ABSTRACT

VELMURUGAN, MURALI. The Role of Fiber Finish in the Conversion of Fiber to Nonwovens.
(Under the direction of Dr. Abdelfattah M. Seyam)

The performance of fiber finish as a mechanical processing aid is assessed with regards to two of the
more highly demanding mechanical processes in the nonwovens industry namely, carding and
needlepunching. The main finish characteristics under study here are finish level (add-on percentage)
and finish uniformity on a macro level. The thesis also aims to study any possible interaction effect of
finish performance with the process variables. The first chapter contains critical review of previous
work done to characterize frictional behavior of fibers and its relevance to asses finish performance.

Chapter two describes the experimental design adopted for investigating finish performance on the
carding machine. Three different fibers are used to study the effect of finish performance and their
interaction with process variables (such as card throughput and cylinder to doffer setting) on key
response parameters like fiber breakage, fiberweb cohesion, nep formation and static charge
generation. The results indicate a significant effect of finish add-on percentage on most of the
response parameters. On the other hand, finish uniformity was found to have less influence on the
response parameters. It has also been found that the influence of finish add on percentage and finish
uniformity varied with the type of fiber used.

In chapter three, the performance of finish in needlepunching and its influence on needlepunched
fabric properties and needling force parameters are investigated. Results indicate a very significant
influence of finish add-on percentage and finish uniformity on the needlepunched fabric properties
and forces experienced by needles during needlepunching. The major conclusions of this project and
the scope for future work are discussed in Chapter four.

Appendix one consists of the result of the finish extraction tests conducted on cotton fibers. The
statistical models and their results are given in appendices two and three for the tests conducted on
chapters two and three respectively. Appendix 4 contains a detailed description of the method used
for analyzing the needling forces studied in chapter three with an example.

i
THE ROLE OF FIBER FINISH IN THE CONVERSION OF FIBER
TO NONWOVENS.
by

MURALI VELMURUGAN

A dissertation submitted to the Graduate Faculty of

North Carolina State University
In partial fulfillment of the
requirements for the Degree of
Master of Science

TEXTILE MANAGEMENT TECHNOLOGY

Raleigh

1999

ii
 ACKNOWLEDGEMENT

I would like to express my sincere gratitude to my advisor, Dr. Abdelfattah M. Seyam, for patiently
guiding me through this project. I would also like to thank Dr. Thomas Kepler for his time and
advice in preparing the statistical models used in this project. I also would like to express my
gratitude to Mr. Thomas J. Proffitt for his useful ideas and support throughout this project. Special
thanks to Dr. William Oxenham , Dr. Subash K. Batra and Dr. Don Shiffler for their advice and
support.

I would also like to thank The Nonwovens Cooperative Research Center for extending financial
support and other facilities to this project. I would like to thank BBA Nonwovens, Fiber Vision Inc.,
and Wellman Inc., for supplying the fibers and Goulston Technologies for supplying the finish used in
this project. I would like to thank Hollingsworth on Wheels Inc. and Lawson-Hemphill, SC for
extending their research facilities for this project. Special thanks to Mr. Chuck Allen of INDA, Mr.
Robert Lanigan of Fiber Vision, Mr. Legrande Crook of Hollingsworth, Mr. Tom Theyson, Dr.
Michael Kutsenko and Mr. Steve Kokenes of Goulston tech, Mr. Robert Avrell of Wellman, and Ms.
Kay McCoy of Henkel Corp, for serving as industrial advisors during the course of this research
work.

I would also like to thank Mr. Mac Mclean, Mr. Sherwood Wallace, Mr. Olin Stewart, Mr. Jim
Watson, Mr. David Curry and Mr. Frank Duek for their assistance during my experimental work. I
also would like to thank my colleagues Hyungsup Kim and Ayad Oumera for their help and support.
I would also like to thank Ms. Janey Bullock and her staff for all their help.

Finally, I would like to thank my parents and sister for their constant love, support and
encouragement.

iii
TABLE OF CONTENTS

LIST OF TABLES……………………………………………...……………….… vi
LIST OF FIGURES…………………………………………………...…………... vii

Chapter 1………..…….……………………………………………….…………... 1
Abstract…………………………………………….………………….…………… 1
1. Introduction………………………………………………….…...……………... 2
2. The Laws of Friction……………………………………………………….…… 2
3. Effect of Fiber and Finish Properties on Frictional Forces…...……………... 4
4. Measurement of Friction……………………………………………..………… 4
4.1. Point Contact Method………….…….………………..……...…….……..…… 6
4.2. Line Contact Method…………………………………………………..…….… 6
4.3. Rotorring®……………………………………………………………………… 8
5. Static Charge……………………………………………………………………. 9
6. Significance of Finish to Carding………………………………………..…….. 11
7. Significance of Finish to Needlepunching……………………………………... 12
8. Shortcomings of conventional method of testing fiber friction to evaluate
finish performance in carding and needlepunching………………………….. 12
References……………………………………………………...………….………. 13

Chapter 2….…..…….……………………………………………………………... 15
Abstract………………………………………………………………………….…. 15
1. Introduction……………………………………………………………………... 16
2. Objectives……………………………………………………………………….. 16
3. Experimental Approach ………………………….…….…..…….…….……… 16
3.1. Measurement of Fiber Breakage…………………………………………...…... 17
3.2. Measurement of Web Uniformity……………………………………………… 17
3.3. Measurement of Neps……………………………………………………….…. 18
3.4. Measurement of Web Cohesion………………………………………………... 19
3.5. Measurement of Static Charge……………………..…………………………... 19
4. Processing Parameters………………………………………………………….. 19
4.1. Cotton…………………………………………..………………………….… 19
4.2. Polypropylene …………………………………………………………………. 20
4.3. Polyester ……………………………………………………………………..… 22
5. Results and Discussion………………………………………………………….. 23
5.1. Cotton ..……………………………………………………………………… 23
5.1.1. Fiber Breakage………………………………………………………….……. 23
5.1.2. Neps …………………………………………………………………………. 24
5.2. Polypropylene………………………………………………………………….. 25
5.2.1. Fiber Breakage………………………………………………………...……... 25
5.2.2. Fiberweb Uniformity………………………………………………………… 25
5.2.3. Fiberweb Cohesion……………………………………………………..……. 26
5.2.4. Neps and Static Charge………………………………………………..…….. 29

iv
5.3. Polyester………………………………………………………………………... 30
5.3.1. Fiber Breakage………………………………………………………….……. 30
5.3.2. Fiberweb Uniformity………………………………………………….……… 32
5.3.3. Fiberweb Cohesion…………………………………………………...……… 32
5.3.4. Neps…………………………………………………………………..……… 32
5.4. Comparison Between Fiber Behavior……………………………………….. 35
6. Conclusion……………………………………………………………………….. 37
References………………………………………………………………………….. 37

Chapter 3………..…….…………………………………………………………… 38
Abstract………………………………………………………………………….…. 38
1. Introduction……………………………………………………………………... 39
2. Objectives……………………………………………………………………….. 39
3. Experimental Approach ………………………….…….…..…….…….……… 39
3.1. Dynamic Needling Forces……………………………………………………… 40
3.2. Fabric Properties……………………………………………………………….. 43
4. Results and Discussion……..…………………………………………………… 43
4.1. Polypropylene…………………………………………………………….……. 43
4.1.1. Dynamic Needling Forces…………………………………………………… 43
4.1.2. Fabric Properties……………………………………………………….…….. 44
4.2. Polyester………………………………………………………………………... 48
4.2.1. Dynamic Needling Forces……………………………………………..…….. 48
4.2.2. Fabric Properties…………………………………………………………….. 49
5. Conclusion……………………………………………………………………….. 52
References………………………………………………………………………….. 52

Chapter 4…………………………………………………………………………… 54
Conclusion…………………………………………………………………………. 54
Scope for future work……………………………………………………………... 54

Appendices 55
Appendix I : Finish Extraction Results for Cotton.………………...…………… 56
Appendix II : Statistical Results of Response Parameters of Carding
Experiments Performed in Chapter 2……………...……………. 57
Appendix III : Statistical Results of Response Parameters of Needlepunching
Experiments Performed in Chapter 3………………………….. 75
Appendix IV : Method Used for Signal Analysis of Needling Forces……….…. 87

v
LIST OF TABLES

Chapter 1
Table 1: Influence of Fiber and Finish Properties on Frictional Force ……...……. 4

Chapter 2
Table 1: Card specifications for cotton…………………………………………….. 19
Table 2: Process parameters used for cotton………………………………………. 20
Table 3: Card specifications for polypropylene…………………………………… 21
Table 4: Blend compositions for polypropylene…………………………………... 21
Table 5: Process parameters used for polypropylene ………………………..……. 21
Table 6: Finish Level and blend composition of polyester fiber…………………... 22
Table 7: Process parameters used for polyester……………………………………. 22

