Mathematical Models of Mixing With Applications of

Viscosity and Load Capacities

A Final Paper

Presented to the

School of General Engineering

Kennedy-Western University

In Partial Fulfillment

Of the Requirements for the Degree of

Bachelor of Science in

General Engineering

Herbert Norman Sr.

Arvada, Colorado
 TABLE OF CONTENTS

CHAPTER 1 INTRODUCTION …………………………………… 1

Statement of the Problem …………………………. 1

Purpose of the Study ………………………………. 2

Importance of the Study …………………………… 3

Scope of the Study …………………………………. 3

Rationale of the Study ……………………………… 4

Definition of Terms …………………………………. 6

Overview of the Study ……………………………… 8

CHAPTER 2 REVIEW OF RELATED LITERATURE …………… 19

Solvents, Oils, Resins & Driers ……………………. 19

Introduction to Paint Chemistry ……………………. 22

Viscosity & Flow Measurement ……………………. 26

Paint Flow & Pigment Dispersion …………………. 29

Printing & Litho Inks …………………..…………….. 34

Physical Chemistry (Suspensions) ……………….. 36

Fluid Mechanics & Hydraulics …………………..…. 37

Chemical Engineering Calculations ………………. 40

Ordinary Differential Equations ……………………. 42

Geometric Series Application …..………………… 49

CHAPTER 3 METHODOLOGY …………………………………… 51

Approach …………………………………………….. 51
 TABLE OF CONTENTS

Data Gathering Method …………………………….. 52

Database of Study …………………………………... 53

Validity of Data ………………………………………. 53

Originality and Limitation of Data ………………….. 54

Summary …………………………………………….. 54

CHAPTER 4 DATA ANALYSIS …………………………………… 56

The Observed Flush Process ……………………... 56

Treatment-I Model A ……………………………….. 58

Treatment-II Model B ……………………………….. 63

Treatment-III Model C ……..……………………….. 69

CHAPTER 5 SUMMARY AND CONCLUSIONS ……………….. 80

BIBLIOGRAPHY ……………………………………………………….. 86

APPENDICES ………………………………………………………….. AB

Flush Formulae ……………………………………... A1

BASIC Program Code (Model-A) …………………. A4

BASIC Program Code (Model-B) …………………. A6

BASIC Program Code (Model-C) …………………. A8

BASIC Program Reports (Model-A) ………………. A11

BASIC Program Reports (Model-B) ………………. A12

BASIC Program Reports (Model-C) ………………. A13

MathCAD (Model-A) ……………….………………. A14

 TABLE OF CONTENTS

MathCAD (Model-B) ……………….………………. A16

MathCAD (Model-C1) …………….………………. A18

MathCAD (Model-C2) …………….………………. A20

Flush Formulae Derivations ………………………. B1

 ABSTRACT

Mathematical Models of Mixing With Applications of

Viscosity and Load Capacities

By

Herbert Norman Sr.

Kennedy-Western University

This is a mathematical algorithm that approximates the total number of

mixing stages (n) required to process optimum amounts of reactants (varnish &

aqueous pigment) in a mixing vessel of fixed capacity (B). In some procedures,

the reactant amounts are calculated in increments (i) by the algorithm to insure

efficient use of the mixer’s capacity, while adhering to a uniform viscosity function

[ή(i) ] for the product.

The viscosity function defines how the paste will thicken over several unit-

flushing stages, 1 ≤ j ≤ n . The distributions can be defined by mathematical

functions or can be manually induced after being determined experimentally. In

each stage of mixing, at least one of the added reactants is a calculated charge

of vehicle (resin, solvent or varnish) or a charge of organic pigment presscake.

The presscake has the physical properties of pigment suspended in water.

The two reactants (presscake and vehicle) will first form a slurry, in which

all of the water, pigment and varnish are suspended. Then the pigment and

varnish will start to adhere to each other, forming a sticky mass in a watery

i
 ABSTRACT

environment, thus displacing the water molecules in the aqueous pigment slurry.

The resin and solvent (varnish) particles are more attracted to the pigment

particles than the water, thus wetting the pigment and displacing the water in an

environment where the vehicle the vehicle-to-pigment ratio is greater than one.

The displaced water can be extracted from the system by means of pour-off and

vacuum. The complete process is known as flushing.

The initial objective of this research is to develop general mathematical

models, which will simulate the observed optimized flushing procedures. Given a

minimum of input parameters, the model calculates the flush output parameters

such as the increments of pigment and vehicle charges as generated by the

viscosity distribution function. The results of the research for this thesis led to the

development of three models, which are referred to as Treatments I, II and III. All

three of the models produce feasible outputs, some of which were verified by

processes used on actual manufacturing work orders. Since the simulations are

math models, the procedures can be programmed on a computer. In this thesis,

all source code for programs will be provided and written in QuickBasic. The

procedures will also be modeled in MathCad worksheets.

Treatment-I requires initial amounts of pigment and vehicle to be charged to the

mixer. The model calculates the amounts of pigment and vehicle charges that

are required for each mixing stage so that the sum of the increment charges will

equal the optimized total charge. In other words, this model distributes the total

charge to agree with the given viscosity distribution. Optimization is the primary

ii
 ABSTRACT

focus of this treatment while adhering to a given viscosity distribution and holding

the mixer capacity constant. The calculated capacity, B(i), is an output parameter

and will be listed at each mixing stage to compare to the constant capacity, B.

The input parameter, E0 (Allowance), is the estimated % of the constant capacity.

Theoretically, E0 is equal to the water displacement in the final mixing stage.

INPUT DATA OUTPUT DATA

Capacity Constant B Calculated Capacity at stage (i) B(i).

Initial Pigment Charge SP(i) Number of mixing stages (n)

Initial Vehicle Charge SV(i) % Vehicle after last stage (xn)

Relative Viscosity of the Pigment (h

h p) System Viscosity Constant (kv)

Relative Viscosity of the Vehicle (h

hv) Viscosity Distribution h(i)

% Solids of Presscake (r) % Pigment per stage xp(i)

Viscosity Distribution Function f(i) % Vehicle per stage xv(i)

Allowance E0 Pigment Charge per stage P(i)

Vehicle Charge per stage V(i)

Water Displacement per stage wd(i)

Total Pigment Charge SP(i)

Total Vehicle Charge SV(i)

iii
 ABSTRACT

Treatment-II requires (xp), the % pigment in the total mix, as an input parameter.

This parameter along with the capacity, B, is used to calculate the initial pigment

and vehicle charges, which are required as input parameters in Treatment-I. The

remaining steps of the procedure and the objectives are identical to Treatment-I.

The model uses the mixer’s capacity along with the viscosity distribution as the

critical input parameters to optimize the loading of each mixing stage and

optimize the yield. The total amount of pigment and vehicle required to charge

the mixer is an output parameter in this procedure.

INPUT DATA OUTPUT DATA

Mixer Capacity (B) Number of mixing stages (n)

% Pigment after last stage (xp) % Vehicle after last stage (xn)

Relative Viscosity of the Pigment (h

h p) System Viscosity Constant (kv)

Relative Viscosity of the Vehicle (h

hv) Viscosity Distribution h(i)

% Solids of Presscake (r) % Pigment per stage xp(i)

Viscosity Distribution Function f(i) % Vehicle per stage xv(i)

Allowance E0 Pigment Charge per stage P(i)

Vehicle Charge per stage V(i)

Water Displacement per stage wd(i)

Total Pigment Charge SP(i)

Total Vehicle Charge SV(i)

Calculated Capacity at stage (i) B(i)

iv
 ABSTRACT

Treatment-III uses the input parameter, Total Pigment Charge SP(i), to create

the pigment distribution, P(i). In this model, the pigment distribution is a

geometric progression, whose sum is equal to the input total pigment charge,

SP(i). The number of terms in the geometric progression, (n), is treated as the

number of mixing stages in the flush procedure. The viscosity distribution is an

output parameter based on the actual % pigment, xp(i), calculated at each

incremental stage (i). The mixer capacity, B, is held constant through out the

procedure. The calculated capacity B(i), is is an output parameter and will be

listed at each mixing stage to compare to the constant capacity, B. In this

treatment, the allowance, E0, is not required or used.

INPUT DATA OUTPUT DATA

Total Pigment Charge SP(i) Number of mixing stages (n)

Total Vehicle Charge SV(i) % Vehicle after last stage (xn)

Relative Viscosity of the Pigment (h

h p) System Viscosity Constant (kv)

Relative Viscosity of the Vehicle (h

hv) Viscosity Distribution h (i)

% Solids of Presscake (r) % Pigment per stage xp(i)

Pigment Distribution Function f(i) % Vehicle per stage xv(i)

In a Geometric Progression model, Pigment Charge per stage P(i)

Capacity B(i) is Constant for all Vehicle Charge per stage V(i)

stages. (1 < i < n) Water Displacement per stage wd(i)

Total Pigment Charge SP(i)

Total Vehicle Charge SV(i)

v
 ABSTRACT

Observed Process Reaction Per Mixing Stage (All Treatments)

A given amount of presscake, PW, is mixed with a given amount of vehicle,

V, to produce a paste, PV (wetted pigment) and displaced water, W.

Formula : PW + V → PV + W

PW ……………………. Aqueous Pigment (Presscake)

W ……………………… Displaced Water

V ………………………. Resin or Resin Solution

PV = P+V ……………… Pigment wetting

P ……………………….. Pigment (Non Aqueous)

Given a mixer of bulk capacity (B), several mixing stages (i = 1, 2, 3, … n) of

aqueous pigment (PW) and vehicle (V) are charged to the mixer in calculated

amounts such that the charge (PW + V) in any given stage (i), plus the paste or

wetted pigment that has already been mixed in prior stages, will always equal or

be less than the bulk capacity (B).

Formula #2: Before Mixing

B ≥ ∑ ( P +V )1,2 ,3,...i −1 + PW +V
i i
i =1

vi
 ABSTRACT

Formula #3: After Mixing B ≥ ∑ ( P +V ) + PV i + W i

1, 2 , 3,...i −1
i =1

The discharge of water, (W i), after any stage of mixing creates the net capacity

for the next stage of additives, (Pi+1 + Vi+1).

vii
LIST OF FIGURES/TABLES

Figure 1.01a
Growth Function

Figure 1.01b
LIST OF FIGURES/TABLES

Figure 1.02
Mix Sequence (Flush)

Figure 2.01
A Report From Model-C
 LIST OF FIGURES/TABLES

Ball-Mill Formulation
Example

Pigment 10.0 Stage I (grinding), Then add:

Resin 1.0
Solvent 3.0

Resin 1.0 Stage II (let down), Empty mill – then add:

Solvent 3.0

Resin 29.0 Stage III (completion of formula)

Solvent 51.5
Additives 1.5
100.0

Types of Viscometers

I. Capillary Viscometers
Absolute viscometers
Relative viscometers

II. Falling Body Viscometers

The Falling Sphere Viscometers
The Rolling Sphere Viscometers
The Falling Coaxial Cylinder Viscometer
The Band Viscometer

III. Rotational Viscometers

Coaxial Cylinder Viscometer
Cone-plate Viscometers

IV. Vibration Viscometers

 LIST OF FIGURES/TABLES

Table 1-1: Typical Viscosities and Shears

Shear Stress Shear Rate Viscosity

. (dynes/cm2) (sec-1) (poises)
Emulsion 280 7 40
500 29 17
625 72 9
Vinyl Plastisol 710 36 20
1430 58 25
2130 77 28

Rotation of Fuid Masses – Open Vessels

Proof that the form of the free surface of the liquid in a rotating vessel
is that of a paraboloid of revolution. The equation of the parabola is
ω2
y= x2
2g
 LIST OF FIGURES/TABLES

EXAMPLE OF EXPONENTIAL DECAY

Radioactive Decay Function

Table 4-2a: Comparison of Viscosity Values for Linseed Oil by Eqs. 2,

3, 4 & 5a with Experimentally Determined Values

Viscosity Values (poises)

Calculated
Temperature . .
(F) (K) Exp. Eq. 2 Eq. 3 Eq. 4 Eq. 5a
50 283 0.60 0.56 0.61 0.59 0.55
86 303 0.33 0.33 (0.33 Used in computation)
122 323 0.18 0.20 0.18 0.19 0.198
194 363 0.071 0.071 (0.071 Used in computation)
302 423 0.029 0.015 0.023 0.019 0.015

η = 773.05e −0.02561(T ) , is used to calculate the data in column Eq. 5a.

 CHAPTER 1

INTRODUCTION

Statement of the Problem:

Most of the written references on pigment dispersion focus on the

chemistry of organic colorants and the physical chemical properties of the

mixes and suspensions. The flushing process has progressed over the

years from grinding in a mixing vessel to movement through conduits to

complex helical mixing chambers.

The former method involves adding aqueous pigment (presscake)

and oil based vehicles into a sigma-blade mixing vessel over several

stages. The mixing displaces the water from the pigment-presscake and

encapsulates the pigment particles with the oil-based vehicles. The water

is poured off of the pigment dispersion and the cycle is repeated until the

vessel is filled to near capacity. The process is called flushing and it is as

much of an art as it is a science. Process operators modify the procedures

much like a cook uses a recipe. Very few processes are identical. Some

pigment and organic ink manufactures still use this process.

Quantifying this flush process is the primary focus of this project. By

using the above general description of the flushing process, models can

be created to simulate the procedure. These models will use bulk load

capacity and viscosity as the major constraints to produce the number of

Page 1
 CHAPTER 1

mixing stages that are required to optimize the quantities of presscake and

vehicle.

There are an infinite number of ways to load the ratio of vehicle to

pigment charges for each addition. The ratio used for each charge, is

usually determined by experimental methods in a laboratory environment.

One of the objectives in this project is to create some models and

methodologies that simulate this experimental process. By using the bulk

load capacity and viscosity as input parameters, these models will

calculate the required quantities of vehicle and pigment needed at each

mixing stage.

