e

:EF"

2.523

1 "s

'lI:,F

9.777

1.t

(.~)())';

1J?6.29'+

U26.291t

100.072

~-:-

( :) ;

, : (1)

_I.':)(J)~

3 (1) "
: ( ) ) ;')

D),S(,))F
(A) ;C:f'

(~.)

1 (i) '-,

:r;),-,
(:)1~(1)G
; ~

( " )

. ..

.,:-

2.8311

19~.O25

1.e

1.e

.277

1911.025'

19.309

1.~

7.31Q

.713

2'+.'+lfi

2'+.lt1&

2.381

12.561)

12.5&0

1 . 2?5

7. 31 I!--

1;.292

- 1 (,

1.043

l.~

19.'+59

1.(

17.'003

f. ~

17.003

1.&58

c::- .-,

1.:t15

I.e

1.315

.126

( :.) :;:- r,

?7'+1

1.~

2.7'+1

.2&7

1~2.1t~5

1.e

182.'+05

1271).693

1 .:

lZ7&.~93

'

.~

( .: ) -J:::-;

..

(: )

"j

r-- -

"

17.78&
1210.508

r
r ~

('. I"

(J) ~F-;

.&5&

1.(

.&56

.0610

('.)-(l):Fr.

1.2';(

1. :

3.200

.312

1.l

.S90

.0117

{'"'II :::;--;

.1191:,

,
,

...
......
......

( ")-

l~

......

. ...

203

;.,r~.'L

"\_U~<;I

Ilf,~,J .l~)=

3.7'\

..

"