Chapter 3
Table 1: Needlepunching process parameters …………………………………….. 40

Appendix IV
Table 1: Regression analysis for calibration of strain gauges……………………... 87
Table 2: Time series model used for fitting curve…………………………………. 89
Table 3: Force equation of needles and their R2 values…………………………… 89
Table 4: Peak penetration forces for polypropylene with 0.61% finish level……... 91
Table 5: Penetration energy for polypropylene with 0.61% finish level…………... 91
Table 6: Matlab Script for removing inertia forces from actual forces and
calculating the peak penetration force and penetration energy of each
needle…………………………………………………………………….. 92

vi
LIST OF FIGURES

Chapter 1
Figure 1: Frictional Behavior of Liquid Lubricated Yarn…………………………... 4
Figure 2: Schematic diagram of Howell’s capstan method……………………..….. 6
Figure 3: Schematic diagram of Roedger’s Apparatus……………………….…….. 7
Figure 4: Schematic diagram of the Sikorski friction measuring device…….……... 8
Figure 5: Schematic diagram of Lord’s Apparatus….……………………………… 8
Figure 6: Schematic diagram of Beard Friction test………………………………... 9
Figure 7: Block diagram of non-contacting voltmeter………………….…………... 10
Figure 8: Two-conductor capacitance problem showing conductor capacitance to
ground C1, NCV probe capacitance to ground C2, and mutual
capacitance CM………………………………………………………….. 10

Chapter 2
Figure 1: Schematic diagram showing the mounting and data acquisition system of 18
the IR-based photometric devices…………………………..……………..
Figure 2: Calibration curve for cotton………………………………………………. 20
Figure 3: Calibration curve for polypropylene……………………………………... 22
Figure 4: Calibration curve for polyester.……………………………………..……. 23
Figure 5: Change in effective fiber length of cotton w.r.t. finish level…..…………. 24
Figure 6: Change in short fiber content of cotton w.r.t. finish level…..……………. 24
Figure 7: Change in nep count of cotton w.r.t. finish level ..….……………………. 25
Figure 8: Change in 2.5% span length of polypropylene fiber w.r.t. finish level…... 26
Figure 9: Change in short fiber content of polypropylene fiber w.r.t. finish level…. 26
Figure 10: Change in 2.5% span length of polypropylene fiber w.r.t. finish
uniformity…..………………………………………………………… 27
Figure 11: Change in short fiber content of polypropylene fiber w.r.t. finish
Uniformity………………………………………………………………. 27
Figure 12: Change in CV % of fiberweb basis weight of polypropylene fiber w.r.t.
finish level……………………………………………………………… 28
Figure 13: Change in CV % of fiberweb basis weight of polypropylene fiber w.r.t.
finish Uniformity……………………………………………………….. 28
Figure 14: Change in drafting force of polypropylene fiber w.r.t. finish level…….. 29
Figure 15: Change in drafting force of polypropylene fiber w.r.t. finish uniformity. 29
Figure 16: Change in 2.5% span length of polyester fiber w.r.t. finish level………. 30
Figure 17: Change in short fiber content of polyester fiber w.r.t. finish level……... 31
Figure 18: Change in 2.5% span length of polyester fiber w.r.t. finish uniformity… 31
Figure 19: Change in 2.5% span length of polyester fiber w.r.t. finish uniformity… 32
Figure 20: Change in CV % of fiberweb basis weight of polyester fiber w.r.t. finish
level .……………………………………………………………………. 33
Figure 21: Change in CV % of fiberweb basis weight of polyester fiber w.r.t. finish
uniformity………………………………………………………………. 33
Figure 22: Change in draft force polyester fiber w.r.t. finish level…………………. 34
Figure 23: Change in draft force polyester fiber w.r.t. finish uniformity…………... 34

vii
Figure 24: Change neps generated polyester fiber w.r.t. finish level……………….. 35
Figure 25: Change in neps count for polyester fiber w.r.t. finish uniformity………. 35

Chapter 3
Figure 1: Schematic Diagram of the NCRC Dynamic Needling Force Measuring
Device…………………………………………………………………….. 41
Figure 2. Locations of the Five-Instrumented Needles in the Needle Board……….. 41
Figure 3: Force experienced by needle during needlepunching…………………….. 42
Figure 4: Change in peak penetration force of polypropylene fibers w.r.t. finish
level………………………………………………………………………. 44
Figure 5: Change in penetration energy of polypropylene fibers w.r.t. finish level... 44
Figure 6: Change in peak penetration force of polypropylene fibers w.r.t. finish
uniformity………………………………………………………………… 45
Figure 7: Change in penetration energy of polypropylene fibers w.r.t. finish
uniformity….…………………………………………………………….. 45
Figure 8: Change in tensile strength of polypropylene fabric w.r.t. finish level……. 46
Figure 9: Change in tear strength of polypropylene fabric w.r.t. finish level………. 46
Figure 10: Change in energy to breaking point of polypropylene fabric w.r.t. finish
level……………………………………………………………………... 46
Figure 11: Change in tensile strength of polypropylene fabric w.r.t. finish
uniformity………………………………………………………………. 47
Figure 12: Change in tear strength of polypropylene fabric w.r.t. finish uniformity.. 47
Figure 13: Change in energy to breaking point of polypropylene fabric w.r.t. finish
uniformity………………………………………………………………. 47
Figure 14: Change in peak penetration force of polyester fibers w.r.t. finish level… 48
Figure 15: Change in penetration energy of polyester fibers w.r.t. finish level……. 48
Figure 16: Change in peak penetration force of polyester fibers w.r.t. finish
uniformity….…………………………………………………………… 49
Figure 17: Change in penetration energy of polyester fibers w.r.t. finish
uniformity………………………………………………………………. 49
Figure 18: Change in tensile strength of polyester fabric w.r.t. finish level………... 50
Figure 19: Change in tear strength of polyester fabric w.r.t. finish level…………... 50
Figure 20: Change in energy to breaking point of polyester fabric w.r.t. finish
level……………………………………………………………………... 50
Figure 21: Change in tensile strength of polyester fabric w.r.t. finish uniformity…. 51
Figure 22: Change in tear strength of polyester fabric w.r.t. finish uniformity…….. 51
Figure 23: Change in energy to breaking point of polyester fabric w.r.t. finish
uniformity………………………………………………………………. 51

Appendices
Appendix IV
Figure 1: Smoothened curve of needle force measurements for location 4………… 88
Figure 2: Frequency domain of force measurements……………………………….. 88
Figure 3: Derivation of Peak penetration force and penetration energy……………. 91

viii
 Chapter 1

“NEED FOR AN UNCONVENTIONAL METHOD TO EVALUATE FINISH PERFORMANCE

AS A MECHANICAL PROCESS AID”

ABSTRACT

The processing of fibers in a free, non-bound state gives rise to a range of frictional behavior for a given
process. The basic laws of fiber friction are reviewed. The finish treatment to the fibers mainly act as a
lubricant and anti static agent. A critical review of the requirements of the finish treatment and the
methods used to evaluate fiber friction and static charge generation are presented. The requirements of
the finish treatment with respect to carding and needlepunching are also discussed. The ineffectiveness
of conventional methods to evaluate performance of finish in such process and the need to measure
performance of finish and optimize it for each of these processes is highlighted. Methods to evaluate the
finish performance in carding and needlepunching will be presented in the papers to follow.

54
1. INTRODUCTION

Almost all synthetic fibers made today cannot be processed without the aid of a spin finish. The
synthetic fibers lack the natural lubricants present in natural fibers such as lanolin in wool, sericin in silk
and waxes on cotton, which protect them from being damaged during textile processing. These natural
lubricants are a mixture of fatty acid esters, fatty alcohols, free fatty acids and free alcohols, which are
almost impossible to duplicate in a laboratory. The fiber finish is essentially a complex mixture of a wide
range of chemicals applied to the surface of the fiber during spinning by means of a kiss roll or metered
applicator to improve the processability of the fiber. The main components of a fiber finish are,
lubricants (determine the retarding or frictional force caused by the fibrous material passing over a
moving or stationary surface), emulsifiers (in case of a water insoluble lubricant), antistatic agent (in
case of hydrophobic fibers), antimicrobial compounds, cohesive agents, humectants and wetting agents.
A more detailed description of the chemistry and function of finish components are given by Proffitt Jr.
and Patterson [20]. The fiber finish has to be specifically developed for the fiber it is to be applied on.
For example, some lubricants such as mineral oils and low molecular weight esters are not
recommended for polypropylene as they can cause the fiber to swell under certain conditions and impair
fiber properties [23].

In the case of nonwovens, cotton is usually bleached as raw stock prior to the manufacturing process
because bleaching of cotton nonwoven fabric is considered highly impractical. This could be due to
possible dimensional changes in the fabric during processing, or due to possible interaction of the bleach
with the adhesive chemicals present in the fabric [5]. Bleaching strips the cotton of its natural waxes and
any soluble salts, which could act as a lubricant. Hence, it is necessary to add a lubricant finish to
bleached cotton to aid the mechanical processing of cotton.