Purpose of the Study:

The purpose of this project is to show how the models are created

and used to predict and analyze the viscosities of resin solutions and

pigment dispersions prior to actual mixing. The models are mathematical

functions, which show how temperature, concentrations and other

parameters relate to the flow of end mixed product.

Further development of these models will show how mathematical

logic can be used to simulate and analyze complex mixing procedures

using relative viscosity and mixing capacity. These models will simulate

the paint flow and pigment dispersion dynamics used in industry.

Page 2
 CHAPTER 1

Importance of the Study:

The procedure for getting projects from concept to production

works much the same as it did decades ago, except for the upgrades in

plant, lab and computer equipment. Hopefully the system is more

productive and efficient. The need for analysis still remains and is even

more important. The experience of the technologist is just as important

now, if not more so. The procedure that is run in the lab is a model of

expected results in a production environment. The skill set of the

technologist, the quality of the lab equipment used and the quality of the

analysis of the results, will determine how well the lab results correlates to

the production application.

Scope of the Study:

Good models will yield plausible results, which can save time and

resources in development and production. If it is useful, it can be a

valuable tool. The models developed in this project have been created

with Math Cad and Microsoft Excel spreadsheets and will be detailed in

the appendices.

This project will refer to calculated or relative values of viscosity

(poise). In no way is it intended for these values to be interpreted as the

Page 3
 CHAPTER 1

absolute viscosity nor the coefficient of viscosity of the dispersion. At best,

the calculated viscosities and yield values are intended to estimate and

quantify the relative thickness of paints, pigments, resins and solutions

with respect to each other.

In this project, the bulk load capacity is the maximum pounds

required to optimize the mixer and produce the desired output. The unit of

measure used for the amounts of vehicle and pigment to be charged to

the mixing vessel will also be pounds.

The treatment of the models uses mathematics, which range from

Summation Algebra to Linear First Order Differential Equations. Most of

the mathematical expressions will be derived from logical statements,

much like postulates and proofs that are used in geometry. The proofs and

derivations, when required, will be detailed in the appendices.

The Rationale of the Study:

A few years ago, the typical industrial coatings development group

consisted of several gifted and creative people with many years of

rheological and analytical backgrounds. Their expertise ranged from the

graphic arts to Ph.D. in Engineering and Chemistry. It has been my

privilege to work with some of these individuals in the pigment

manufacturing and finished ink industry. At that time, microcomputer

Page 4
 CHAPTER 1

technology was being introduced into the color and coatings industry. A

typical pigment design problem would have required a senior technical

person to outline or sketch a flush color procedure and assign it to a junior

technician or engineer to work on. The technologist would review the lab

procedure, make the final calculated adjustments and gather the materials

needed to complete the lab procedure. Upon completion of the lab work,

the technologist reviews the results, completes the analysis and returns

the document to the senior technologist.

The primary objective of the methodologies and models that are

created in this project is to emphasize their importance and improve the

quality of the analysis and project management in a laboratory

environment.

More specifically, this project will show how models are created and

used to estimate viscosities of resin solutions. The models are comprised

of mathematical functions, which show how temperature, concentrations

and mixer capacity affect the flow of resin and pigment dispersions.

Further development of these models, show how mathematical

logic is used to simulate and analyze complex mixing procedures using

relative viscosities and mixing capacities. These models simulate the paint

flow and pigment dispersion dynamics that are currently used in industry.

Page 5
 CHAPTER 1

Definition of Terms:

Apparent Building up the body with respect to viscosity

Viscosity
Binding The maximum load in pounds a flush mixer will handle.
Bulk Capacity Property of certain pigment dispersion systems which causes them to
exhibit an abnormally high resistance to flow when the force which
causes them to flow is suddenly increased.
Colloid Dispersions of small particles of one material in another.
Dilatant The movement of wetted particles into the body of the liquid or
suspension.
Dispersion Same as wetting.
Encapsulate The reciprocal of Newtonian viscosity. Unit of measure is (Rhe)
Flocculation In the flushing process the moist cakes from the filter press are
introduced into a jacketed kneading type mixer together with the
calculated quantity of vehicle. During subsequent mixing, the oil or
vehicle displaces the water by preferential wetting, the separated water
being drawn off periodically; the final traces of water being removed,
when necessary, by heat and partial vacuum. The batch is then
sometimes given several grinds through a roller mill to complete the
process.

Fluidity The mechanical breakup and separation of the particle clusters to

isolated primary particles.
Flushing "True liquid:" A liquid in which the rate of flowis directly proportional to
the applied force
Grinding The solid portion of printing inks which impart the characteristics of color,
opacity, and to a certain part of the printing ink that is visible to the eye
when viewing printed matter.
Newtonian A viscous liquid which exhibits Plastic Flow. A liquid that has yield value
Liquid in addition to viscosity, and a definite finite force must first be applied to
the material to overcome the static effect of the yield value before the
material may be made to flow.
Oil Absorption The minimum amount of oil or varnish required to “wet” completely a unit
weight of pigment of dry color. Raw linseed oil is the reference vehicle in
the plant industry, while litho varnish of about twelve poises viscosity (#0
varnish) is the testing vehicle more commonly used in the printing ink
industry.
Pigment The moist cakes from the filter press are used in the flush process

Page 6
 CHAPTER 1

Plastic Material that has variable fluidity and no yield value.

Presscakes The science of plastic flow
Pseudoplastic Characteristic of false body or high yield value at rest. Applied aggitation
breaks down the false body to near newtonian flow, but will return to high
yield upon standing
Rheology The proportionality constant between a shearing force per unit area (F/A)
and velocity gradient (dv/dx).
Thixotropy Wetting refers to the displacement of gases (such as air) or other
contaminants (such as water) that are absorbed on the surface of the
pigment particle with subsequent attachment of the wetting medium to
the pigment surface.
Viscosity The action of a dispersed particles coming back together and forming
clusters. As a result, the body builds up thus causing a higher viscosity
or yield value.
Wetting The permanent property of an ink that is a measure of its inherent
rigidity. It refers to a certain minimum shear stress tha must be exceeded
before flow takes place
Yield Value Term used to indicate that the viscosity is that of a non-Newtonian liquid.
The adjective apparent is not meant to imply that the viscosity is an
illusory value, but rather that the viscosity pertains to only one shear rate
condition.

Page 7
 CHAPTER 1

OVERVIEW OF THE STUDY

Elementary science and basic chemistry taught us that mater

existed in one of three states; solid, liquid or gas. As we grew older, we

learned that substances exist in physical states, which are none of these

three basic states, but fall somewhere in between. Smoke, molasses,

varnish and paint are examples. P. W. Atkins, Physical Chemistry (1982),

p. 842, a college textbook, defines a colloid as “… dispersions of small

particles of one material in another.”

This project will focus on the methodology and model development

to approximate the flow and general rheological parameters combined

with the load capacities of the mixing vessel using aqueous displacement.

Herbert J. Wolfe, Printing and Litho Inks, (1967), p. 90, describes aqueous

displacement (flushing), “In the flushing process the moist cakes from the

filter press are introduced into a jacketed kneading type mixer together

with the calculated quantity of vehicle. During subsequent mixing, the oil

or vehicle displaces the water by preferential wetting, the separated water

being drawn off periodically; the final traces of water being removed, when

necessary, by heat and partial vacuum. The batch is then sometimes

given several grinds through a roller mill to complete the process.”

Page 8
 CHAPTER 1

The part of the above definition, which refers to the “… calculated

quantities of vehicle.”, is the primary focus of this project. Given a

quantity of pigment paste, there are an infinite number of given quantities

of vehicle that can be mixed with the paste, such that the ratio of vehicle to

pigment solids is greater than one. The definition of wetting, according to

Temple C. Patton, Paint Flow and Pigment Dispersion, 1st edition, (1963),

p. 217, “Wetting refers to the displacement of gases (such as air) or other

contaminants (such as water) that are absorbed on the surface of the

pigment particle with subsequent attachment of the wetting medium to the

pigment surface.”

This mixing process is repeated until a mass of flushed pigment,

suspended in vehicles (oils, varnishes and resin). The relative viscosity of

the end product is usually greater than the viscosity or yield value of the

first mixing stage. The first and early mixing stages are usually where

wetting takes place. Vehicle to pigment ratio is at its highest values during

wetting, to maximize the dispersion and encapsulation of the pigment

particles. Wetting is followed by a series of grinding and binding stages,

where the vehicle to pigment ratio is gradually decreased. Sometimes

vehicles of higher relative viscosities are used in these later stages in

order to build the body of the mix.

Page 9
 CHAPTER 1

The relative viscosity increases sharply in the early stages and

levels off as the number of mixing stages approaches the final stage (n). A

function that will model the building of the incremental viscosities, (hi),

over the stages, (1≤ i ≤ n), could be an exponential function (1 – ex) or a

logarithmic function, a[ln(x)]. Refer to Figure 1.01 below.

Figure 1.01a

Figure 1.01b

Page 10
 CHAPTER 1

In theory, there is no limit to the number of mixing stages that could

be used, but in reality, mixing capacity and the capacity to mix, is one the

key parameters, which implies a logical end point to stop the process.

Given a beaker and a spatula as the mixing utility, the capacity (B),

of the beaker and the ability to apply shear to the mixture of paste and

vehicles, tends to identify the some of the practical limits of the process.

The contents of the beaker and the energy required to mix the vehicle and

displace the water, should not exceed the beaker volume of the mixing

unit and cause overflow. Once the water is squeezed from the sticky mass

of wetted pigment, the water is discarded.

If the beaker volume is optimized prior to mixing, the new volume

for the next addition is equal to the volume of water discarded. This mixing

cycle is repeated until the working capacity of the mixer is reached and

there is no more room to mix without overflow. The number of mixing

stages (n) required to flush (P) amount of pigment is also determined

experimentally and is one of the parameters that will be used in this

project. For the sake of symbolic variables, (PW) will be assigned to

aqueous pigment paste, since it is composed of pigment, (P), and water,

(W). The variable assigned to vehicle is (V). The colloidal suspension or

pigment dispersion is assigned the variable (PV). Refer to Figure 1.02

below.

Page 11
 CHAPTER 1

Figure 1.02 (Flush Sequence)

The general mixing reaction equation is expressed as follows:

Formula #1: PW + V = PV + W

Given a mixer of capacity (B), several increments (n) of aqueous pigment

(PW), and vehicle (V), are charged to the mixer in amounts such that the

incremental charge (PW + V), will not overflow the mixer vessel. At the

end of each mixing stage, the water (W), becomes insoluble in the mixture

(PV + W), and is discharged from the vessel leaving only a sticky mass of

pigment dispersed in the vehicle (PV).

Page 12
 CHAPTER 1

i −1
Formula #2: Before Mixing: B ≥ ∑ ( P + V ) i + PWi + Vi
i =1

i −1
Formula #3: After Mixing: B ≥ ∑ ( P + V ) i + PVi + Wi
i =1

In the mixing scenario given above, (n), the number of mixing

sages required, has a direct relationship with the total pigment charge, (P),

water displacement, P(1/r-1), and total vehicle amount, (V). The number of

mix stages, (n), varies indirectly with the final % vehicle (xv) and the

working capacity of the mixer, (B). An empirical expression,

 P 
 ∑ +∑ V 
n= r  , will serve as an algorithm to estimate the parameter (n).
 xv B 
 
 

Viscosity:

The difficulty of mastering rheology, the science of flow and

deformation, is best summarized by T. C. Patton, Paint Flow and Pigment

Dispersion, 2nd edition, (1979), p. 1, “Unfortunately, flow phenomena can

become exceedingly complex. Even such a simple action as stirring paint

in a can with a spatula involves a flow pattern that challenges exact

mathematical analysis. However, simplifications and reasonable

approximations can be introduced into coating rheology that permit the

development of highly useful mathematical expressions. These in turn

Page 13
 CHAPTER 1

allow the ink or paint engineer to proceed with confidence in controlling

and predicting the flow performance of inks or paint coatings.”

Viscosity is defined as the opposition to fluidity. Water passes

through a funnel quickly; boiled oil slowly, while treacle would pass

through very slowly. An explanation for such varied rates of liquid

movement is as follows. When a liquid is caused to move, a resistance to

the motion, is set up between adjacent layers of the liquid, just as when a

block of wood is dragged along the floor. In the latter case, friction arises

between the two solid surfaces; in the case of a liquid, friction arises

between moving surfaces within it. This internal friction is called viscosity.

The frictional force, which opposes motion is felt when one moves a hand

through a tub of water. All liquids show a resistance to flow. Although

forces applied externally, affect the rate of liquid flow, viscosity is

concerned only with the internal frictional effect.

If two layers of a liquid are moving at different speeds the faster

moving layer experiences resistance to its motion, while the slower

moving layer experiences a force which increases its velocity. The

coefficient of viscosity is defined as the force in dynes required per square

centimeter to maintain a difference in velocity of 1 cm/sec between two

parallel layers of the fluid, which are ( ∆d ), 1 cm apart. This is best

represented in the following expression from James F. Shackelford,

Page 14
 CHAPTER 1

f∆d
Introduction to Materials Science for Engineers, (1985), p.329, η = ,
∆va

where (η ) is the coefficient of viscosity in poise, ( a ) is the area in cm2,

( ∆ν ) is change in velocity in cm/sec and ( f ) is the applied force in dynes.

stress
η=
rate _ of _ shear

The liquids for whose rate of flow varies directly with the applied

force ( f ), are called Newtonian Liquids. However, Non-Newtonian flow

is observed when the dispersed molecules are elongated, when there are

strong attractions between them or when dissolved or suspended matter is

present, as in resin and paint solutions. Most paint and pigment solutions

show Non-Newtonian viscosity to some degree.

Newtonian (Simple Flow):

An ideal liquid having a constant viscosity at any given temperature for low

to moderate shear rates.