While there is a clear understanding of the effect of finish on the frictional properties of well defined
textile structures such as yarn or fabric, very little is understood about how finish affects the frictional
properties of fibers when they are processed in a free form. Here the fibers are subjected to very little
cohesive force and the frictional behavior varies as the fiber speed changes due to various accelerating
and decelerating forces. In many such cases, the finish treatment will have to be balanced to allow
maximum fiber processability given the different requirements of the frictional property at various stages
of a single process. To get a better understanding of the frictional behavior of fibers during such
processing, a thorough study of the fundamental laws of friction, methods to evaluate friction and the
interaction of the finish film at such an interface is necessary for each of these processes. Two such
processes are studied in this series of investigations namely, carding and needlepunching.

2. THE LAWS OF FRICTION

Two of the basic laws of sliding friction of solids were proposed by Leonardo da Vinci (1452-1519)
[30],

1. Frictional force is independent of the area of contact between two solids.

2. Frictional force is directly proportional to the load that is normal to the sliding direction.

55
The laws were redefined by the French engineer Amontons and were also verified by C.A. Coulomb
[30]. A simple representation of friction force is given by Coloumb and is valid only in the case of
metals or hard surfaces sliding against each other.

3. The static coefficient of friction is greater than the kinetic coefficient of friction.
4. Coefficients of friction are independent of sliding speed.

F = µR, (1)

Where, F is the frictional force, R is the normal load and µ is the coefficient of friction. The frictional
force is mainly governed by the asperities on the sliding surface, which in turn are dictated by the
smoothness of the surface. The frictional force can be further broken down into adhesive force and
deformation force. Adhesive force is the attractive force exhibited by the two surfaces in contact against
each other and deformation force is the force with which one surface, usually the harder one abrades the
other surface. Unlike metals, which generally exhibit plastic deformation, polymers undergo a significant
amount of elastic deformation. This viscoelastic nature of polymers greatly influences its frictional
properties. Several equations have been proposed to define the frictional properties of textile material
[30]. The generally accepted equation for textile materials was proposed by Bowden and Young [4].

F = aRn, (2)

Where, F and R are the same as in Equation [1], ‘a’ is a constant equal to µ when n = 1 and n is a
measure of elasticity or plasticity of the materials in contact. For two perfectly plastic surfaces, n = 1,
and for two perfectly elastic surfaces, n = 0.67. Several models have been proposed to characterize
friction behavior in fibrous materials [14,17]. But for all practical purposes, the empirical equation
proposed by Bowden and Young seems to be well suited.

However, the relationship does not hold for fibers or textile surfaces to which a finish has been added.
The value of n is determined based on the direct interaction of the two surfaces in contact. The finish
tends to separate these two surfaces with a complex multiple surface interactions making it very difficult
to evaluate “n”. Hence, the general practice is to plot “µ” (coefficient of friction) against processing
parameters such as speed, viscosity, pressure and so on. The frictional behavior of textile materials can
further be broken down into three categories based on the relative speed between the surfaces in
contact. The frictional behavior at slow speed (10-4 to 10-1m/min) is called boundary friction, as it is
more influenced by the mechanical, physical and chemical forces existing at the interface between the
two surfaces. The frictional behavior at high speed (above 5m/min) is called hydrodynamic friction, as it
is more influenced by the flow or shear properties of the lubricant. Finally, the friction behavior at
intermediate speed (10-1 to 5m/min) is called as semi-boundary friction, as it forms a transition between
the boundary and hydrodynamic friction. The coefficient of friction varies depending on the above three
categories as shown in figure 1.

56
 Figure 1: Frictional Behavior of Liquid Lubricated Yarn [19].

3. EFFECT OF FIBER AND FINISH PROPERTIES ON FRICTIONAL FORCES

The main finish parameters that would affect the frictional properties of a fiber are type of lubricant,
viscosity and cohesive agent added. The cohesive agent determines the magnitude of fiber to fiber
friction. As the finish is adsorbed on the fiber surface, the arrangement of molecules will alter the
frictional properties. The effect of some of the fiber and finish properties on frictional forces is
summarized in table 1.

Table 1: Influence of Fiber and Finish Properties on Frictional Force

Fiber Properties Effect on Frictional Forces
Material Different material have a frictional behavior of their own and hence require
specific lubrication.
Denier Friction increases significantly with increase in denier of the fiber, especially in
the hydrodynamic region. This is due to an increase in area of contact at the
fiber to metal or fiber to fiber surface interface [29]. The contact area of coarse
fibers is less than that of a fine fiber of same fiber type and mass.
Moisture Regain Hydrodynamic friction increases with an increase in moisture regain, especially
at higher relative humidity. This is due to an increase in area of contact when
the fibers swell. The high relative humidity is also known to affect finish
composition and makes it more non-lubricating [30].
Cross Section Experiments performed on certain tire cords show no effect of fiber cross
section on frictional behavior, these cross section were varied by means of

57
 manufacturing parameters or raw material and not by changing spinneret hole
design [30].
Crystallinity At low levels of crystallinity, an increase in crystallinity results in a decrease in
dry coefficient of friction. The fiber was found to harden and increase abrasion
resistance at higher levels of crystallinity [30].
Luster Experiments on nylon yarn showed that friction in the boundary and semi-
boundary region increases in the order, bright, semi-dull and dull [28].
Finish Properties Effect on Frictional Properties
Viscosity Friction is independent of lubricant viscosity at the boundary region and highly
dependent on viscosity at the hydrodynamic region. Fiber to metal friction is
directly proportional to the viscosity of the lubricant[16,27].
Surface Tension A study on silicone finishes, show The lubricative properties of silicone is due
to its low surface and interfacial tensions [7].
Amount of Several experiments show that the coefficient of friction leveled off after
Lubricant reaching a certain level of finish on the fiber [30].
Uniformity of finish There is very little understanding of the effect of uniformity of the finish applied
to the fibers for a given mass of fiber.

4. MEASUREMENT OF FRICTION

Several factors have to be taken into consideration while performing friction tests such as speed, surface
roughness, surface temperature, ambient temperature and humidity, finish viscosity, uniformity of finish
application, finish concentration of the fiber and fiber size and shape [22]. There are several methods to
measure hydrodynamic friction when the textile material is in the form of a yarn, fabric or other well-
defined textile structures. Some of the most commonly used techniques for measurement of
hydrodynamic friction are, Rothschild F-Meter Model R-1188 [24], ATLAB friction tester [31],
capstan method (ASTM D 3412–89) of measuring friction of textile yarn. Measurement of
hydrodynamic friction is found to be very much applicable at all aspects of yarn processing as it governs
the speed of that particular operation which in turn determines the production speed. Generally, High-
speed friction tests are used primarily for measuring fiber to hard surface friction and slow-speed
friction tests are used to study both fiber to hard surface and fiber to fiber friction at the boundary
region.

The measurement of boundary friction is less appreciated as it does not have any immediate industrial
implication in yarn and fabric formation or processing. It is also more difficult to measure due to its
small magnitude. Hence, it is also tough to reproduce the same results. Nevertheless, boundary friction
is a very fundamental parameter in fiber processing especially in process where the fibers are held
loosely like in carding and needlepunching. In carding though the collective fiber mass appear to travel
at high speeds from feed to delivery, the fibers as an individual experience high acceleration and
deceleration forces relative to one another, especially while being transferred from one carding element
to another. Similar forces act on fibers while needlepunching. The techniques used to measure boundary
friction can be divided into two general principles [15]. The first is the point contact method, in which
one fiber is rubbed against another fiber at right angle at constant speed. The second technique is the

58
extended line contact method, in which the fiber is in contact with more than one fiber, by means of
some cohesive form like twist or pressed against each other by means of a weighted grip.

4.1. Point Contact Method

A more convenient and proven method of the point contact type is the capstan method of Howell [13].
Moghazhy and Gupta [18] used a version of the capstan method to study the effects of structural and
morphological factors on fiber friction. A schematic diagram of the point contact method is given in
figure 2. One fiber is attached to the load cell, while the other fiber attached to the bow (U), whose
position can be changed relative to vertical fiber by micrometer gauge (M). The bow is attached to the
frame (P5) which in turn is attached to an Instron cross head. P6 is kept stationary and a vessel (C)
comes handy to test the frictional behavior when the fibers are immersed in a fluid. The fiber is loaded
by a known tension T0 and the tension T1 on the fibers is directly measured from the Instron. As the
fiber is deflected by an angle θ, the coefficient of friction is calculated by measuring the exponential of
the ratio between the two tensions. Another point contact method is the TRI/SCANTM Surface Force
Analyzer manufactured and sold by the Textile Research Institute.

Figure 2: Schematic diagram of Howell’s capstan method [13].

4.2. Line Contact Method

A variation of the capstan method was used by Roedger to measure fiber to fiber and fiber to metal
friction based on line contact. The schematic diagram of Roedger’s apparatus [21] is given in figure 3.
The instrument consists of a cantilever to which a fiber is attached on the free end. The fiber passes over
a capstan roller and is weighted on the other side. The roller is capable of being operated at nine
different levels of speed and the deflection of the cantilever would provide a measure of the frictional
force between the fiber and the metal surface of the roller. The roller can also be coated with parallel
fibers to measure fiber to fiber friction.

While performing experiments using the capstan method on monofilaments, Grossberg and Plate [10]
found that the rigidity of the fiber greatly influences the coefficient of friction of the fiber and the
difference in the coefficient varies more with a higher value of n, in the equation F = aRn. Hence the
rigidity of the fiber should be taken into account before making any comparison of frictional properties
of different fibers tested using the capstan method. A method to measure fiber to fiber friction was

59
proposed by Gralen and Lindberg based on the line contact principle [9]. The method tested the
frictional force between two fibers twisted together by a certain number of twists and the fibers are
individually weighted and pulled against each other. The test method required a certain minimum length
of fibers (not applicable to all staple lengths) and is very cumbersome.

A modification of the Sirkoski friction-measuring device was used by Basu, Hamza and Sirkorski, to
study the friction of cotton fibers [3]. The device consists of Two brass cantilever (A,B), a probe (C),
the fiber is pulled at a known velocity by grip (D) and the force is measured by means of two
semiconductor strain gauges on the brass cantilevers. The schematic diagram of the device is given in
figure 4. The device requires a very high precision drive and is unable to detect the stick-slip motion
found at all boundary friction.

Figure 3: Schematic diagram of Roedger’s Apparatus [21].

Lord proposed a method to study the friction between fibers in a bundle of fibers [15] (figure 5). Two
Parallel fiber fringe of about 1 inch wide with a density of about 5 mg/cm2, were pulled against each
other between two smooth surfaces. One of the two fringes was attached to a cantilever by means of
clamp (G) and the other fringe is atteched to the lower block (K) by means of clamp (H). The upper
block (J) exerts a known force (R) on the fiber fringe, normal to the direction of motion. The lower
block (K) is displaced with constant velocity towards direction (D). The displacement of the cantilever
(d) provides the frictional force measurements. A modification of the lord’s test is the beard test
developed by Broughton et al. [6], in which the fibers are vertically mounted on any tensile testing
device and a normal force is applied to the sides. This test can be used on any tensile tester with little
modification. The only drawback is the normal force, which needs to be maintained uniform for all test
and varied with the number of fibers and type of fiber. Another modification of Lord’s test is the sled
test used by cotton Inc., and pioneered by ICI [2]. A layer of carded fiber is normally loaded by a
known weight, and the friction between the carded fiber and the hard surface is measured by means of a
load cell.

The pin drafting force measurement is another method of measuring friction [26]. The apparatus
consists of two beds of intermeshing pins. An eight inch long sliver is mounted in an upper clamp. The
remaining sliver is engaged between the two boards of pins and held on the other side by a stationary
lower clamp. The upper clamp is attached to a load-measuring device and is moved at a constant rate of

60
10mm/min. The device measure the force necessary to draft the sliver. This gives a good measure of the
overall fiber to metal friction and the fiber to fiber friction. But the two components cannot be identified
separately.

Figure 4: Schematic diagram of the Sikorski friction measuring device [3].

Figure 5: Schematic diagram of Lord’s Apparatus [15].

4.3. Rotorring
The Rotoring is a custom manufactured modification of a sample-opening device [8]. The principle of
this device does not fall under the above two classifications and hence has to be discussed separately.
The instrument attempts to provide a good simulation of the opening process. The device is a large
open-end spinning system consisting of a feed roller, an opening roller and a rotor. The fiber is passed
through rotor and is collected in a groove in the rotor instead of being made into a yarn. The energy
readings on the opening roller provide a good measure of the resistance by the fiber to individualization.
Running the opening roller at lower speed, enables one to measure fiber to fiber friction, as the work is
concentrated more in pulling the fibers from the tangled mass. Fiber to metal friction can be measured

61
by running the roller at high speed where the work is more concentrated in wire teeth sliding through
the fibers than the fibers being actually pulled out of the tangled mass. However, the geometry and
normal force during friction measurements are still unclear. The instrument is also fairly expensive and a
production model is not yet available.

Figure 6: Schematic diagram of Beard Friction test [6].

5. STATIC CHARGE

It has been observed that whenever two materials are rubbed against each other, they become
electrically charged as electron from one material is passed on to the other. The donor or receiver
among the two is decided based on their position in the triboelectric series. The material listed first on
the series acquires a positive charge and vice versa. The charge is usually dissipated by means of a
conductive medium like moisture. In the case of hydrophobic fibers this charge is seldom dissipated fast
enough causing the fibers to be either attracted towards oppositely charged materials or repelled from
likely charged materials. This causes the fibers to cling to certain roller or machine surfaces and repel
from each other. Hence, controlling static charge is very essential to ensure processability. Often an
anti-static chemical is added to the finish to dissipate this charge. Controlling the ambient humidity is
also another mean of keeping a check on static charge generations. The effectiveness of the anti-static
treatment can be measured by comparing the charge generation before and after the treatment.

There are several methods to measure static charge generation based on either resistivity or charge
dissipation time. Some of the methods that measure static charge based on dissipation time is the half-
time (the time taken for half the charges to be dissipated) measurement [11], Fabric to metal cling test
[1], Sail test [35], and Ash test [32]. The half-time measurement method is more suitable for measuring
static charge generated in carpets than in fiber web. This is because the exact conditions of charge
generation in the case of fiber web cannot be replicated in a laboratory and also the fiber web is part of a
dynamic system, undergoing changes at all time during processing. Resistivity can be measured by an
electrostatic voltmeter. The electrostatic voltmeter must also be of the non-contacting type so as not to
disturb the fibers being processed. Usually the Reynolds number Re of the system, which is the ratio
between the charge relaxation time and the material transport time [36], is considered to be less than
one (Re << 1). Therefore for all practical purposes, web movement is assumed to be negligible. The
variation in static charge due to speed of the material is accounted for by increasing the frequency of the
reading. This provides almost a continuous monitoring of static charge generation. The problem with a

62
non-contacting voltmeters is, the voltmeter is calibrated in units of potential, or volts, while surface
charge must be measured in units of Coul/m2, or charge per unit area. There are several such voltmeters
commercially available like the RothscildTM Static Voltmeter, TantecTM non-contacting electrostatic
voltmeter, MonroeTM and TrekTM Monroe and Trek electrostatic voltmeters. . Noncontacting voltmeters
(NCV) such as MonroeTM or TrekTM have been used in the industry to measure surface charge in many
electrostatic applications. The working principle of Noncontacting Electrostatic Voltmeter and the
relation between the reading of a NCV probe and the value of measured surface charge is given by M.
N. Horenstein[12]. The device consists of a field-mill type probe with viewing aperture, a detection
circuit, and a high-voltage amplifier, as illustrated in fig.7. When the probe is exposed to a surface of
non-zero potential, the detected field signal is amplified by the high voltage amplifier and a feedback to
the probe raises the probe potential toward that of the surface reducing the net field magnitude at the
aperture to zero. This can be represented to a simple capacitance problem as shown in figure 8.

Figure 7: Block diagram of non-contacting voltmeter [12].

Figure 8: Two-conductor capacitance problem showing conductor capacitance to ground C1, NCV
probe capacitance to ground C2, and mutual capacitance CM. [12].

63
The relationship between conductor potential and charge is given by the following matrix equation. This
gives the null field condition i.e., zero stored charge in the capacitance CM, which is equivalent to the
condition V1 – V2 = 0 or V2 = V1 = Q1/C1 ( for CM << C1 ) .
V1 = Voltage between Conductor and ground
V2 = Voltage between Probe and ground
C1 = Conductor capacitance to ground
C2 = Probe capacitance to ground
CM = Mutual capacitance

An array of such NCV can be fixed over the width of the card for example, and measure the static
charge distribution on fiberweb. The relationship between the reading of NCV and charge density
depend mainly on distance between the probe and the surface charge, the geometric distribution of the
surface charge and the probe geometry. Therefore, the instrument has to be calibrated based on the
above factors.

6. SIGNIFICANCE OF FINISH TO CARDING

Carding is a very crucial step of fiber processing in yarn manufacturing as well as nonwovens
manufacturing. There are a few variations on the design of a card based on fiber length, fineness and
rigidity, but the basic principle remains the same. Here individualization of the fibers is achieved by
breaking down the lumps of fiber fed into the card by means of repeated action between two saw
toothed surfaces. The surfaces move against each other with the saw tooth facing opposite direction.
The distance between the surfaces is small enough only to allow the passage of a few fibers. This
restricted passage causes the fibers to be caught on the saw teeth and separated from the lumps. The
fibers undergo tremendous stress as they are tugged and pulled. The frictional forces play an important
role in the whole process of carding. As the fibers are in a free state, there is no fixed demarcation
between boundary friction and hydrodynamic friction. Generally, higher fiber to fiber friction is
necessary in ensuring a carded web with good cohesion, uniformity and integrity, while a low fiber to
metal friction is necessary to ensure less damage to the fibers especially with smaller diameters. A very
high fiber to metal friction will load the carding elements and jam it causing major down time to clean it.