Page 15
 CHAPTER 1

Non-Newtonian (Plastic Flow):

Flow with a yield value. This is a minimum shear stress value that must be

exceeded before flow will take place. Below yield value, the substance has

elastic properties. (Pigment-Resin-Solvent Solutions.)

Non-Newtonian (Pseudoplastic Flow):

A hybrid flow, which simulates plastic flow at moderate to high shear rates,

and Newtonian flow at low shear rates. (Paint and Ink Solutions.)

Non-Newtonian (Dilatant Flow):

Viscosity is reduced as shear stress is increased. This type of solution

gets thicker on increased agitation. (Rare Paint Systems)

Non-Newtonian (Thixotropic Flow):

Much like Pseudoplastic flow, but more complex and plasticized. In

general, thixotropic breakdown (loss of viscosity) is fostered by an

increase in the shear stress, by prolonging the shear time. When the

shear stress is removed, recovery of thixotropic viscosity ensues as

thixotropic structure is again built up throughout the paint system.

Page 16
 CHAPTER 1

Dispersions:

As a vehicle is incorporated with pigment by a mixing action, a

good dispersion initially displays significant resistance to sudden pressure,

turning dull in appearance. With further vehicle addition, the mixture

reaches a point where it coalesces into a smooth glossy mass. A small

additional increment of vehicle converts the mass into a mobile dilatant

dispersion.

Physically, this dispersion is characterized by deflocculated

particles, fully separated by a minimum of dispersion vehicle to give a

relatively closely packed system. If the shear stress applied to this

dispersion is low, sufficient time is allowed for the particles to slip and slide

around each other without contact. As a result of this action, a minimum

viscosity resistance results. If the shear stress is high, then adjacent

particles ram through the mix barrier separating themselves to establish

solid-to-solid contact. Without the lubrication afforded by the intervening

dispersion vehicle, major viscous resistance is exhibited.

Besides mixing the vehicle with the pigment particles, there is

another phenomena taking place which affects the body and consistency

of the dispersion. This action is absorption. The amount of absorption that

takes place depends on the interactive properties of the surface of the

pigment and the properties of the vehicle. The absorption causes some

Page 17
 CHAPTER 1

puffiness about the surface of the pigment particles and thus the same is

observed on a larger sampled mass. This puffiness causes a slight build

up in viscosity of the dispersion and also contributes to the flocculation.

The above properties will provide the basic resource for

constructing the logic and math models to simulate the flushing process.

Page 18
 CHAPTER 2

REVIEW OF RELATED LITERATURE

Paint Technology Manuals

PART TWO – Solvents, Oils, Resins and Driers

Published on behalf of The Oil & Colour Chemists’ Association – 1961

This manual covers the chemistry and physical chemical

characteristics of oils and resins. The book was very popular with

technologists in the coatings industry because it covered the chemical

derivations and practical applications with regard to paint manufacturing

and ink making. Regarding this project, it was a very useful resource for

information on resins and solvents.

Sometimes the technologist encounters significant chemical

reactions when mixing certain resin solutions such as driers. Without

taking into account the basic chemistry of solvents and resins, one might

assume that just mixing some oil with resin, a varnish like substance will

result. And by adding more oil or solvent to the mix, one would expect the

result to be a thinner solution, which should flow more easily. But what if

there is a reaction with the oxygen in the air, solvent and the resin and the

mix begins to thicken. This is what happens when a drier is created.

Page 19
 CHAPTER 2

Coatings of all resin solutions have a tendency to dry because of a

chemical process called oxidation. But what categorizes a resin solution

as a drier is the relative rate of drying, resin concentration and sometimes

temperature.

“Paints have been made for centuries by mixing pigments such as

red lead, white lead and umber with drying oils, and it became

obvious that these paints dried faster than the raw oils. Eventually it

was discovered that oils stored in the presence of lead or

manganese compounds, e.g. red lead or manganese dioxide, or

better still if heated in the presence of these compounds so as to

produce oil-soluble products, developed improved drying

properties; this formed the basis of the production of boiled linseed

oil; one of the foundations of paint formulation. “ (Atherton, 1961, p.

31)

Drying is just one of the many challenges that a coating technologist will

encounter. Because of various degrees of chemical reactions, there are

numerous levels of compatibility of solvents, resins and pigments. Today

these dispersions are classified into solvent and resin systems.

This project employs non-drying dispersions, which will allow

wetting to take place without rapid oxidation and aggregation. The mixing

Page 20
 CHAPTER 2

methodology assumes ideal systems of resins and solvents. This manual

on resins and solvents gave me a great appreciation on the complexity

and sophistication of the behavior of pigment and resin dispersions. There

is much room for further development of this mixing model using non-

idealistic resin solutions as vehicles. The ink chemistry and physics

involved in the actual rheology of pigment dispersions go far beyond the

level of mathematics used in this project.

Page 21
 CHAPTER 2

Introduction to Paint Chemistry

By G. P. A. Turner - 1967

Turner’s treatment of paint chemistry is somewhat of a general

treatment of the physics and chemistry of paint. It reviews the inorganic

and organic systems of paint chemistry. Turner also incorporates some of

the information that was previously covered on oils, solvents, resins and

driers. This book on paint chemistry is more of a general textbook on the

manufacturing and production of paint. It covers general atomic theory as

it relates to molecular bonding of compounds used in paint. He discusses

viscosity of suspensions and colloids. There is an introduction to

substrates and color theory, where the science of polymer coating is

explained quite clearly.

The chapter on pigmentation describes the dough mixer, which are

used by many pigment manufactures to produce distributions.

“A fourth type of mill is the heavy duty or ,‘pug’ mixer, in which

roughly S-shaped blades revolve in opposite directions and at

different speeds in adjacent troughs. A stiff paste is required.

Several alternative mills are available, which the reader may

discover elsewhere.” (Turner, 1967, p. 119)

Page 22
 CHAPTER 2

The dough mixers mentioned above is an accurate description of the

mixing vessels used in processing flush-color dispersions. The S-shaped

blades are called sigma blades. The paint mixing procedure is much like

the flush procedure. The primary stages and their functions are described

as follows.

“It is obvious from the mention of stiff pastes that the whole paint is not

charged into the mill. In fact, the paint maker aims to put in the

maximum amount of pigment of pigment and the minimum amount of

varnish to get the largest possible paint yield from his mill. This

mixture forms the grinding or first stage. When the dispersion is

complete (after a period varying from10 minutes to 48 hours according

to the materials and machinery involved), the consistency is reduced

with further resin solution or solvent, so that the mill can be emptied as

cleanly as possible. This is the ‘let-down’ or second stage and may

take up two hours. The third or final stage (carried out in a mixing

tank) consists of the completion of the formula by addition of the

remaining ingredients. A break-down of a possible ball mill formula

looks like this:

Page 23
 CHAPTER 2

Pigment 10.0 Stage I (grinding), Then add:

Resin 1.0

Solvent 3.0

Resin 1.0 Stage II (let down), Empty mill – then add:

Solvent 3.0

Resin 29.0 Stage III (completion of formula)

Solvent 51.5

Additives 1.5

100.0

The exact composition of Stage I is found by experiment, to give the

minimum grinding time and the most stable and complete dispersion.

Stages II and III also require care, as hasty additions in an incorrect

order can cause the pigment to re-aggregate (flocculate).

The amount of pigment in the formula is that required for the

appropriate colour, hiding power, gloss, consistency and durability. As

a rough guide, the amount might vary from one third of the binder

weight to an equal weight (for a glossy pastel shade).” (Turner, 1967,

p. 119)

Page 24
 CHAPTER 2

The above procedure is very much like the total flush procedure. This

project is focused on the referenced grinding Stage I where the pigment is

introduced into the system. In flushing, several stages are required to

introduce all of the pigment into the system. The first stage of the series of

grinding stages is called the wetting stage. In the wetting stage of flushing,

usually the largest charge of pigment and vehicle is introduced to the

mixer, where the vehicle charge is greater than the pigment. The purpose

for the vehicle-to-pigment ratio being greater than one is to allow for the

encapsulation of the pigment particles and maximum displacement of

water. A low viscosity, due to the large amount of vehicle present,

generally characterizes the wetting stage.

The subsequent stages are grinding stages, where the rest of the

pigment is charged to the mixer in lesser amounts. The vehicle-to-pigment

ratio for these stages is usually less than one. A graph of the viscosity of

the dispersion with respect to the number of stages, usually looks similar

to an exponential growth function. Refer to Figures 1.01a and 1.01b.

The 10.0% of pigment in the total paint dispersion shown above is

71.4% of Stage I. In flush procedures after the last stage of pigment

charge, the percent pigment is usually in the range of 50% to 60%.

Page 25
 CHAPTER 2

Ferranti Instrument Manual

The Measurement and Control of Viscosity

And Related Flow Properties

McKennell, R., Ferranti Ltd., Moston & Mancheser (1960)

The Ferranti Instrument Manual was written to give the ink

technician an overview of the complexity of measuring viscosity. The

manual lists four major types of viscometers and some examples of each.

Several types of non-Newtonian fluids are discussed. Different types of

non-Newtonian measurements are exemplified and matched with the best

type of viscometer. There are suggestions and examples of experimental

techniques for measuring various types of non-Newtonian substances for

experimental purposes as well as calibration.

A brief overview is given of how viscometers generate automatic

flow-curve recordings and the curves are analyzed. Basic viscosity

formulae are listed and discussed. Specific flow problems and suggested

solutions are discussed.

The list of the four major types of viscometers is listed below. It is

taken from the table of contents of the manual.

Page 26
 CHAPTER 2

Types of Viscometer:

I. Capillary Viscometers

Absolute viscometers

Relative viscometers

II. Falling Body Viscometers

The Falling Sphere Viscometers

The Rolling Sphere Viscometers

The Falling Coaxial Cylinder Viscometer

The Band Viscometer

III. Rotational Viscometers

Coaxial Cylinder Viscometer

Cone-plate Viscometers

IV. Vibration Viscometers

Besides being a great source for viscosity terminology, the section

on special flow problems, the suggestion of using the function of percent

solids content against apparent viscosity, is a major corner stone of the

methodology of this project.

The percentage solids content of slurries and similar

suspensions may be rapidly determined by constructing a

curve of percent solids content against apparent viscosity. A

Page 27
 CHAPTER 2

suitable shear rate must be chosen and adopted as standard

and equilibrium apparent viscosity readings taken on a number

of slurries of known percentage solids content. Determination

can be made in a fraction of the time required using

conventional gravimetric techniques, with an accuracy which is

acceptable for many applications. (Ferranti and McKennell,

1955, “Liquid Flow Problems and Their Solution”: Reprint from

Chemical Product)

Figure 2.01 is a report that was generated from one of the computer

programs written by the author for this project. It shows how percent

content is compared to apparent viscosity can be used as part of the flush

dispersion analysis.

Figure 2.01

A Report From Model-C

Page 28
 CHAPTER 2

Paint Flow and Pigment Dispersion

Patton, T. C., 1st edition (1963) & 2nd edition (1979)

Patton takes more of a mathematical approach to explore the

dynamic properties of resin and pigment dispersions. Both editions

provide a practical and comprehensive overview of rheological aspects of

paint and coatings technology. The second edition includes expanded

material on pigment-binder geometry, the theoretical aspects of

dispersion; and a more detailed breakdown of grinding equipment.

The sections that are most referenced for this project are the ones

which elaborate on viscosity, the effects of temperature and resin

concentration on viscosity and pigment dispersion theory.

Viscosity

The treatment of viscosity theory is the same as the other resources.

Patton uses tables which lists various substances and their viscosities to

help the reader better understand the concept of flow. He also uses tables

to show how well viscosity formulae correlate to actual experimental data.

The models in this project will also use tables. I considered this illustrative

technique to be very effective especially in showing results for analysis.

Page 29
 CHAPTER 2

Table 1-1

Shear Stress Shear Rate Viscosity

. (dynes/cm2) (sec-1) (poises)
Emulsion 280 7 40
500 29 17
625 72 9
Vinyl Plastisol 710 36 20
1430 58 25
2130 77 28

(Patton, 1963, 1st edition, p. 9, Table 1-1)

Temperature and Viscosity

Patton refers to a formula based on experimental data of temperature and

related viscosity of a liquid.

“It has been found experimentally that for any given viscosity h

the change in viscosity dh produced by a change in

temperature dT is substantially the same for most liquids.

Furthermore, the function f(h) of Eq. 1 depends primarily on the

magnitude of the viscosity only (it does not depend appreciably

on the nature of the liquid).”

dη
= f (η ) Eq. 1
dT

(Patton, 1979, 2nd edition, p. 91)

Page 30
 CHAPTER 2

“Of the many equations that have been proposed for relating

viscosity to temperature, one appears to represent the

viscosity/temperature relationship most accurately. It is commonly

referred to as Andrade’s equation (Eq. 2).

η = A(10 B / T Eq. 2

Equation 2 can be expressed alternatively in logarithmic form as Eq. 3.

B
log η = log A + Eq. 3
T

Temperature T must be expressed in absolute units (K = 273 + C or

R = 460 + F), and A and B are constants for the liquid in question.

If subscripts 1 and 2 are used to denote the conditions for two

different temperatures, it can be readily shown (by subtraction) that the

two conditions are related by Eq. 4.”

η1 1 1
log = B( − ) Eq. 4
η2 T1 T2

(Patton, 1979, 2nd edition, p. 93)

The following table shows how temperature effects the viscosity of linseed

oil and also how well the above equations fit actual experimental data.

This table will also serve as a resource to measure the accuracy of the

formulae and models which will be developed in this project.