Lubrication has the ill effect of making the fibers too slippery. Inadequate coefficient of friction may
cause the web to break down and also clumps of fibers to be drawn in past the feed roll and feed plate
by the lickerin. These clumps may also overload the card elements and may even damage the card wire
[33]. Improper selection of finish also lead to filling of card wire by the finish [33]. The thermal
behavior of the finish dictates the finish’s tendency to deposit residue on surfaces or cause toxic fumes.
The finish forms a powdery coating on the carding elements and more often tend to disperse into the
atmosphere and may lead to an unhealthy working condition. The thermal behavior of the finish can be
assessed by means of either a thermogravimetric analysis or an isothermal analysis. Some times
problems attributed to finish can also be due to other reasons. A study by Scardino et al.[25], on carding
of glass fiber found that the choking of lickerin could also be induced by a very close feed roller to nose
plate setting.

64
The optimization of the finish treatment to the fiber also depends on various fiber factors like type of
fiber, fiber fineness, fiber length and rigidity, and also carding parameters like speed of the card and
setting of the carding elements. Another factor that may affect the finish performance in carding is
uniformity of the finish on the fiber on a macro scale i.e., it deals with the variation in finish add on
percentages from one unit mass of fiber to another rather than the finish film on a single fiber. A study
of the effect of finish uniformity on fiber processability will enable to better explain the cause of
applicator effect if any. It can open possibilities for blending fibers with different finish levels in case
finish uniformity has no effect on fiber processability.

7. SIGNIFICANCE OF FINISH TO NEEDLEPUNCHING

In the case of needlepunching, the fiber web from the card is cross-lapped to form a bat of higher basis
weight and the fibers are further entangled and bound to each other. The entanglement results in re-
orientation of fiber from one another and hence, the fibers undergo friction between them and between
the needles. In this case the finish treatment should be optimized to ensure a friction low enough to
permit realignment of fibers and high enough to impart sufficient strength to the fabric after
needlepunching. The fiber to metal friction also needs to be optimized to enable the needles to slide
through the web and the fibers to be grabbed by the barbs on the needle. Earlier studies indicate that the
use of fiber finish have resulted in a reduction of needling forces and needle breakage in the case of
kevlar aramid fiber [34]. As the stiffness of Kevlar aramid fibers (500-900gpd) is much more than that
of polyester (Dacron, 12-17gpd), the same results may not be applicable for polyester fibers. The
properties of the needlepunched fabric may also be affected by finish performance as fiber to fiber
friction plays an important role in determining inter fiber cohesion.

8. SHORTCOMINGS OF CONVENTIONAL METHOD OF TESTING FIBER FRICTION TO

EVALUATE FINISH PERFORMANCE IN CARDING AND NEEDLEPUNCHING

Measuring of fiber friction is difficult, as a prohibitively large amount of testing may be required to
achieve some desired degree of accuracy. Further more the deviation from a smooth surface and circular
cross section makes it more difficult to interpret experimental observations. Both carding and
needlepunching are complex processes where the friction measurements cannot be directly translated
into a cause and effect equation. This is because the fibers undergo rapid acceleration and deceleration
causing the conventional measure of boundary and hydrodynamic friction meaningless. Exact laboratory
simulation of the carding or needlepunching process to measure friction has not been very successful.
This calls for a full-scale study of the carding and needlepunching process with respect to finish
performance.

Key response parameters can be identified and tested for effect of finish parameters and their possible
interaction with the process parameters. Such a study is expected to provide a better understanding of
the finish performance. The next two papers in this series presents the experimental design and results of
two such studies conducted on carding and needlepunching.

65
REFERENCES

[1] AATCC Test Method, “Electrostatic Clinging of Fabrics: Fabric to Metal Test:, AATCC
Technical Manual, 115-1995 (1998)
[2] Allen, C., Cotton Inc, Personal Communication
[3] Basu, S.C., Hamza, A.A., Sikorski, J., “The Friction of Cotton Fibers”, J. Text. Inst., #2/3, p68
(1978)
[4] Bowden, F.P. , Young, J. E., “Friction of Clean Metals and the Influence of Adsorbed Films”,
Proc. Roy. Soc. (London), A208:444 (1951)
[5] Broughton, J., R. M., El Mogahzy, Y.E., “The Measurment of Fiber Friction and its Importance in
Manufacturing Cotton Nonwovens”, TAPPI Proceedings, Nonwovens Conference, p 33-39
(1992)
[6] Broughton, R.M., Moghazy , El Y.E., “the Measurement and Importance of Cotton Fiber
Friction”, Paper presented at the 1992 Beltwide Cotton Conference, Nashville, TN (1992)
[7] Buckingham, A., Avery, D., “Silicone Treatments for Nonwovens”, TAPPI Proceedings,
Nonwovens Conference, 331-334 (1993)
[8] Ehler, P., Mavely, J., “Crimping of Polyester Staple Fibers”, Melliand Textilberichte (English
Edition), 97-106 (Feb 1983)
[9] Gralen, N., and Lindberg, J., “Measurement of Friction between Single Fibers, Part II: Frictional
Properties of Wool Fibers Measured by the Fiber-Twist Method”, Textile Res. J., 18, p287-301
(1948)
[10] Grossberg, P., Plate, D.E.A., “Capstan Friction for Polymer Monofilaments With Rigidity”, J.
Text. Inst., p268-283 (1969)
[11] Hayek, M., and Chromey, F.C., “The Measurement of Static Electricity on Fabrics”, Am. Dyestuff
Rep., 40:164 (1951)
[12] Horenstein, Mark N., “Measuring Surface Charge with a Noncontacting Voltmeter”, IEEE Trans,
Power Apparutus and Syst., p 1811-1668 (1993)
[13] Howell, H.G., “ The Friction of a Fiber Round a Cylinder and its Dependence upon Cylinder
Radius”, J. Textule Inst. 45, T575-579 (1954)
[14] Howell, H.G., Mazur, J., “Amontons Law and Fiber Friction”, J. Textile Inst., $$, T59-T69 (1953)
[15] Lord E., “Frictional Forces Between Fringes of Fibers”, J. of Textile Institute p41-59 (1955)
[16] Lyne, D.G., “the Dynamic Friction Between Cellulose Acetate Yarn and a Cylindrical Metal
Surface”, J. Text. Inst., 46:P112 (1955)
[17] Moghazy El, Yehia, E., Gupta, S.B., “Friction in Fibrous Materials. Part I: Structural Model “,
Textile Research Journal 61(9),p547-555 (1991)
[18] Moghazy El, Yehia, E., Gupta, S.B., “Friction in Fibrous Materials. Part II: Experimental Study of
the Effects of Structural and Morphological Factors”, Textile Research Journal 63(4),p219-230
(1993)
[19] Olsen J.S., “Frictional Behavior of Textile Yarns”, Textile Res. J., 39:31 (1969) [9]
[20] Proffitt Jr., T.J., Patterson, H.T., “Oleochemical Surfactants and Lubricants in the Textile
Industry”, JAOCS, Vol.65 No.10 (Oct 1988)
[21] Robins, M.M., Rennell, R.W., Arnell, R.D., “The Friction of Polyester Staple Fibers: The Effect of
Speed, Processing, Spin Finish, Humidit and Counter-surface Roughness”, J. Text. Inst., 1: p126-
139 (1988)

66
[22] Ross Stan, “Important Properties of Fibers that are affected by Finish”, NCRC Short Course on
Fiber Finish, Feb, (1998)
[23] Ross, S., “Important Properties of Fibers tha are affected by Finish”, NRCRC Short-course (1998)
[24] Rothschild Messinstrumente, Zurich. F-Meter R-1188, Electronic Instrument for Automatic
Measurement of the Friction Coefficient, Lawson-Hemphill Sales, Inc., Spartanburg, SC (1995)
[25] Scardino, F.L., Frank, K.K., Williams P.M., “Characterization of the Cohesion of Carded
Fiberglass Webs”, TAPPI Proceedings, Nonwovens Conference, 169 (1987)
[26] Scardino, F.L., Lyons, J.W., “Influence of Fiber Geometry on the Mechanical Properties of
Assemblies During Processing”, Textile Res. J., Vo. 40, No.6, p559-570 (1970) [
[27] Schick, M. J., “Friction and Lubrication of Synthetic Fibers. Part IV: Effect of Fiber Material and
Lubricant Viscosity and Concentration”, Text. Res. J., 43:342 (1973)
[28] Schick, M.J., “Friction and Lubrication of Synthetic Fibers. Part V: Effect of Fiber Luster, Guide
Material, Charge, and Critical Surface Tension on Fiber Friction.”, Textile Res. J., 44:758, (1974)
[29] Schick, M.J., “Friction and Lubrication of Sythetic Fibers. Part III: Effect of Guide Temperature,
Loop Size, Pretension, Denier, and Moisture Regain of Fiber Friction"” Textile Res. J., 43:254
(1973)
[30] Slade, Philip E., “Handbook of Fiber Finish Technology”, Marcel Dekker, Inc. New York (1998)
[31] Specialities from ICI for Fiber Producers., Laboratory Test Methods for Fiber Finishes. Friction
Test, ICI Americas, Wilmington, DE (1987) [
[32] Steiger, F.H., “Evaluating Antistatic Finishes”, Textile Res. J., 28:721 (1958)
[33] Wagner J.R., “Carding Nonwovens: A Quarter Century of Experience”, Nonwovens Conference
(1995)
[34] Wagner, J.R., Ahles, L.J., “The Effect of Fiber Finish and Needle Type on Needle Breakage and
Compaction of Needle Punched Kevlar Aramid Fiber Nonwovens”, Nonwovens Conference
(1987)
[35] Wilkinson, P.R., “Static Electricity: Domestic and Commercial Implications in Textile Materials”,
Modern Textiles, 51:35 (1970)
[36] Woodson, H.H. and Melcher, J.R., “Electromechanical Dynamics (Part II)”, Wiley, New York,
Section 7.0, p.355 (1968)

67
 Chapter 2

“FINISH PERFORMANCE AS A MECHANICAL PROCESSING AID IN CARDING.”