Page 31
 CHAPTER 2

Table 4-2: Comparison of Viscosity Values for Linseed Oil by Eqs. 2,

3 & 4 with Experimentally Determined Values

Viscosity Values (poises)

Calculated
Temperature . .
(F) Exp. Eq. 2 Eq. 3 Eq. 4
50 0.60 0.56 0.61 0.59
86 0.33 0.33 (0.33 Used in computation)
122 0.18 0.20 0.18 0.19
194 0.071 0.071 (0.071 Used in computation)
302 0.029 0.015 0.023 0.019

(Patton, 1979, 1st edition, p. 85, Table 4-2)

Resin Concentration

“A common viscosity problem calls for calculating the change in a solution

viscosity produced by a change in resin concentration. Such a change

may be due to addition of let-down thinner, or it may occur as a result of

blending together two compatible resin solutions.” (Patton, 1979, 1st

edition, p. 88)

Equations Relating Viscosity to Resin Concentration

“The simplest expression and possibly a fully adequate one for most

purposes for relating solution viscosity to resin concentration takes the

Page 32
 CHAPTER 2

form of Eq. 9, where x is the fractional content of nonvolatile resin in the

resin solution and A and B are constants.

η = A(10 Bx ) or log η = log A + Bx Eq. 9

To evaluate the constants A and B, solution viscosities at two different

resin concentrations must be known. Once A and B are determined, a

viscosity for any third resin concentration is obtained by straightforward

substitution in Eq. 9.” (Patton, 1979, 1st edition, p. 88)

The data in the above tables will be referenced in later chapters to

illustrate how other methodologies compare for accuracy and use in

several models. Patton uses logarithms to the base-10 in the above

methodologies. The formulae that will be used in the development of the

flush models will use logarithms to the base-e or natural logs.

Page 33
 CHAPTER 2

Printing and Litho Inks

Wolfe, H. J. (1967)

Printing and Litho Inks blends the history and art of ink making with

the world of science and technology. It is the source of many ink-making

terms that are used in this project. Flushing is described as follows.

“As is well known, the kneading type of mixer also is employed in the

“flushing” of pulp colors, i.e., the production of pigment-in-oil pastes

directly from pigment-in-water pastes, by introducing the water-pulp

color and the varnish into the mixer and agitating until the varnish has

displaced the water. Steam-jacketed mixers are generally employed

for this purpose. Air-tight covers also may be fitted to these mixers so

that vacuum may be employed to remove the water from the pulp

more rapidly.” (Wolfe, 1967, p. 445)

The viscosity of the dispersion, relative particle size and oil absorption of

the pigment are very important characteristics that help determine the

point at which to stop mixing. Wolfe provides tables and details about the

properties of the different classes and types of pigment in dry color state.

These dry properties directly relate to the flush procedure (pigment

suspended in water) because the resulting dispersion after displacing the

water (flushing), will have the same properties as if it were mixed dry. If

Page 34
 CHAPTER 2

the pigment particle size and the vehicles are the same, then the end

result should be the same. The only difference will be the grinding

methods. Regarding pigments, resins and solvents, Wolfe lists standard

testing procedures, test equipment and test methods used in the printing

ink industry. For example:

The term “oil absorption” as used in the dry color and printing ink

industries refers to the minimum amount of oil or varnish required to

“wet” completely a unit weight of pigment of dry color. Raw linseed oil

is the reference vehicle in the plant industry, while litho varnish of

about twelve poises viscosity (#0 varnish) is the testing vehicle more

commonly used in the printing ink industry. (Wolfe, 1967, p. 472)

Viscosity is without a doubt, the most important characteristic of a printing

ink vehicle, since it determines the length, tack and fluidity of the vehicle;

which in turn in a large measure, determines the working qualities of the

resulting inks. Although listed separately, the properties of viscosity are

directly related to those of oil absorption and particle size. These three

topics are key to the flush models developed in this project.

Page 35
 CHAPTER 2

Physical Chemistry

Atkins, P. W. (1982)

The concepts of suspensions and viscosity are covered quite

extensively in physical chemistry. To get more clarification on these terms,

Atkins’ “Physical Chemistry” is a good resource. For example:

“A major characteristic of liquids is their ability to flow. Highly viscous

liquids, such as glass and molten polymers, flow very slowly because

their large molecules get entangled. Mobile liquids like benzene have

low viscosities. Water has a higher viscosity than benzene because its

molecules bond together more strongly and this hinders the flow.

We can expect viscosities to decrease with increasing

temperatures because the molecules then move more energetically

and can escape from their neighbors more easily. (Atkins, 1982, p.18)

Regarding the relationship of viscosity to particle size as stated

above in Wolfe, Atkins confirms this as follows:

The presence of macromolecules affects the viscosity of the

medium, and so its measurement can be expected to give

information about size and shape. The effect is large even at low

concentrations, because the big molecules affect the surrounding

fluid’s flow over a long range.” (Atkins, 1982, p.825)

Page 36
 CHAPTER 2

Fluid Mechanics and Hydraulics

Giles, R. V. (June, 1962)

Fluid mechanics and hydraulics explained the viscosity in such an

abstract manner that it was somewhat limited as a resource in this project.

However, the viscosity units of measure were clearly explained and came

in very handy when the property of density was introduced.

Absolute or Dynamic Viscosity (m)

Viscosity of a fluid is that property which determines the amount of

its resistance to a shearing force. Viscosity is due primarily to

interaction between fluid molecules. (Poise, lb-sec/ft2)

Kinematic Viscosity (n)

Kinematic coefficient of viscosity is defined as the ratio of absolute

viscosity to that of mass density (r). (Stokes, ft2/sec = m/r)

(Giles, 1962, p.3)

Page 37
 CHAPTER 2

Rotation of Fuid Masses – Open Vessels

The form of the free surface of the liquid in a rotating vessel is that

of a paraboloid of revolution. Any vertical plane through the axis of

rotation which cuts the fluid will produce a parabola. The equation

ω2
of the parabola is, y = x 2 where x and y are coordinates, in feet,
2g

of any point in the surface measured from the vertex in the axis of

revolution and w is the constant angular velocity in rad/sec. Proof

of this equation is given in Problem 7. (Giles, 1962, p.42)

Problem 7. An open vessel partly filled with a liquid rotates about

a vertical axis at constant angular velocity. Determine the equation
of the free surface of the liquid after it has acquired the same
angular velocity as the vessel.

Page 38
 CHAPTER 2

Solution: Fig. (a) represents a section through the rotating

vessel, and any particle A is at a distance x from the axis of

rotation. Forces acting on mass A are the weight W vertically

downward and P which is normal to the surface of the liquid since

no friction is acting. The acceleration of mass A is xw2, directed

toward the axis of rotation. The direction of the resultant of forces

W and P must be in the direction of this acceleration, as shown in

Fig. (b).

W
From Newton’s second law, Fx = Max or (1) P sin θ = xω 2
g
From SY = 0 (2) P cos θ = W

xω 2
Dividing (1) by (2), (3) tan θ =
g

Now q is also the angle between the X-axis and a tangent drawn to
the curve of A in Fig. (a). The slope of this tangent is tan θ or dy .
dx
Substituting in (3) above,

dy xω 2 ω2
= from which, by integration, y= x 2 + C1
dx g 2g

To evaluate the constant of integration, C1: When x = 0, y = 0 and C1 = 0.

(Giles, 1962, p.42, Problem 7)

Page 39
 CHAPTER 2

Manual of Chemical Engineering Calculations & Shortcuts

New Analysis Provides Formula to Solve Mixing Problems

Brothman, A, Wollan, G, & Feldman, S. (1947)

Universally used by the process industries, mixing operations have

been the subject of considerable study and research for several years.

Despite these efforts, mixing has remained an empirical art with little

foundation of scientific analysis as found in other important unit

operations. A new approach based on a study of kinetics and on the

concept that mixing is essentially an operation of three-dimensional

shuffling, has resulted in a formula for solving practical problems.

“Mixing is that unit operation in which energy is applied to a mass of

material for the purpose of altering the initial particle arrangement so

as to effect a more desirable particle arrangement. While the object of

this treatment is usually to blend two or more materials into a more

homogenous mixture, it may also serve to promote accompanying

reactions, or it may support other unit operations such as heat

transfer.” (Brothman, Wollan and Feldman, 1947, p.175)

Page 40
 CHAPTER 2

The above referenced chapter is about applications of analytical

methods, which brings forth a new relationship between mixing time and

mixing completion. Based on the theory of probability and resulting from a

study of mixing kinetics, the derived expression and its implications may

well lead the way to closer and more reliable correlation of mechanical

design and functional performance of mixers. The concepts and mixing

methodologies are explained in shuffling operation, blending, turbulence

and liquid mixing. Most of the concepts and methodologies referenced in

this book included the mixing time.

After careful consideration, the author decided to omit these

methodologies from this book because their complexity was beyond the

scope of this project. However, the time of mixing, though not used directly

in my models, will have an indirect correlation to the algorithm that will be

used to estimate number of mixing stages required. These mixing

concepts related to functions that are continuous, where my models are

more mathematically discrete.

Page 41
 CHAPTER 2

Advanced Engineering Mathematics

Kreyszig, E. (August 1988)

Modeling Physical Applications

Differential equations are of great importance in engineering, because

many physical laws and relations appear mathematically in the form of

differential equations. Referring to T. C. Patton’s expression (Eq. 1), which

describes the physical relationship of viscosity (h) and temperature (T).

dη
= f (η ) Eq. 1
dT

(Patton, 1979, 2nd edition, p. 91)

Although Patton uses the differential expression (Eq. 1), to describe the

relationship between viscosity and temperature, the development of the

formulae that are used in Andrade’s equations (Eq. 2, Eq. 3, and Eq. 4), is

not shown.

η = A(10 B / T ) Eq. 2

B
log η = log A + Eq. 3
T

η1 1 1
log = B( − ) Eq. 4
η2 T1 T2

(Patton, 1979, 2nd edition, p. 93)

Page 42
 CHAPTER 2

Kreyszig describes the development process in a detailed step by step

example of a radioactive decay problem below.

EXAMPLE 5. Radioactivity, exponential decay

Experiments show that a radioactive substance decomposes at a rate

proportional to the amount present. Starting with a given amount of

substance, say, 2 grams, at a certain time, say, t = 0, what can be

said about the amount available at a later time?

Solution. 1st Step. Setting up a mathematical model (a differential

equation) of the physical process.

We denote by y(t) the amount of substance still present at time t. the

rate of change is dy/dt. According to the physical law governing the

process of radiation, dy/dt is proportional to y.

dy
(9) = ky
dt

Here k is a definite physical constant whose numerical value is known

for various radioactive substances. (For example, in the case of

226
radium 88Ra we have k ~ -1.4 x 10-11 sec-1.) Clearly, since the

amount of substance is positive and decreases with time, dy/dt is

negative, and so is k. We see that the physical process under

Page 43
 CHAPTER 2

consideration is described mathematically by an ordinary differentia

equation of the first order. Hence this equation is the mathematical

model of that process. Whenever a physical law involves a rate of

change of a function, such as velocity, acceleration, etc., it will

lead to a differential equation. For this reason differentia

equations occur frequently in physics and engineering.

2nd Step. Solving the differential equation. At this early stage of our

discussion no systematic method for solving (9) is at our disposal.

However, (9) tells us that if there is a solution y(t), its derivative must

be proportional to y. From calculus we remember that exponential

functions have this property. In fact the function ekt or more generally

(10) y (t ) = ce kt

where c is any constant, is a solution of (9) for all t, as can readily be

verified by substituting (10) into (9). [We shall see later (in Sec. 1.2)

that (10) includes all solutions of (9); that is (9) does not have singular

solutions.]

3rd Step. Determination of a particular solution. It is clear that our

physical process has a unique behavior. Hence we can expect that by

using further given information we shall be able to select a definite

Page 44
 CHAPTER 2

numerical value of c in (10) so that the resulting particular solution will

describe the process properly. The amount of substance y(t) still

present at time t will depend on the initial amount of substance given.

This amount is 2 grams at t = 0. Hence we have to specify the value

of c so that y = 2 when t = 0. This condition is called an initial

condition, since it refers to the initial state of the physical system. By

inserting this condition

(11) y (0) = 2

in (10) we obtain

y(0) = ce0 = 2 or c=2

If we use this value of c, then the solution (10) takes the particular form

(12) y (t ) = 2e kt

Page 45
 CHAPTER 2

This particular solution of (9) characterizes the amount of substance

still present at any time t ≥ 0 . The physical constant k is negative, and

y(t) decreases, as shown in Fig. 5.

4th Step. Checking. From (12) we have

dy
= 2ke kt = ky and y (0) = 2e 0 = 2
dt

We see that the function (12) satisfies the equation (9) as well as the

initial condition (11). The student should never forget to carry out

this important final step, which shows whether the function is (or

is not) the solution of the problem.

(Kreyszig, 1988, p.8)

dη
Based on the Kreyszig modeling example, (Eq. 1), = f (η ) , has the
dT

following solution.

dη
∝η
dT

dη
= kdT
η

dη
∫ η
= k ∫ dT

ln η = kT + c

Page 46
 CHAPTER 2

η = e kT +c

η = Ce kT given; C = ec

Since molecular motion approaches zero, at absolute zero at T=0oK, ho.

(Temperature in degrees Kelvin)

η = e k ( 0 )+c

ηo = ec = C at T = 0oK

η = η o e kT Eq. 5a

Note: Refer to T.C. Patton’s experimental data Table 4-2a listed below.

Convert degrees Fahrenheit (F), to absolute, degrees Kelvin (K) and add

the additional columns (K) and Eq. 5a. The viscosity and temperature data

from the table (Used in computation) was plugged into my new model

equation, Eq. 5a, to create a pair of simultaneous equations;

.33 = η o e k ( 303) and 0.071 = η o e k ( 363)

When this pair of simultaneous equations are solved for the constants, k

and ho, their calculated values are; k = -0.02561 and ho = 773.05.

η = 773.05e −0.02561(T ) , is used to calculate the data in column Eq. 5a.