ABSTRACT

This part of the study aims to investigate role of fiber finish in the carding process. Three different fibers namely, cotton,
polyester and polypropylene with different levels of finish and finish uniformity, are processed on carding machine.
Certain key response parameters such as fiber breakage, fiberweb cohesion, fiberweb uniformity, nep generation and static
charge generation, are chosen and the effect of finish level and its uniformity on these parameters are observed. The finish
performance is also studied for possible interaction effects between other carding process parameters. The results indicate
a significant effect of finish add-on percentage on most of the response parameters. On the other hand, finish uniformity
was found to have less influence on the response parameters. It has also been found that the response parameters behave
differently for each of the three fibers and hence results based on one type of fiber may not be applicable for another.

68
INTRODUCTION

Carding is a very crucial stage in the processing of staple fibers. Many a times, problems in
carding have been attributed to the percentage add-on of finish on the fiber and its uniformity.
This has impelled to investigate the actual effect of finish and its uniformity in terms of
mechanical processability of fibers. The function of fiber finish in the carding process is to reduce
the fiber to metal friction and thus prevent abrasion and fiber damage. The finish must also
optimize fiber to fiber friction for better web cohesion. In the case of low moisture content fibers
like polyester, the finish should also control the static charge generation which otherwise would
lead to web bellowing and fibers clinging to machine parts and dust accumulation.

Fiber finish uniformity at a macro level can be defined as the level of distribution of finish on the
fiberweb. Previous work in this field has identified many causes for finish non-uniformity on fibers such
as applicator effects, spreading of finish film on fiber surface which in turn depends on the surface
tension of the finish and surface characteristics of the fiber[3]. Many problems caused by improper
selection of finish like low cohesion between the fibers in the fiberweb, overloading of card wire and
excessive fiber breakage in the fiberweb have also been reported in the previous chapter. This calls for
an optimization of the finish application for a given fiber type. For this, a thorough understanding of the
role of fiber finish on carding processability is necessary.

The experimental design used in this project aims to study the effect of finish level and finish uniformity
on a macro level on various response parameters such as fiber breakage, web uniformity, neps and static
charge generation. The experiment was performed using three different fibers namely cotton, polyester
and polypropylene. The processing parameters are specific to each fiber. The fibers made with different
finish levels and uniformity were processed on a carding machine under controlled conditions. The
interaction of finish performance with respect to carding variables such as throughput and cylinder to
doffer setting was also investigated.

OBJECTIVES

The overall objective of this part of the study is to investigate the role of fiber finish in the
processability of fiberwebs for nonwovens in carding. The change in fiberweb characteristics
such as fiberweb web uniformity and quality, web cohesion and the ability to run at high speed,
fiber breakage and neps are used as parameters to asses the finish performance in terms of
fiberweb quality.

EXPERIMENTAL APPROACH

69
The effect of finish level was studied by using a minimum of five different finish levels. The range of
finish levels was arbitrarily chosen after consulting with the industrial advisors. The fiber producers
supply at least 3 bales of fibers of different finish levels. The remaining finish levels were obtained by
blending these 3 primary levels. To study the effect of finish uniformity, the fibers from the high (H) and
low (L) levels of finish are blended in the right proportion to achieve a theoretically same finish level as
that of the medium (M). This blend (HL) should have a more non-uniform finish application than that of
the medium level (M). A comparison of the response parameters of the blend (HL) and medium (M)
should provide a good indication of the effect of finish uniformity on finish performance. Two different
levels of throughput and cylinder to doffer settings were used to determine any interaction effect of
finish with respect to carding variables. The treatments were replicated thrice to eliminate bias. The
throughput is expressed in terms of doffer surface speed rather than mass of fibers produced per unit
time. As the mass of the fibers is dependent on the type of fiber, the use of doffer speed to represent
throughput helps to perform more meaningful comparisons. The speed of the feed roller was kept at a
fixed ratio to the speed of the doffer so as to maintain a constant basis weight of the fiberweb. Varying
the carding parameters, changes the amount of recycling time the fibers spend on the cylinder i.e., the
carding points per fiber per unit time is changed, which in turn affects the work done on the fibers by
the cylinder and flat interaction. By having a wider cylinder to doffer setting, the recycling time of the
fiber on the cylinder increases (carding points per fiber per unit time increases) and the effect is vice
versa for a closer cylinder to doffer setting. In the same manner, by decreasing the throughput while the
speed of the cylinder is left unchanged, the effective time spent by the fiber on the cylinder increases and
vice versa for an increase in throughput with a constant cylinder speed. In effect, a combination of these
two parameters would present a good representation of the finish performance with varying degrees of
carding on the fiber.

The experiments were performed on a Hollingsworth M 2000 high-speed stationary flat top card (for
cotton and polyester) and on a Saco-Lowell Card Master card (for polypropylene) at the Hollingsworth
research laboratory, Greenville SC. Five main response parameters have been identified. They are
fiberweb uniformity, fiber breakage, web cohesion, neps and static charge generation. The test method
used for collecting each of the above four response parameters is described below.

3.1. Measurement of Fiber Breakage

The fiber breakage during the carding process can be estimated by measuring the fiber length and its
distribution of the fiber mass before and after its passage through the card. The AFIS (Advanced Fiber
Information System, provided by cotton Inc.) was used to measure the fiber length distribution of
cotton and the Keisokki Classifiber, model KCF/ LS (provided by Lawson-Hemphill) was used to study
the fiber length distribution of polypropylene and polyester. Comparisons are made based on the 2.5%
span length and the short fiber content for each treatment. The 2.5% span length is the length exceeded
by 2.5% of the fibers in the entire distribution. It is the parameter used in the spinning industry to
determine the settings of the various machine elements and hence is a good representation of fiber
length. The short fiber content is described as the percentage of fibers in the distribution whose length is
less than 1.27cm (0.5"). In the case of cotton 6 samples were collected after processing for each
treatment. For polypropylene and polyester, 4 samples were collected for each treatment and tested on
the Keisokki Classifiber KCF/LS.

70
3.2. Measurement of Web Uniformity
The fiberweb produced at the card has very little cohesion and therefore requires great care not to
disturb the web while measuring the basis weight distribution of the fiberweb. Offline measurement
techniques are often inaccurate due to poor handling of material. Measurement of fiber loading on
doffer is a good indicator of the fiberweb uniformity. For this many effective online measuring
techniques have been developed. A thorough analysis of the various available techniques and
instruments is given by Meng et al[4]. The Nonwoven Cooperative Research Center (NCRC) IR-based
photometric instrument is used to measure the fiber load on the card elements. The device works on the
principle that light generated by the IR LED is focused on the carding surface and the reflected light
(specular reflection) is sensed by the photometric device. The voltage reading of the sensor is a direct
function of the amount of light reflected which in turn depends on the amount of fibers on the carding
element. The schematic diagram of the IR-based device and the possible mounting positions on the card
is given in figure 1. A more detailed description of the principle and working of the NCRC IR-based
photometric device is given by Seyam et al [5].

Seven IR LED and sensor assembly was mounted on the doffer at the recommended distance of 25 mm
from the doffer surface and 25ο to the axis perpendicular to the doffer surface [5]. The device needs to
be calibrated for each fiber type used, as the reflectance property varies depending on the type of fiber
and surface characteristics of the fiber. The calibration procedure consists of the following steps. The
fiber was processed on the card at five different levels of web density. Carded sliver was collected over
a period of 2 min. and the corresponding voltage reading from the IR device is recorded. The data is
collected at a rate of 100 scans/sec. The linear density of the carded sliver was noted. At least five
readings were collected for each level of web density. The output voltage was regressed on the weight
per linear yard of the carded sliver to arrive at the calibration equation. Care should be taken to reach a
steady state condition of carding before collecting any data. This calibration equation was later used to
convert the voltage readings into weight per unit area of the fiberweb. The equation also determines the
sensitivity of the IR sensors to the change in web density of the given fiber. It was found that the
calibration equation did not vary with finish level for all three fibers used in this study.

71
Figure 1: Schematic diagram showing the mounting and data acquisition system of the NCRC IR-based
photometric devices.

3.3. Measurement of Neps

Neps by definition are small entanglements of fibers, which cannot be separated. The neps in the card
web is measured by manually counting the number of neps in a specific area of the card web and
calculating the neps per unit area of cardweb. An alternative method to measure neps is given by ASTM
standards D 5866-95. This method is based on AFIS-N instrument and can be used for cotton alone.
For polyester and polypropylene, the manual method of measuring neps was adopted. Six measurements
were made for each treatment. The results are expressed in terms of neps per gram.

3.4. Measurement of Web Cohesion

Two methods for measurement of web cohesion were considered. The first method requires the
fiberweb to be passed between two pairs of draft rollers. The draft between the rollers is slowly
increased till the point of failure (breakage) of fiberweb. The draft forces were recorded. The fiber
specimen was weighed and density calculated. The draft force per unit density of fiberweb was
calculated and used as measure of fiberweb cohesion. The second method requires the fiber assembly in
sliver form. Fiber cohesion test for sliver is given by the ASTM standard tests[1,2]. It consists both a
static (< 1 cm/min) and dynamic test. The static test involves the measurement of the tensile loading of
the sliver by a very sensitive load cell. The dynamic test is similar to the draft test mentioned above
except that the specimen is in the form of a sliver.

In this study the Rothschild Cohesion-Meter R-2020 was used to measure the sliver cohesion. The
instrument measures the force required to draft a given sliver between two pairs of rollers. The draft
between the pairs of rollers was kept constant at a recommended draft of 1.25 and the speed of the

72
delivery rollers was 10 m/min. The draft force was recorded as an average force (N) over 2 minutes of
drafting. Three readings were taken for each treatment.

3.5. Measurement of Static Charge

As static charge generated on the cardweb gets dissipated before they can be collected and tested, the static charge
measuring device, must be of an online and non-contacting type. For this purpose, the Tantec non-contacting electrostatic
voltmeter was used [6]. It is a hand held self-calibrating electrostatic voltmeter making it convenient to measure the static
generation at any region of the web without disturbing it and while the card is running.

PROCESSING PARAMETERS

Cotton
Bleached cotton with 3 different finish levels were provided by Veratec. The fiber length is 27.75mm
(1.09”, effective length) and its fineness is 4.7 micronaire. The finish consists of butoxyethyl stearate
and sodium acetate (anti-stat). It has to be noted that the cotton fibers retained a significant amount of
moisture while performing the fiber length measurements on the Advanced Fiber Information System
(AFIS). The fibers were processed on a Hollingsworth M 2000 high speed card whose specifications are
given in table 1.

Table 1: Card specifications for cotton

Element Diameter, cm (in) Speed, rev/min Card Clothing, teeth/cm2
(teeth/in2)
Feed Roller 7.62 (3)
Licker-in 25.40 (10) 6.2 (40)
Cylinder 129.54 (51) 380 max 124.0 (800)
Doffer 68.58 (27) 22 – 85 46.5 (300)

The three finish levels were labeled high, low and no finish as the finish extraction results were not
available at the time of the experiments. Later, the extraction results showed a high variation in finish
levels (appendix I). The low finish cotton failed to run in the high speed card. The sliver was found to be
“scroopy” and repeatedly broke down at the calendar rollers after the autoleveller mechanism. It also
generated static charge causing the fiber web to fray and stick to the carding elements. Hence, we were
unable to run the low finish and the no finish fibers. The high and low finish fibers were blended to give
four different finish levels, 100% high, 75/25 high/low, 50/50 high/low and 25/75 high/low. The choice
of blend ratio was purely arbitrary. To obtain a relationship between the fiber load on doffer and the
output voltage, the IR sensors were calibrated using 100% high finish cotton at a throughput of 45.72
m/min (50 yds/min). The resulting calibration curve and equation is given in figure 2. The high R2 value
indicates a good linear relationship between the web density and output voltage. The process
parameters used for cotton are given in table 2. The treatments were applied in a random manner. Data
was collected for each treatment only after the card reached a steady state condition in terms of finish
level and throughput.

73
 Figure 2: Calibration curve for cotton

Table 2: Process parameters used for cotton

Process parameter Levels
Finish level H25L75, H50L50, H75L25, H100L0
Finish Uniformity -
Fiberweb basis weight 5.83256 g/m2
Cylinder to doffer setting 0.127 mm, 0.229 mm
Throughput (doffer speed) 16 kg/hr (45.72 m/min)
32 kg/hr (91.44 m/min)

Polypropylene
Polypropylene with 3 different finish levels were provided by Fibervision Inc. The fiber length is 38mm
(1.5”) and fiber fineness is 9 denier. The finish levels were determined by performing 20 extractions on
each fiber type. The finish levels are 0.11% (low), 0.61% (medium) and 1.97% (high). The finish
applied on the polypropylene fibers is of a generic type prepared by Goulston Tech Inc. The finish is
composed of 90% lubricant (PEG(400) Monolaurate) and 10% antistat (Amine neutralized linear alkyl
phosphate). The fibers were processed on a Saco Lowel - Card Master flat top card whose
specifications are given in table 3.

Table 3: Card specifications for polypropylene

Element Diameter, cm (in) Speed, rev/min Card clothing, teeth/cm2
(teeth/in2)
Feed Roller 7.62 (3)
Licker-in 25.40 (10) 6.2 (40)
Cylinder 129.54 (51) 382 (max) 124.0 (800)
Doffer 68.58 (27) 14 – 73 26.0 (168)

A less dense card clothing was used for the doffer to reduce damage to the 9 denier fiber. The IR
sensors were calibrated using polypropylene with 0.61% finish level. The resulting calibration curve and

74
equation is given in figure 3. The high R2 value indicates a good second order relationship between the
web density and output voltage. Except for the low finish fibers, none of the other blends generated
static charge. We did not use the high finish fibers directly as it deposited a lot of finish on the carding
element. This may affect subsequent runs in the card. Hence, the 3 finish levels were blended to provide
5 different finish levels and one of the blends consist of the high and low finish fibers in exact proportion
to arrive at a final finish level equal to that of the medium finish. We call this 0.61%(B) to differentiate it
from the medium finish. A comparison of 0.61%(B) and 0.61% finish fibers should give a fair idea about
the effect of finish uniformity on fiber performance. The finish levels and their blend composition is
given in table 4. The process parameters used for polypropylene are given in table 5. The treatments
were applied in a random manner. Data was collected for each treatment only after the card reached a
steady state condition in terms of finish level and throughput.

Table 4: Blend compositions for polypropylene

Blend # Desired finish level High finish Medium finish Low finish
1.97% 0.61% 0.11%
1 0.36% - 50% 50%
2 0.61% - 100% -
3 0.61%B 27% - 73%
4 0.86% 25% 56.5% 18.5%
5 1.11% 50% 13.5% 36.5%
6 1.36% 60% 26.8% 13.2%

Table 5: Process parameters used for polypropylene

Process parameter Levels
Finish level 0.36%, 0.61%, 0.86% 1.11% 1.36%
Finish Uniformity 0.61%, 0.61% (B)
Fiberweb basis weight 9.67108 g/m2
Cylinder to doffer setting 0.127 mm, 0.229 mm
Throughput (doffer speed) 15.9 kg/hr (27.4 m/min)
30.75 kg/hr (53 m/min)

Figure 3: Calibration curve for polypropylene

75
4.3. Polyester
Polyester fiber of 3 denier and 38mm staple length, with four different levels of finish were provided by Wellman Inc. The
finish used for the polyester fiber is supplied by Goulston Tech. and is the same as that on the polypropylene fibers. The
remaining finish levels are obtained by blending these 4 primary levels. The final finish levels and their blend
composition are given in table 6. Here the blend # 3 and # 4 correspond to fiber with uniform finish distribution and fiber
with less uniform finish distribution respectively.

The experiments were performed on a Hollingsworth M 2000 high speed stationary flat top card. The
process parameters used for polypropylene are given in table 7. The treatments were applied in a
random manner. Data was collected for each treatment only after the card reached a steady state
condition in terms of finish level and throughput. The sensors were once again calibrated for the
polyester fiber and the resulting calibration curve and equation is given in figure 4.

Table 6: Finish Level and blend composition of polyester fiber

Blended Original Finish Levels
Blend #
Finish Level 0.038% 0.111 % 0.132 % 0.22 %
1 0.075 % 50 % 50 % - -
2 0.111 % - 100 % - -
3 0.137 % - - 100 % -
4 0.137 % (B) 10 % 60 % - 30 %
5 0.175 % - - 50 % 50 %
6 0.220 % - - - 100 %

Table 7: Process parameters used for polyester

Process parameter Levels
Finish level 0.075%, 0.11%, 0.137%, 0.175%, 0.22%
Finish Uniformity 0.137%, 0.137% (B)
Fiberweb basis weight 5.2076gm/m2
Cylinder to doffer setting 0.127 mm, 0.229 mm
Throughput (doffer speed) 25.72 kg/hr (82.3 m/min)
51.43 kg/hr (164.6 m/min)

76
 Figure 4: Calibration curve for polyester

5. RESULTS AND DISCUSSION

The statistical results for all the response parameters observed are given in appendix II. The page number of the statistical
results for the corresponding parameter under discussion is given in brackets, following the title of the graph.