Page 47
 CHAPTER 2

Table 4-2a: Comparison of Viscosity Values for Linseed Oil by Eqs. 2,

3, 4 & 5a with Experimentally Determined Values

Viscosity Values (poises)

Calculated
Temperature . .
(F) (K) Exp. Eq. 2 Eq. 3 Eq. 4 Eq. 5a
50 283 0.60 0.56 0.61 0.59 0.55
86 303 0.33 0.33 (0.33 Used in computation)
122 323 0.18 0.20 0.18 0.19 0.198
194 363 0.071 0.071 (0.071 Used in computation)
302 423 0.029 0.015 0.023 0.019 0.015

After comparing the calculations of Eq. 2 and Eq. 5a, I conclude that the

same methodology was used to solve Patton’s differential equation. The

exception is that T. C. Patton used logarithms to the base “10,” where the

author uses natural logs or logarithms to the base “e.”

Since the calculus is much more straightforward, the author will

also be using the natural logarithmic methodology for solutions in the flush

models to show the relationship between dispersion concentrations and

viscosity distributions.

Page 48
 CHAPTER 2

Engineering Mathematics

Stroud, K.A., 5th edition (2001)

One of the models, Model-C, that will be created is based on the

geometric progression. The author found Stroud to be an excellent source

for reviewing the concepts relating to the Geometric Series. The

applications of geometric progression that first came to mind were

problems of finance, like compound interest.

Model-C uses sequences the pigment concentrations after each

break as a geometric series. If the pigment concentrations are geometric

in nature, then their pigment charges are geometric. A geometric model

will allow the progression elements to be summed by formula. The very

first illustration that Stroud uses in his chapter, Series 1, “Geometric series

(geometric progression), denoted by GP,” problem 11:

An example of a GP is the series:

1 + 3 + 9 + 27 + 81 + … etc.

Here you can see that any term can be written from the previous term

by multiplying it by a constant factor 3. This constant factor is called

the common ratio and is found by selecting any term and dividing it

by the previous one:

e.g. 27 ÷ 9 = 3 ; 9 ÷ 3 = 3 ; … etc.

Page 49
 CHAPTER 2

A GP therefore has the form:

a + ar + ar 2 + ar 3 + ... etc.

where, a = first term, r = common ratio.

(Stroud, 2001, p. 752)

When the author saw this example, he thought of the flush distribution of

pigment charges. The graph of the geometric sequence 1, 3 ,9, 27, 81, is

shown below.

90
80
70
60
50
Series1
40
30
20
10
0
1 2 3 4 5

To the author, it looks like the upside-down version of the exponential

graph in chapter 1. Therefore he felt that he should be able to manipulate

the sequence, mathematically and produce the exponential pattern. This is

what led to the creation of “Model-C, The Geometric Series.”

Page 50
 CHAPTER 3

METHODOLOGY

Approach

The initial idea which led to the development of this project was

conceived while serving in the capacity of a formulator whose duties were

to prepare the work orders and process the procedures in a pigment

manufacturing plant. The senior technologist was the person who initiated

the process by preparing the small research batches in the laboratory and

in the plant. Upon completion of the development stage, the plant

formulator would scale the procedure up so that it could be run in a

production size mixer.

After preparing numerous plant procedures for flush dispersions, a

common pattern was noticed about all of the procedures and that was the

pigment and vehicle charge distribution. As the mixing stages progressed

from the initial stage to the final stage, both the pigment and the vehicle

charges were always decreasing in amount. It was also noticed that the

viscosity of the dispersions seemed to be building in an exponential

growth curve pattern. It was also noticed that as the skeletal development

procedures were scaled up to production capacities, adjustments had to

be made because of a non-linear relationship between the material

distribution and the production mixer capacities.

Page 51
 CHAPTER 3

Data Gathering Method

The need to make the connection between material distribution and

the mixer capacity sparked an interest and curiosity, which led to an in-

depth journey into the research of pigment and paint dispersions. The

subject matter was a scientific and mathematical excursion into the world

of measurement of viscosity and its applications.

Creating a mathematical model to simulate the flush procedure was

the best way try to produce the same patterns that kept showing up in the

plant work orders. The method by which the models were created can

best be described as mathematical. Most of the math focused on the

dynamics of growth functions and their applications. The viscosity

applications required the use of first order differential equations and the

algebra of exponential functions. The mixer capacity applications involved

summation algebra.

The methodology of the flush models had to be created with pure

mathematics and then modeled into programs. Once the models were

coded into program logic and math scripts, it was easier to experiment

with the parameters and their inter-relationship.

Page 52
 CHAPTER 3

Database of Study

Most of the research literature was focused on dispersions and the

viscosity of resin solutions. Temple C. Patton’s “Paint Flow and Pigment

Dispersion” was the best resource for this project because of his

mathematical treatment of the subject.

Most of the math that was used in this project was a result of

cumulated mathematical training over the years. Some of the advanced

math required some review in the area of applied differential equations.

The mathematics of finance was a great resource for reviewing

applications of the infinite series.

MathCAD and Basic Programming were very useful tools in

creating and testing the models. They were very good resources for

producing quick results with minimum effort. My programming experience

came in very handy.

Validity of Data

The resource literature contained tables of experimental data that

served as a target for the models to reproduce. The primary strategy was

to use logical empirical modeling to reproduce the experimental results.

The validity of the output from the models will rely entirely on analysis. A

good model will closely mimic the experimental data.

Page 53
 CHAPTER 3

Originality and Limitations of Data

There was very little literature found that would imply that this

approach to creating flush models has been attempted. The concept is

quite simple in nature, but because of the complexity of the technology

related to viscosity measurement and fluid dynamics, it gets quite involved

mathematically. The model output is empirical and it’s objective is to serve

as a tool for the ink technologist when analyzing flush procedures.

After the models were completed, they only opened the door to

more questions. These models only address the mixing stages as a

discrete function. There is so much more to be learned from the mixing

dynamics that take place between the stages. The focus of this project

limited and simplified the units of measure to obtain its objective.

However, there is much potential to advance this project and incorporate

the concepts of energy usage and manufacturing cost analysis.

Summary of Chapter 3

Most of the methods and techniques used in creating the models

for this project are simple and straightforward. The output of the models

can only be analyzed and compared to data that is documented within the

resource literature.

Page 54
 CHAPTER 3

The core of the creation of the models lies within the chapters that

show the steps in the longhand mathematical development of the logical

functions and relationships from which the models are built. The proof

development will be shown in the appendix.

Page 55
 CHAPTER 4

DATA ANALYSIS

Observed Process Reaction Per Mixing Stage (All Treatments)

A given amount of presscake, PW, is mixed with a given amount of vehicle,

V, to produce a paste, PV (wetted pigment) and displaced water, W.

Formula #23: PW V  PV W

PW ……………………. Aqueous Pigment (Presscake)

W ……………………… Displaced Water

V ………………………. Resin or Resin Solution

PV = P+V ……………… Pigment wetting

P ……………………….. Pigment (Non Aqueous)

Given a mixer of bulk capacity (B), several mixing stages (i = 1, 2, 3, … n) of

aqueous pigment (PW) and vehicle (V) are charged to the mixer in calculated

amounts such that the charge (PW + V) in any given stage (i), plus the paste or

wetted pigment that has already been mixed in prior stages, will always equal or

be less than the bulk capacity (B).

Formula #24: Before Mixing

B ( P V )1,2,3,...i 1 PW i V i

i 1

Page 56
 CHAPTER 4

Formula #25: After Mixing

B (P V) PV i W i

1, 2 , 3 ,...i1
i1

The discharge of water, (Wi), after any stage of mixing creates the net capacity

for the next stage of additives, (Pi+1 + V i+1).

Wi PWi1 Vi1

Wi1 PWi2 Vi 2

Wi2 PWi 3 Vi3

W1 PW V W given 1 i 

Theoretically, this process could go on forever; i  , but a point is rapidly

approached where a decision must be made to end the process. This final stage

is designated as the nth or last stage (n). So the final expression that shows Wn

is; Wn 1 PWn V n Wn given 1 i n

The function or algorithm which approximates the number of stages required to

n
mix a total amount of pigment, P , having a solids contents of (r), with a total
i
i
1

n
amount of vehicle or resin solution, V , into a mixer vessel of bulk capacity (B),
i
i1

n n
(1 / r)(P Vi )
is Formula #8; n i
1 i
1
. The ratio of total pigment charge
x pn B

Page 57
 CHAPTER 4

P i
to total charge is designated as Formula #11, xpn. x pn  n i
1
. This ratio
( P V ) i i
i
1

also is indirectly related to the number of stages (n) in Formula #8 above, which

is required to completely mix the pigment with the resin solution and displace all

of the water.

Treatment – I: MODEL (A) requires initial amounts of pigment and vehicle (non-

optimized) to be charged to the mixer. The model calculates the amounts of

pigment and vehicle charges that are required for each mixing stage so that the

sum of the increment charges will equal the optimized total charge. In other

words, this model distributes the total charge to agree with the given viscosity

distribution. Optimization is the primary focus of this treatment while adhering to

a given viscosity distribution and holding the mixer capacity constant. The

calculated capacity, B(i) , is is an output parameter and will be listed at each

mixing stage to compare to the constant capacity, B. The input parameter, E 0

(Allowance), is the estimated % of the constant capacity. Theoretically, E0 is

equal to the water displacement in the final mixing stage.

INPUT DATA OUTPUT DATA

Capacity Constant B Calculated Capacity at stage (i) B(i).

Initial Pigment Charge SP(i) Number of mixing stages (n)

Initial Vehicle Charge SV(i) % Vehicle after last stage (x n)

Relative Viscosity of the Pigment (hp) System Viscosity Constant (k v)

Relative Viscosity of the Vehicle (hv ) Viscosity Distribution h(i)

Page 58
 CHAPTER 4

INPUT DATA (Continued) OUTPUT DATA (Continued)

% Solids of Presscake (r) % Pigment per stage xp(i)

Viscosity Distribution Function f(i) % Vehicle per stage xv(i)

Allowance E0 Pigment Charge per stage P(i)

Vehicle Charge per stage V(i)

Water Displacement per stage w d(i)

Total Pigment Charge SP(i)

Total Vehicle Charge SV(i)

Example Problem MODEL-A:

Given the non-optimized charge, SP(i) = 1350 lbs. SV(i) = 1200 lbs.

Relative Viscosity of the Pigment (hp ) = 240,000 poise. Relative Viscosity of the

Vehicle (h v) = 100 poise. % Solids of Presscake (r) = 20%: (0.20). Capacity

Allowance E0 = 15%: (Eff% = 85% Refer to Formula #22). Optimize the flush

procedure to be mixed in a vessel of Capacity, B = 3000 lbs.

Solution Steps MODEL-A:

Step #1

Using Formula #11: Calculate the % pigment (xp)in total charge.

n
n
 Pi
P i
1347 . 42
x p  (% i 1
 i 1

(1347 . 42 1202 . 58 ) = 0.5294
Eff )( B ) n

( P
i 1
i V i )

xv x p 1 ; xv 1 x p xv 10.5294 xv 0.4706

Step #2

Page 59
 CHAPTER 4

Using Formula #8: Estimate the number of required mixing stages (n)

(1 E0 ) rp xv (0.85)(0.05294

.20 ) 
x )
0.4706
Formula #8 n=  = 5.7813
xv n 0.4706

Round (5.7813) up to n = 6 Stage counter i, (1 to n); 1 i n

Step #3


Using Formula #15 k v ln( v ) Viscosity Constant for the mix


100
k v ln( ) 7.7832
240000

Step #4

Using Formula #16 n p e kvxv Viscosity of the mix

n 240000e ( 7.7832 )(0 .4706 ) 6159.19

Step #5

n
Using Formula #17 a  Relative Viscosity Constant
ln( n 1)

6110.871
a 3165.1975
ln(6 1)

Step #6

Using Formula #18 i a[ln(i 1)] Viscosity Distribution 1 i n

Viscosity ( i ) = (1) 2193.9477, (2) 3477.3249, (3) 4387.8955,

(4) 5094.1889, (5) 5671.2726, (6) 6159.1900

Page 60
 CHAPTER 4

Step #7


ln( i )
p
Using Formula #26 xvi  % Vehicle Distribution given (1 i n )
kv

xvi (1) 0.6032, (2) 0.5440, (3) 0.5142, (4) 0.4950, (5) 0.4812, (6) 0.4706

xv x p 1 ; x p 1 xv

x p i (1) 0.3968, (2) 0.4560, (3) 0.4858, (4) 0.5050, (5) 0.5188, (6) 0.5294

Step #8

i
1
Bx pi  Pi
Using Formula #19 Pi 1r xv (i111r ) Pigment Distribution for 1 i n
i

Pi (1) 460.1058, (2) 321.4687, (3) 229.6540, (4) 166.8274,

n
(5) 123.0372, (6) 92.1022 P 1393.1953
i
1
i Optimized

Step #9

P
Using Formula #6 Wi  i Pi Water displacement distribution 1 i n
r

Wi (1) 1840.4232, (2) 1285.8750, (3) 918.6159, (4) 667.3095,

(5) 492.1489, (6) 368.4088

Step #10

Using logic, V1 B P1 W1 : V1 3000 460.1058 1840.4232 V1 699.4710

P 321.4687
V2 B V1  2 : V2 3000 699.4710  V2 233.0795
r 0.20

Page 61
 CHAPTER 4

Step #11

i 1

Using Formula #20 Vi B ( Pi V i )  r : Vehicle Distribution

i P

i 1

Vi (1) 699.4710, (2) 233.0795, (3) 137.6051, (4) 84.4791,

n
(5) 52.1233, (6) 31.6379 V i 1238.3959 Optimized
i
1

Refer to the model, MatCad MODEL_A in the appendix (A14)

Refer to BASIC Program Reports: MODEL-A1 in the appendix (A11)

BASIC PROGRAM REPORTS

REPORT-MODEL-A

Page 62
 CHAPTER 4

Treatment – II- MODEL-B requires (xp), the % pigment in the total mix, as an

input parameter. This parameter along with the capacity, B, is used to calculate

the initial pigment and vehicle charges, which are required as input parameters in

Treatment-I. The remaining steps of the procedure and the objectives are

identical to Treatment-I. The model uses the mixer’s capacity along with the

viscosity distribution as the critical input parameters to optimize the loading of

each mixing stage and optimize the yield. The total amount of pigment and

vehicle required to charge the mixer is an output parameter in this procedure.

INPUT DATA OUTPUT DATA

Mixer Capacity (B) Number of mixing stages (n)

% Pigment after last stage (x p) % Vehicle after last stage (x n)

Relative Viscosity of the Pigment (p ) System Viscosity Constant (k v)

Relative Viscosity of the Vehicle (v ) Viscosity Distribution (i )

% Solids of Presscake (r) % Pigment per stage xp(i)

Viscosity Distribution Function f(i) % Vehicle per stage xv(i)

Allowance E0 Pigment Charge per stage P(i)

Vehicle Charge per stage V(i)

Water Displacement per stage w d(i)

Total Pigment Charge SP(i))

Total Vehicle Charge SV(i)

Calculated Capacity at stage (i) B(i)

Page 63
 CHAPTER 4

Example Problem MODEL-B:

Given the Pigment Content of the mix (xp)=0.5294, Relative Viscosity of

the Pigment (hp ) = 240,000 poise. Relative Viscosity of the Vehicle (h v) = 100

poise. % Solids of Presscake (r) = 20%: (0.20). Capacity Allowance E0 = 15%:

(Eff% = 85% Refer to Formula #22). Optimize the flush procedure to be mixed in

a vessel of Capacity, B = 3000 lbs.

Note: Given the same input parameters of MODEL-A with the exception

being (xp), which is a calculated output parameter, the output of MODEL-B is

expected to be the same as the optimized output of MODEL-A.

Solution Steps MODEL-B:

Step #1

Calculate the % vehicle (xv)in the total charge.

xv x p 1 ; xv 1 x p xv 10.5294 xv 0.4706

Step #2

Using Formula #8: Estimate the number of required mixing stages (n)

(1 E0 ) rp xv (0.85)(0.05294 ) 0.4706

x )

#8 n=  .20
= 5.7813
xv n 0.4706

Round (5.781) up to n = 6 Stage counter i, (1 to n); 1 i n

Step #3


Using Formula #15 k v ln( v ) Viscosity Constant for the mix


100
k v ln( ) 7.7832
240000

Page 64
 CHAPTER 4

Step #4

Using Formula #16 n p e kvxv Viscosity of the mix

n 240000e ( 7.7832 )(0 .4706 ) 6158.626

Step #5

n
Using Formula #17 a  Relative Viscosity Constant
ln( n 1)

6158.626
a 3164.9077
ln( 6 1)

Step #6

Using Formula #18 i a[ln(i 1)] Viscosity Distribution 1 i n

Viscosity ( i ) = (1) 2193.7468, (2) 3477.0065, (3) 4387.4937,

(4) 5093.7224, (5) 5670.7533, (6) 6158.6260

Step #7


ln( i )
p
Using Formula #26 xvi  % Vehicle Distribution given (1 i n )
kv

xvi (1) 0.6032, (2) 0.5441, (3) 0.5142, (4) 0.4950, (5) 0.4812, (6) 0.4706

xv x p 1 ; x p 1 xv

x p i (1) 0.3968, (2) 0.4559, (3) 0.4858, (4) 0.5050, (5) 0.5188, (6) 0.5294

Page 65
 CHAPTER 4

Step #8

i
1
Bx pi  Pi
Using Formula #19 Pi 1r xv (i111r ) Pigment Distribution for 1 i n
i

Pi (1) 460.1005, (2) 321.4635, (3) 229.6492, (4) 166.8233,

n
(5) 123.0339, (6) 92.0995 P 1393.1700
i Optimized
i
1

Step #9

P
Using Formula #6 Wi  i Pi Water displacement distribution 1 i n
r

Wi (1) 1840.4021, (2) 1285.8539, (3) 918.5970, (4) 667.2934,

(5) 492.1356, (6) 368.3981

Step #10

Using logic, V1 B P1 W1 : V1 3000 460.1005 1840.4021 V1 699 .4973

P 321.4635
V2 B V1  2 : V2 3000 699.4973  V2 233.0848
r 0.20

Step #11

i 1

Using Formula #20 Vi B  ( P i

P
V i )  ri : Vehicle Distribution
i 1

Vi (1) 699.4973, (2) 233.0848, (3) 137.6077, (4) 84.4803,

n
(5) 52.1238, (6) 31.6380 V i 1238.4318 Optimized
i
1

Refer to the model, MatCad MODEL_B in the appendix (A16)

Refer to BASIC Program Reports: MODEL-A2 in the appendix (A11)

Page 66
 CHAPTER 4

BASIC PROGRAM REPORTS

REPORT-MODEL-B

Graph Actual Relative Viscosity vs Stage

Stage Viscosity
1 2191
2 3473
3 4382
4 5087
5 5664
6 6151

Page 67
 CHAPTER 4

MODEL A&B follow mathematical theory growth curve shown in Chapter-1.

Figure 1.01a

Analysis of Vehicle-to-Pigment Ratio:

Stage Viscosity % Pigment Pigment Vehicle Water V-P Ratio

1 2191 0.3966 460 700 1840 1.52
2 3473 0.4558 321 235 1284 0.73
3 4382 0.4857 230 134 920 0.58
4 5087 0.5048 167 85 668 0.51
5 5664 0.5186 123 53 492 0.43
6 6151 0.5292 92 32 368 0.35

This model’s first stage has a vehicle-to-pigment ratio of 1.52, which is

typical of the flush process (Expecting V/P Ratio 1.3 to 1.8). An excess of vehicle

is required in the early stage for proper pigment wetting and oil absorption. The

following stages are grinding stages where the body builds up and the viscosity

increases to the desired value.

Given the above output results and analysis for MODEL-A & B, the

algorithms (A & B) seem to produce output distributions that appear to be

acceptable flush parameters.

Page 68
 CHAPTER 4

Treatment – III-MODEL-C uses the input parameter, Total Pigment Charge

SP(i), to create the pigment distribution, P(i). In this model, the pigment

distribution is a geometric progression, whose sum is equal to the input total

pigment charge, SP(i) . The number of terms in the geometric progression, (n), is

treated as the number of mixing stages in the flush procedure. The viscosity

distribution is an output parameter based on the actual % pigment, xp(i) ,

calculated at each incremental stage (i). The mixer capacity, B, is held constant

through out the procedure. The calculated capacity B(i) , is is an output parameter

and will be listed at each mixing stage to compare to the constant capacity, B. In

this treatment, the allowance, E0, is not required or used.

INPUT DATA OUTPUT DATA

Total Pigment Charge SP(i) Number of mixing stages (n)

Total Vehicle Charge SV(i) % Vehicle after last stage (x n)

Relative Viscosity of the Pigment (hp) System Viscosity Constant (k v)

Relative Viscosity of the Vehicle (hv ) Viscosity Distribution h (i)

% Solids of Presscake (r) % Pigment per stage xp(i)

Pigment Distribution Function f(i) % Vehicle per stage xv(i)

In a Geometric Progression model, Pigment Charge per stage P(i)

Capacity B(i) is Constant for all Vehicle Charge per stage V(i)

stages. (1 < i < n) Water Displacement per stage w d(i)

Total Pigment Charge SP(i)

Total Vehicle Charge SV(i)

Page 69
 CHAPTER 4

Example Problem MODEL-C1 (Non-Optimized):

Given the charge, SP(i) = 1350 lbs. SV(i) = 1200 lbs. Relative Viscosity of

the Pigment (hp ) = 240,000 poise. Relative Viscosity of the Vehicle (h v) = 100

poise. % Solids of Presscake (r) = 20%: (0.20). Capacity Allowance E0 = 15%:

(Eff% = 85% Refer to Formula #22). Optimize the flush procedure to be mixed in

a vessel of Capacity, B = 3000 lbs.

Solution Steps MODEL-C1 (Non-Optimized):

Step #1

Using Formula #11: Calculate the % pigment (xp)in total charge.

 Pi
n
P
i 1
i
1347 . 42
x p  (% i 1
 
(1347 . 42 1202 . 58 ) = 0.5294
Eff )( B ) n

( P
i 1
i V i )

xv x p 1 ; xv 1 x p xv 10.5294 xv 0.4706

Step #2

Using Formula #8: Estimate the number of required mixing stages (n)

(1 E0 ) rp xv (0.85)(0.05294 ) 0.4706

x )

Formula #8 n=  .20
= 5.7813
xv n 0.4706

Round (5.781) up to n = 6 Stage counter i, (1 to n); 1 i n

Step #3

w
Using Formula #22 E 0  n ; wn E0 B wn 450
B

Page 70
 CHAPTER 4

Step #4

wr
Using Formula #21 Pn  n , the Final (nth) Pigment Charge, Pn 112.5
1 r

Step #5

Given the series, a, aR, aR 2 , aR 3 ,..., aR n , the sum of a series function,

a (R n 1)
S , substitute the variables a Pn , (first element in the series),
R 1

n
aR (second element in the series), … , and the sum of the elements, S Pi .
i1

Pn ( R n 1)
n
After substitution, the function is, Pi  . To reverse the order of the
i1 R 1

n
Pn ( R n 1)
series, aR n , aR n 1 , aR n2 ,..., aR , a use P 
i1
i
R 1
Pn . Using an iterative

algorithm to solve for the series variable (R), which is a different variable than the

(r, % solids). Using the Root Solver function in MathCad, R 1.2755 ; n=6. The

pigment Distribution Series is generated by, Pn i1 Pn R i 1 :

i1
P6i 1 (112.5)(1.2755)

P6 = 112.5, P5 = 143.4926, P4 = 183.0233, P3 = 233.4442, P2 =297.7557,

n
P1 = 379.7842 P 1350
i
i

Step #6

P
Using Formula #6 Wi  i Pi Water displacement distribution 1 i n
r

Wi (1) 1519.1370, (2) 1191.0227, (3) 933.7770, (4) 732.0930,

(5) 573.9703, (6) 450.0000

Page 71
 CHAPTER 4

Step #7

Using, V1 B P1 W1 : V1 3000 379.7842 1519.1370 V1 1101.0788

P 297.7557
Using, V2 B V1  2 : V2 3000 379.7842  V2 30.3586
r 0.20

i
1
P
Using Formula #20 Vi B (Pi Vi )  i , given 1 i n ; generates
i
1 r

Vi = (1); 1101.0788, (2); 30.3586, (3); 23.8015, (4); 18.6607, (5); 14.6302,

n
(6); 11.4703 V i
i 1200

Step #8


Using Formula #15 k v ln( v ) Viscosity Constant for the mix


100
k v ln( ) 7.7832
240000

Step #9

Using Formula #16 n p e kvxv Viscosity of the mix

n 240000e ( 7.7832 )(0 .4706 ) 6159.19

Step #10

P i

Using x pi  i
i1
, xvi 1 x p generates,
( P V )
i
i i
i
1

xpi = (1); 0.2585, (2); 0.3745, (3); 0.4409, (4); 0.4824, (5); 0.5101, (6); 0.5294

xvi = (1); 0.7435, (2); 0.6255, (3); 0.5591, (4); 0.5176, (5); 0.4899, (6); 0.4706

Page 72
 CHAPTER 4

Step #10

Using i pe
k v x vi
generates,

hi = (1); 736.0271, (2); 1845.1630, (3); 3092.5231, (4); 4271.3767,

(5); 5299.3304, (6); 6159.1899

Refer to the model, MathCad MODEL_C1 (Non-Optimized) in the appendix (A18)

Analysis of Vehicle-to-Pigment Ratio:

Model-C1

Stage Viscosity % Pigment Pigment Vehicle Water V-P Ratio

1 736 0.2565 380 1101 1519 2.90
2 1845 0.3745 298 30 1191 0.10
3 3093 0.4409 233 24 934 0.10
4 4271 0.4824 183 19 732 0.10
5 5299 0.5101 143 15 574 0.10
6 6159 0.5294 113 11 450 0.10

A high vehicle-to-pigment ratio in the first stage implies excess wetting

which usually results in a very long period for that mixing stage. The displaced

water, excess vehicle and wetted pigment will create a slurry that is very difficult

to separate. This condition is also characterized by the subsequent vehicle

additions being very low or sometimes going negative.

Stage Viscosity
1 736
2 1845
3 3093
4 4271
5 5299
6 6159

Page 73
 CHAPTER 4

The graphic characteristic is a slow rise from the low viscosities to the end

mix viscosity. Too sharp of an increase implies a low vehicle-to-pigment ratio and

aggregation or large pigment clusters result. In summary, any drastic departure

from the exponential growth pattern below, is an indication of abnormality.

Figure 1.01a

Example Problem MODEL-C2 (Optimized):

Given the charge, SP(i) = 1393.1953 lbs. SV(i) = 1238.3959 lbs. Relative

Viscosity of the Pigment (hp ) = 240,000 poise. Relative Viscosity of the Vehicle

(h v) = 100 poise. % Solids of Presscake (r) = 20%: (0.20). Capacity Allowance E 0

= 15%: (Eff% = 85% Refer to Formula #22). Optimize the flush procedure to be

mixed in a vessel of Capacity, B = 3000 lbs.

Solution Steps MODEL-C2 (Optimized):

Page 74
 CHAPTER 4

Step #1

Using Formula #11: Calculate the % pigment (xp)in total charge.

 Pi
n
P
i 1
i
1347 . 42
x p  (% i 1
 
Eff )( B ) n
(1347 . 42 1202 . 58 ) = 0.5294
( P
i 1
i V i )

xv x p 1 ; xv 1 x p xv 10.5294 xv 0.4706

Page 75
 CHAPTER 4

Step #2

Using Formula #8: Estimate the number of required mixing stages (n)

(1 E0 ) rp xv (0.85)( 0.20 ) 0.4706

x 0.5294)

Formula #8 n=  = 5.7813
xv n 0.4706

Round (5.781) up to n = 6 Stage counter i, (1 to n); 1 i n

Step #3

w
Using Formula #22 E 0  n ; wn E0 B wn 368.4088
B
Step #4

wr
Using Formula #21 Pn  n , the Final Pigment Charge, Pn 92.1022
1 r

Step #5

Given the series, a, aR, aR 2 , aR 3 ,..., aR n , the sum of a series function,

a (R n 1)
S , substitute the variables a Pn , (first element in the series),
R 1

n
aR (second element in the series), … , and the sum of the elements, S Pi .
i1

n
Pn ( R n 1)
After substitution, the function is, 
i1
Pi 
R 1
. To reverse the order of the

n
Pn ( R n 1)
n
series, aR , aR n
1
, aR n2
,..., aR , a use Pi  Pn . Using an iterative
i1 R 1

algorithm to solve for the series variable (R), which is a different variable than the

(r, % solids). Using the Root Solver function in MathCad, R 1.2755 ; n=6. The

pigment Distribution Series is generated by, Pn i1 Pn R i 1 :

P6 i1 (92.1022)(1.3689) i1

Page 76
 CHAPTER 4

P6 = 92.1022, P 5 = 126.0796, P4 = 172.5917, P3 = 236.2626, P2 =323.4222,

n
P1 = 442.7360 P 1393.1943
i
i

Step #6

P
Using Formula #6 Wi  i Pi Water displacement distribution 1 i n
r

Wi (1) 1770.9440, (2) 1293.6889, (3) 945.0502, (4) 690.3668,

(5) 504.3186, (6) 368.4088

Step #7

Using, V1 B P1 W1 : V1 3000 442.7360 1770.9440 V1 786.3201

P 323.4222
Using, V2 B V1  2 : V2 3000 442.736  V2 153.8328
r 0.20

i
1
P
Using Formula #20 Vi B (Pi Vi )  i , given 1 i n ; generates
i
1 r

Vi = (1); 786.3201, (2); 153.8328, (3); 112.3761, (4); 82.0917,

n
(5); 59.9686, (6); 43.8076 V i 1238.3969
i

Step #8


Using Formula #15 k v ln( v ) Viscosity Constant for the mix


100
k v ln( ) 7.7832
240000

Step #9

Using Formula #16 n p e kvxv Viscosity of the mix

Page 77
 CHAPTER 4

n 240000e ( 7.7832 )(0 .4706 ) 6159.19

Step #10

P i

Using x pi  i
i1
, xvi 1 x pi generates,
( P V )
i
1
i i

xpi = (1); 0.3602, (2); 0.4490, (3); 0.4878, (4); 0.5087, (5); 0.5213, (6); 0.5294

xvi = (1); 0.6398, (2); 0.5510, (3); 0.5122, (4); 0.4913, (5); 0.4787, (6); 0.4706

Step #11

Using i pe
k v x vi
generates,

hi = (1); 1650.5725, (2); 3294.3023, (3); 4455.4729, (4); 5244.0024,

(5); 5784.0311, (6); 6159.1715

Refer to the model, MathCad MODEL_C2 (Optimized) in the appendix (A20)

Analysis of Vehicle-to-Pigment Ratio:

Model-C2

Stage Viscosity % Pigment Pigment Vehicle Water V-P Ratio

1 1651 0.3602 443 786 1771 1.77
2 3294 0.4490 323 154 1294 0.48
3 4455 0.4878 236 112 945 0.47
4 5244 0.5087 173 82 690 0.47
5 5784 0.5213 126 60 504 0.48
6 6159 0.5294 92 44 368 0.48

Page 78
 CHAPTER 4

This model’s first stage has a vehicle-to-pigment ratio of 1.77, which is

within the acceptable range (V/P Ratio 1.3 to 1.8). Comparing the ratio to

MODEL-A & MODEL-B, the expectation would be an adequate, but slightly

longer mixing period in the first wetting stage.

Given the above output results and analysis for MODEL-C2, the algorithm

(A & B) seems to produce the best output distributions if the pigment and vehicle

input charges are optimized.

Page 79
 CHAPTER 5

SUMMARY, CONCLUSION AND RECCOMENDATIONS

Restatement of the Problem:

Quantifying the flush process is the primary focus of this project. By using

the general description of the flushing process, models can be created to

simulate the procedure. The three models, (MODEL-A, MODEL-B, MODEL-C),

uses bulk load capacity and viscosity as the major constraints to produce the

number of mixing stages and the component quantities that are required to

optimize the mixing process.

Since the vehicle-pigment ratio for each mixing stage is usually

determined by experimental methods in a laboratory environment, the main

objective in this project is to create some models and methodologies that

simulate this experimental process. If this objective is met, then this methodology

could possibly serve as a design and analysis tool which will increase the

productivity of the technologist.

SUMMARY

MODEL-A and MODEL-B

These two models are essentially the same in that they produce the same

output. The difference between the two models is that MODEL-B uses the end

mix pigment content and the mixer capacity to calculate the initial (non-

optimized) quantities. MODEL-B is more efficient in that it requires one less input

parameter while producing the same output. The end mix pigment content is also

Page 80
 CHAPTER 5

a constant parameter, in the since that its value remains the same from (non-

optimized) input to (optimized) output. Refer to the program reports below.

(MODEL-A & MODEL-B)

REPORT-MODEL-A

REPORT-MODEL-B

Page 81
 CHAPTER 5

Using Formula #11: Calculate the % pigment (xp)in total charge in MODEL-A is:

n
n
 Pi P
i 1
i
1350
x p  (% i 1
 
(1350 1200 ) = 0.52941,
Eff )( B ) n

( P
i 1
i V i )

which is the same as the INPUT % Pigment Charge xp in MODEL-B. Also note

the (non-optimized) INPUT Pigment Charge = 1349.996 & Vehicle Charge =

1200.005, are calculated by the program. Both program OUTPUT distributions

are very nearly identical.

The OUTPUT distributions of the two models resemble the beaker-spatula

mixing example shown below from Chapter-1 as Figure 1.02

Figure 1.02 (Flush Sequence)

Page 82
 CHAPTER 5

MODEL-C1 & C2 (Geometric Series)

These two models, MODEL-C1 and MODEL-C2, use the geometric

function to determine the pigment content distribution, xp(i), instead of the

exponential growth function used in MODEL-A & B. The MODEL-C series does

not optimize the input charges, but creates the output distributions based on the

bulk capacity, B, and the allowance, E. The sum of the output charge distribution

is always equal to the total input charge.

n n
INPUT : P; V equals OUTPUT : Pi ; Vi
i1 i1

For comparative purposes, MODEL-C1 was created to use the (Non-

Optimized) pigment and vehicle charges that are inputs in MODEL-A & B, and

MODEL-C2 was created to use the (Optimized) charges that are outputs from

MODEL-A & B. Refer to the analysis of MODEL-C1 and MODEL-C2.

CONCLUSION

The output distributions generated form the flush models show that the

empirical derivations and implied relationships are accurate enough to serve as a

general outline for more complex models, which will provide further in-depth

analysis. I am certainly convinced that it is possible to model the flush procedure

with mathematical algorithms. There is much room for expansion of the models

to include more useful constraints such as temperature, time and energy

requirements. More detail design is needed prior to committing laboratory labor

Page 83
 CHAPTER 5

and equipment to correlate and test the theoretical results to real dispersion

procedures.

As the project progressed into the analysis and summary phase, more

questions than answers were generated. I plan to continue working on this

project by fine-tuning the models and programs to be user-friendlier. There are

so many conditions, which need to be analyzed, but time and project format

constraints do not permit this at this time. I am very please with the development

of the mathematical logic and procedures, because the math is the foundation of

the modeling efforts.

RECOMENDATIONS

This phase of the project focused on the end result of the mixing (flushing)

stages and can best be characterized as empirical. The next phase of this project

is to do further analysis by testing more input conditions. There seems to be

some input values that will generate errors in the program during processing.

This needs to be investigated. The algorithm, which estimates the number of

required, mixing stages, needs more development.

Other conditions that were not addressed were the use of multiple input

vehicles and pigments of various viscosities. The use of solvents that evaporate

and agents that serve as catalysts, are also potential development

enhancements.

Page 84
 CHAPTER 5

The related literature shows that temperature, shearing, pigment

absorption rates, evaporation rates, particle size and mixing speed are just some

of the many parameters that are directly related to the energy of mixing. The

action that occurs between the mixing stages is the most important part of the

flushing procedure and will require more detail treatment in the subsequent

phases. These and all of the items that were mentioned above are the

recommendations for future expansion and development. After all, the action that

occurs between the mixing stages is what is called “flushing.”

Page 85
 BIBLIOGRAPHY

Atherton, D., Hedley, B, Greaves, J., Marks, S., Martin, S. & Smith, M.
(1961). Paint Technology Manuals: PART TWO - Solvents, Oils
, Resins and Driers: Published on behalf of The Oil & Colour
Chemists' Association

Atkins, P. W. (1982). Physical Chemistry. (2nd ed.). New York, San

Francisco: W. H. Freeman and Company

Brothman, A, Wollan, G, & Feldman, S. (1947). Manual of Chemical

Engineering Calculations & Shortcuts: New Analysis Provides
Formula to Solve Mixing Problems. New York: McGraw-Hill
Publishing Co., Inc.. Page 175

Giles, R. V. (June, 1962). Fluid Mechanics and Hydraulics: Schaum's

Outline Series. (2nd ed.). New York, St. Louis, San Francisco,
Toronto, Sidney: McGraw-Hill Book Company

Kreyszig, E. (August 1988) Advanced Engineering Mathematics. (6th ed.).

New York, Chichester, Brisbane, Toronto, Singapore: John Wiley &
Sons

McKennell, R., Ferranti Ltd., Moston & Mancheser. (1960). A Reprint from
the "Instrument Manual", 1960, Section XI: Ferranti Instrument
Manual: The Measurement and Control of Viscosity And Related
Flow Properties.

Patton, T. C. (1963). Paint Flow and Pigment Dispersion. (1st ed.). New
York, Chichester, Brisbane, Toronto: John Wiley & Sons.

Patton, T. C. (1979). Paint Flow and Pigment Dispersion. (2nd ed.). New
York, Chichester, Brisbane, Toronto: John Wiley & Sons.

Stroud, K.A., (2001). Engineering Mathematics. (5th ed.). New York:

Industrial Press, Inc.

Turner, G. P. A (1967). Introduction to Paint Chemistry. London: Chapman

and Hall

Wolfe, H. J. (1967). Printing and Litho Inks. New York City: MacNair-
Dorland Company

Page 86
 APPENDIX

FLUSH MODEL FORMULAE

#1 r (0 < r < 1) :constant Solids content of aqueous pigment

n
Pi
#2 r ∑ P +W
i =1
Calculation of solids content
i i

#3 i ( 0 ≤ i ≤ n) :integer Incremental flush stages

#4 P Pi Pigment charge at stage, (i)

#5 V Vi Vehicle charge at stage, (i)

#6 W Wi Water displacement at stage, (i)

n n
#7 Wn = B − ( ∑ Pi + ∑ V ) i Water displacement at last stage, (n)
i =1 i =1

n n
(1 / r )( ∑ P + ∑ Vi )
i =1 i =1
#8 n= Calculation of the number of stages
x pn B
required to flush the total charge of pigment
and vehicle

n
#9 ∑ (P +V )
i =1
i i
Total charge after water displacement

∑ (P + V )
i =1
i i
#10 B= Bulk capacity or working mixer capacity
% Eff
at % Effective (~ 85%; Decimal)

n
∑ Pi
#11 xp = i =1
(% Eff )( B ) % pigment in total charge

A1
 APPENDIX

FLUSH MODEL FORMULAE

#12 x p = 1 − xv % pigment in total charge

n
∑Vi
#13 xv = i =1
(% Eff )( B ) % vehicle in total charge

#14 xv = 1 − x p
η
#15 k v = ln( ηvp ) Viscosity constant in the Exponential
Viscosity Distribution

#16 ηn = ηp e k x v v
Relative End-Viscosity of the mix at stage (n)

ηn
#17 a= ln( n +1) Relative Viscosity Distribution Constant

#18 ηi = a ln(i + 1) Viscosity Distribution Function 1≤ i ≤ n

i −1
Bx pi − ∑ Pi
#19
Pi = r + xvi
1
i =1
(1− 1r ) Pigment Charge Distribution given 1≤ i ≤ n

A2
 APPENDIX

FLUSH MODEL FORMULAE

i−1
Vi =B−∑(Pi +Vi )− ri
P
#20 Vehicle Charge Distribution given 1≤ i ≤ n
i=1

wnr
Pn = Pn
= Pn +wn
#21
1− r Final (nth) Pigment Charge given r

E0 =
wn
#22 B Allowance = (100 - % Effective)

#23 PW + V → PV + W Physical reaction of presscake (PW),

mixing with vehicle, (V), to produce a
paste of wetted pigment, (PV), and displaced
water, (W).
i −1
#24 B ≥ ∑ ( P + V ) i + PWi + Vi Expression of capacity (B), before
i =1
mixing and water displacement.

i −1
#25 B ≥ ∑ ( P + V ) i + PVi + Wi Expression of capacity (B), after
i =1
mixing and water displacement.

A3
 APPENDIX

BASIC PROGRAMS
MODEL-A

REM Pigment Distribution Bsaed On Viscosity Function Algorithm

REM Created by Herb Norman Sr. for Mixer Problem Project (MODEL-A)
REM 03/02/2005 - MODEL_A.BAS
REM*******************************************************
CLS
REM Input Parameters
REM =================
INPUT "Mixer Capacity (B) .............. B ="; B
INPUT "Total Pigment Charge (P) ........ P ="; P
INPUT "Total Vehicle Charge (V) ........ V ="; V
INPUT "% Solids of Pigment (r) ......... r ="; r
INPUT "Vehicle Viscosity (nv) ......... nv ="; nv
INPUT "Pigment Viscosity (np) ......... np ="; np
INPUT "Prior Residual (W) .............. W ="; W
INPUT "W Vehicle Content [xv(0)] ... xv(0) ="; xv(0)

REM B = 3000
REM P = 1350
REM V = 1200
REM r = .2
REM nv = 100
REM np = 240000
REM W = 0: xv(0) = 0
REM xv(0) = 0

REM Calculate Constants

REM ===================
P(0) = W * (1 - xv(0))
V(0) = W * xv(0)
kv = LOG(nv / np)
xv = V / (P + V): xp = (1 - xv)
nmix = np * EXP(kv * xv)
n = INT((P / r + V) / (xv * B) + .5)
n0 = (P / r + V) / (xv * B)
a = nmix / (LOG(n + 1))

REM Calculate Viscosity Distribution n(j)

A4
 APPENDIX

BASIC PROGRAMS
MODEL-A
(Continued)

REM =====================================
PRINT "# "; "Viscosity", "% Pgmt", "Pigment", "Vehicle", "Water"
PRINT "== "; "=========", "======", "=======", "=======", "====="

FOR j = 1 TO n
n(j) = INT(a * LOG(j + 1))
xv(j) = INT(((LOG(n(j) / np)) / kv) * 10000 + .5) / 10000
xp(j) = 1 - xv(j)
K1 = K1 + P(j - 1): K2 = K2 + V(j - 1)
P(j) = INT((B * xp(j) - K1) / ((1 / r) + xv(j) * (1 - 1 / r)) + .5)
V(j) = INT(B - (K1 + K2) - P(j) / r + .5)
wd(j) = P(j) * (1 / r - 1)
SumP = SumP + P(j): SumV = SumV + V(j): SumW = SumW + wd(j)
PRINT j; n(j), xp(j), P(j), V(j), wd(j)
NEXT j

PRINT : PRINT "Pigment Charge (INPUT)"; P

PRINT "Vehicle Charge (INPUT)"; V
PRINT "Sum of Pigment Charge"; SumP
PRINT "Sum of Vehicle Charges"; SumV
PRINT "Sum of Water Displacement"; SumW
PRINT "Original n .... n0"; n0
REM PRINT SumP, SumV, SumW

A5
 APPENDIX

BASIC PROGRAMS
MODEL-B

REM Pigment Distribution Bsaed On Viscosity Function Algorithm

REM Created by Herb Norman Sr. for Mixer Problem Project (MODEL-B)
REM 03/08/2005 - MODEL_B.BAS
REM*******************************************************
CLS
REM Input Parameters
REM =================
INPUT "Mixer Capacity (B) .............. B ="; B
INPUT "% Pigment Charge (xp) ...........xp ="; xp
INPUT "% Solids of Pigment (r) ......... r ="; r
INPUT "Vehicle Viscosity (nv) ......... nv ="; nv
INPUT "Pigment Viscosity (np) ......... np ="; np
INPUT "Prior Residual (W) .............. W ="; W
INPUT "W Vehicle Content [xv(0)] ... xv(0) ="; xv(0)

REM B = 3000
REM P = 1350
REM V = 1200
REM r = .2
REM nv = 100
REM np = 250000
REM W = 0: xv(0) = 0
REM xv(0) = 0

REM Calculate Constants

REM ===================
xv = 1 - xp
P = B * .85 * xp
V = B * .85 - P
P(0) = W * (1 - xv(0))
V(0) = W * xv(0)
kv = LOG(nv / np)
nmix = np * EXP(kv * xv)
n = INT((P / r + V) / (xv * B) + .5)
n0 = (P / r + V) / (xv * B)
a = nmix / (LOG(n + 1))

A6
 APPENDIX

BASIC PROGRAMS
MODEL-B
(Continued)

REM Calculate Viscosity Distribution n(j)

REM =====================================
PRINT "# "; "Viscosity", "% Pgmt", "Pigment", "Vehicle", "Water"
PRINT "== "; "=========", "======", "=======", "=======", "====="

FOR j = 1 TO n
n(j) = INT(a * LOG(j + 1))
xv(j) = INT(((LOG(n(j) / np)) / kv) * 10000 + .5) / 10000
xp(j) = 1 - xv(j)
K1 = K1 + P(j - 1): K2 = K2 + V(j - 1)
P(j) = INT((B * xp(j) - K1) / ((1 / r) + xv(j) * (1 - 1 / r)) + .5)
V(j) = INT(B - (K1 + K2) - P(j) / r + .5)
wd(j) = P(j) * (1 / r - 1)
SumP = SumP + P(j): SumV = SumV + V(j): SumW = SumW + wd(j)
PRINT j; n(j), xp(j), P(j), V(j), wd(j)
NEXT j

PRINT : PRINT "Pigment Charge (INPUT)"; P

PRINT "Vehicle Charge (INPUT)"; V
PRINT "Sum of Pigment Charge"; SumP
PRINT "Sum of Vehicle Charges"; SumV
PRINT "Sum of Water Displacement"; SumW
PRINT "Original n .... n0"; n0
REM PRINT SumP, SumV, SumW

A7
 APPENDIX

BASIC PROGRAMS
MODEL-C

REM Iteration to find Ri factor in the Geometric Series

REM Calculate Compoment Distribution & Relavive Viscosity Distribution
REM Created by Herb Norman Sr. for Mixer Problem Thesis
REM QBASIC PROGRAM - 03/15/2005
REM 1st Draft Thesis - 07/04/2005
REM Mathematical Model For A Mixing Optimizing Algorithm
REM With Aqueous Displacement And Extraction
REM Using Relative Viscosity And Mixer Capacity
REM As the Primary Physical Constraints
REM - An Application Of Geometric Series Distributions
REM************************************************************************
CLS
INPUT "Mixer Capacity (B) .............. B ="; B
INPUT "Total Pigment Charge (P) ........ P ="; P
INPUT "Total Vehicle Charge (V) ........ V ="; V
INPUT "% Solids of Pigment (r) ......... r ="; r
INPUT "Vehicle Viscosity (nv) ......... nv ="; nv
INPUT "Pigment Viscosity (np) ......... np ="; np

e = 2.7183
ex = .5
np = 240000!
nv = 100
kv = LOG(nv / np)
xp = P / (P + V)
n = INT(((P / r + V) / (xp * B)) + .5)
pn = (r / (1 - r)) * (B - (P + V))

REM *** Start Iteration - Solve for (Ri) & Final Pigment Charge P(n)
FOR x = 1 TO 1000
Ri = 1 + (x / 1000)
y = (pn * ((Ri ^ n) - 1)) / (Ri - 1)
IF y >= (P - ex) THEN
IF y <= (P + ex) THEN
Rx = Ri
END IF
END IF
NEXT x

A8
 APPENDIX

BASIC PROGRAMS
MODEL-C
(Continued)

REM *** Setup for Pigment Disribution

px = pn
P(n) = pn
xp(n) = P / (P + V): xv(n) = 1 - xp(n)
nj(n) = np * e ^ (kv * xv(n))
tp = tp + pn

CLS
PRINT "Total Pigment Charge ... (P) ="; P, "Total Vehicle Charge (V) ="; V
PRINT "Mixer Capacity ......... (B) ="; B, "% Solids of Pigment (r) ="; r
PRINT "Calculated Series Ratio (Ri) ="; Rx, "Pigment Viscosity (Np) ="; np
PRINT "Number of Mixing Stages (n) ="; n, "Vehicle Viscosity t (Nv) ="; nv

FOR s = 1 TO n - 1
px = px * Rx
P(n - s) = px
tp = tp + px
xp(s) = (tp / (P + V)): xv(s) = 1 - xp(s)
nj(s) = np * e ^ (kv * xv(s))
NEXT s
V(0) = 0
P(0) = 0
tp = 0
tv = 0
PRINT
PRINT " j"; " Pigment ", " Vehicle ", "Cum Pigment", "Cum Vehicle", " Pour-Off"
PRINT "=="; " =========", " ==========", "===========", "===========",
"========="

FOR s = 1 TO n
tp = tp + P(s - 1)
tv = tv + V(s - 1)
wd(s) = (P(s) / r) - P(s)
V(s) = B - (tp + tv + P(s) / r)
PRINT s; P(s), V(s), tp + P(s), tv + V(s), wd(s)
NEXT s

A9
 APPENDIX

BASIC PROGRAMS
MODEL-C
(Continued)

REM *** Pigment & Viscosity Distribution

PRINT
PRINT " j"; " Pigment ", " Vehicle ", " % Pigment ", " % Vehicle ", "Viscosity"
PRINT "=="; " =========", " ==========", "===========", " ==========",
"========="

FOR s = 1 TO n
PRINT s; P(s), V(s), xp(s), xv(s), nj(s)
NEXT s

BEEP: BEEP: PRINT

INPUT a$
SYSTEM

A 10
 APPENDIX

BASIC PROGRAM REPORTS

REPORT-MODEL-A1

REPORT MODEL-A2

A 11
 APPENDIX

BASIC PROGRAM REPORTS

REPORT MODEL-B1

REPORT MODEL-B2

A 12
 APPENDIX

BASIC PROGRAM REPORTS

REPORT MODEL-C1

REPORT MODEL-C2

A 13
A 14
A 15
A 16
A 17
A 18
A 19
A 20
A 21
 APPENDIX – B

FLUSH FORMULA DERIVATIONS

Solids Content of Pigment Presscake (r)

Formula #2
n
Given the total pigment charge, P , the presscake is;
i
i1
n

P i n n
i
1
Pi Wi
r i 1 i
1
n

P i

n
i
1
r
( P W )
i 1
i i

Water Displacement at stage (i)

Formula #6
Pi
Given the charge of presscake at stage (i),
r
Pi
Wi Pi
r
Pi
Pi Wi
r
1
Pi ( )
1 Wi
r

B1
 APPENDIX – B

Water Displacement (W) at stage (n)

Formula #7
n
1
Given the capacity, B: B ( Pi Vi ) Pn Wn Vn
i
1
n
B ( Pi Vi ) Wn
i
1
n n
Wn B (Pi Vi )
i
1 i1

Bulk Capacity, Allowance & % Effective (B), (E0) & (%Eff)

Formula #10
n
Given capacity, B ( Pi Vi ) Wn and %Eff 1 Eo )
i
1
n n

 ( Pi Vi ) ( Pi Vi )
B i
1
 i1
%Eff 1 E 0
n
(% Eff )B ( Pi Vi )
i
1
n

( P V ) i i

%Eff  i1 1 E0

B
n

(P V ) %Eff ( B ) (1 E ) B

i i 0
i
1

% Pigment in Total Charge xp

Formula #11
n n n
Given Total Charge, ( Pi Vi ) Pi Vi
i1 i
1 i
1
n n

P i P i
xp  n i
1
 i1

% Eff ( B )
( P V )
i
1
i i

B2
 APPENDIX – B

% Vehicle in Total Charge xv

Formula #13
n n n
Given Total Charge, ( P V ) P V
i i i i
i1 i
1 i
1
n n

V i V i
xv  n
i1
 i 1
%Eff ( B)
(P V ) i i
i
1

% Pigment (xp) & % Vehicle (xv)

Formula #12
n
Given the total charge, ( P V )i i
i1
n n n n

Pi  Vi
P i Vi
n
i1
n i1
 n
i1 i1
1
(Pi Vi ) ( Pi Vi ) (Pi Vi )
i
1 i
1 i
1

x p x v 1
x p 1 x v
x v 1 x p
Pigment Charge Distribution P(i)
Formula #19
n 1
P
Given capacity B: B (Pi Vi )  n Vn
i1 r
i1
P
B (Pi Vi )  i Vi
i1 r
i
1
P
B (Pi Vi )  i Vi
i
1 r
i
1
P
Vi B ( Pi V i )  i
i
1 r

B3
 APPENDIX – B

Pigment Charge Distribution P(i) (Continued)

j1

V i V j
xv j  i1
n

( P V )
v

i i
i
1

j  j1
xv j i i( P  V ) Vi V j
i 1  i1
j  j 1
xv j i i
( P  V ) Vi V j
i1  i 1
Equation Set:
i
1
P
Vi B (Pi Vi )  i
i
1 r
j  j1
V j xv j i i
( P V ) Vi
i1  11

V j V j
j
1
Pj j 1 j1
Pj  j 1
B ( Pi Vi )  xv j ( Pi V i ) Pj B  ( Pi Vi )  Vi
i1 r i1 i1 r  i 1
j
1
P  Pj  j 1
B (Pi Vi )  j xv j Pj  B Vi
i1 r  r  i 1
j1 j1
Pj  Pj  j1
B Pi Vi  xv j  Pj  B Vi
i
1 i
1 r  r  i1
j j j
1 1
P 1
 1 
B Pi Vi  j V i xv j  Pj (1  ) B 
i
1 i1 r i1  r 
j1
Pj 1
B Pi  xv j Pj (1  ) xv j B
i1 r r
j1
P 1
B xv j B Pi  j xv j Pj (1  )
i 1 r r
j
1 1 
1
B(1 xv j ) Pi Pj  xv j (1  )
i1 r r 

B4
 APPENDIX – B

Pigment Charge Distribution P(i) (Continued)

j1
(B ) xv j Pi
i
1
Pj
1 1 
r xv j (1 r )
 

Vehicle Charge Distribution V(i)

Given: xp j 1 xv j
j1
P
Then: V j B ( Pi Vi )  j
i
1 r

B5