5.1. Cotton
The finish extraction results of cotton show a significant variation in the finish uniformity for each finish level. These
variations could be either due to application effect or due to error in extraction method. Nevertheless, the results for fiber
breakage and nep generation represent good trend with respect to finish level and are discussed below. Static charge
generation was observed during the trial running of bleached cotton with no finish and low finish content. The magnitude
of the static charge generated on the electrostatic voltmeter was as high as 24.3 volts. But, due to processing difficulties,
the above two blends were not used in the experimental design and the other blend ratios did not generate any static
charge. As for web cohesion and fiberweb uniformity, the statistical analysis show no significant effect of finish level on
any of the above two response parameters. This may be due to the high variation present in the finish level itself and/or
the actual difference between the two finish levels was too small to cause significant impact.

5.1.1. Fiber Breakage

There seem to be a significant effect of finish level on the effective fiber length of cotton. The effective length of cotton is
similar to the 2.5% span length of the man made staple fibers (2.5% span length = 0.9xeffective length). Figure 5 shows
that the effective fiber length of cotton increases with an increase in finish level. Since the effective lengths of the original
fibers were same for all finish level, this indicates that fiber breakage decreases with an increase in finish levels. Here, the
effective lengths of the processed cotton fibers were found to be greater than that before processing. As mentioned earlier,
the cotton fibers were damp while measuring fiber length prior to carding. This dampness may have caused the fibers to
swell and hence result in a lower estimate of the fiber length prior to carding. The setting and throughput have no
significant effect on fiber breakage. Figure 6 shows the change in short fiber content of the cotton fibers with respect to
finish level. The short fiber content decreases with an increase in finish level indicating a drop in fiber breakage with an
increase in finish level.

5.1.2. Neps
There is a significant effect of finish level on the nep count of the cotton. Usually the feed material of the card has some
amount of neps in them as a result of prior handling of fibers in the previous processes. In the carding process the amount
of neps generated should be less than the amount of neps eliminated. So good carding performance should lead to a
maximum amount of neps eliminated and is very much dependant upon the carding parameters. The analysis of variance

77
indicates a very significant effect of finish level, cylinder to doffer setting and throughput on nep count. From figure 7, we
can see that the reduction in neps increases with finish level.

Figure 5: Change in effective fiber length of cotton w.r.t. finish level (58)

Figure 6: Change in short fiber content of cotton w.r.t. finish level (58)

Figure 7: Change in nep count of cotton w.r.t. finish level (59)

78
5.2 Polypropylene

5.2.1. Fiber Breakage

Figure 8 shows the change in 2.5% span length of polypropylene fibers with respect to finish level. There is a significant
effect of finish level on fiber breakage. From the graph below it is evident that fiber breakage decreases with an increase
in finish level. Figure 9 shows the change in short fiber content of polypropylene fibers with respect to finish level for all
treatments. There is a significant effect of finish level and throughput on short fiber content. In general the short fiber
content is a good indication of fiber breakage. But if fiber breakage occurs and the broken fibers are longer than 0.5 inches
then short fiber content will not be able to reflect these broken fibers. Nevertheless it is still useful in understanding how
the fiber breaks. Here the fiber breakage can be inferred to occur at higher fiber length and mostly the broken fiber end up
being longer than 0.5".

On comparing the effect of finish uniformity on fiber breakage, it is found that finish uniformity has no
significant effect on both 2.5% span length and short fiber content of polyester. Figure 10 and 11 show
the change in 2.5% span length and short fiber content of polypropylene with respect to finish
uniformity.

5.2.2 Fiberweb Uniformity

Figure 12 shows the change in CV % of the fiberweb basis weight of polypropylene. Statistical analysis, show a significant
effect of finish level and throughput on fiberweb uniformity. On performing multiple comparisons between the means of
finish level, show the difference in fiberweb uniformity is evident only between the lowest finish level (0.36%) and the
rest of the finish levels. In other words, fiberweb uniformity is affected only at very low levels of finish. On comparing the
effect of finish uniformity on fiberweb uniformity, it is found that finish uniformity did not have any effect on fiberweb
uniformity. This is shown in figure 13.

Figure 8: Change in 2.5% span length of polypropylene fiber w.r.t. finish level (61)

79
 Figure 9: Change in short fiber content of polypropylene fiber w.r.t. finish level (62)

5.2.3. Fiberweb Cohesion

Figure 14 shows the change in drafting force of the polypropylene sliver with respect to finish level. Once again there is a
significant effect of finish level and throughput on the drafting forces. In genreal, the drafting force initially drops with an
increase in finish level and then appears to increase for further increase in finish level and then level off.

On comparing the 0.61% (B) and the 0.61% finish level for any effect of finish uniformity on drafting forces, it is found
that finish uniformity has no effect on the drafting forces of polypropylene fibers. This is shown in figure 15.

Figure 10: Change in 2.5% span length of polypropylene fiber w.r.t. finish uniformity (61)

80
 Figure 11: Change in short fiber content of polypropylene fiber w.r.t. finish Uniformity (63)

Figure 12: Change in CV % of fiberweb basis weight of polypropylene fiber w.r.t. finish level (63)

81
Figure 13: Change in CV % of fiberweb basis weight of polypropylene fiber w.r.t. finish uniformity (64)

Figure 14: Change in drafting force of polypropylene fiber w.r.t. finish level. (65)

82
 Figure 15: Change in drafting force of polypropylene fiber w.r.t. finish uniformity (65)

5.2.4. Neps and Static Charge

The polypropylene fibers owing to their high fiber diameter (9 denier) and hence high bending rigidity, did not generate
any neps during processing on the card. This may be because the higher diameter made them more prone to fiber breakage
rather than being entangled. There was also no static charge generation observed for the range of finish level used in these
experiments.

5.3. Polyester

5.3.1. Fiber Breakage

Figure 16 shows the change in 2.5% span length of polyester fibers with respect to finish level. There is
a significant effect of finish level and its interaction with cylinder to doffer setting. Multiple comparisons
of the means show that the fiber breakage initially increases and then follows a wavy pattern with an
overall drop in fiber breakage for further increase in finish level. The wavy pattern can be due to the
interaction effect of finish and cylinder to doffer setting. Figure 17 shows the change in short fiber
content of polyester fibers with respect to finish level. Here too there is a significant effect of finish level
and the difference in mean effect of finish level is evident only at lower levels of finish. The short fiber
content shows a higher fiber breakage at lower levels. This show that though the fiber breakage in not
evident while observing the 2.5% span length, it is clearly shown in the short fiber content. This may be
because the fiber breakage occurs for fibers whose length is less than the 2.5% span length, as they
spend more time on the cylinder than the fibers with length equal to or greater than 2.5% span length.

Figure 18 and 19 show the change in 2.5% span length and short fiber content with respect to finish uniformity. No
significant effect of finish uniformity on either 2.5% span length or short fiber content could be found.

83
 Figure 16: Change in 2.5% span length of polyester fiber w.r.t. finish level (67)

Figure 17: Change in short fiber content of polyester fiber w.r.t. finish level. (68)

84
 Figure 18: Change in 2.5% span length of polyester fiber w.r.t. finish uniformity (67)

Figure 19: Change in short fiber content of polyester fiber w.r.t. finish uniformity (69)

5.3.2. Fiberweb Uniformity

Figure 20 show the change in CV% of fiberweb basis weight of the polyester fibers with respect to finish level. There is
only a significant effect of throughput and cylinder to doffer setting on the CV% of the fiberweb basis weight and no effect
of finish level. Though for one particular treatment (setting 0.127mm and doffer speed 164.6 m/min), the fiberweb
uniformity appear to improve with an increase in finish level.

On comparing the effect of 0.137% (B) finish level and the 0.137% finish level on fiberweb uniformity
in figure 21, there appear to be no main effect of finish uniformity on fiberweb uniformity. But there is a

85
significant effect of throughput and cylinder to doffer settings along with a slight interaction effect of
finish uniformity and cylinder to doffer setting. This indicates that finish uniformity by itself has no effect
on fiberweb uniformity and maybe at adverse settings they affect the fiberweb uniformity.

5.3.3. Fiberweb Cohesion

Figure 22 shows the change in draft force of polyester sliver with respect to finish level. There is a
significant effect of finish level and its interaction with cylinder to doffer setting on the draft force of
polyester sliver. Overall, the draft force decreases with an increase in finish level. Figure 23 shows the
effect of finish uniformity on draft force. Statistical analysis show no effect of finish uniformity on the
draft force of polyester fibers though the graphs indicate a higher drafting force for the original medium
finish than for the blended medium finish.

5.3.4. Neps
Figure 24 shows the change in nep count of polyester fiber with respect to finish level. Only cylinder to
doffer setting is found to have a significant effect on neps generated. On comparing the effect of finish
uniformity of neps generated (figure 25), there is no significant effect of any of the factors on nep
generation. This shows that nep generation is independent of finish level and its uniformity.

Figure 20: Change in CV % of fiberweb basis weight of polyester fiber w.r.t. finish level (69)

86
Figure 21: Change in CV % of fiberweb basis weight of polyester fiber w.r.t. finish Uniformity.
(70)

Figure 22: Change in draft force polyester fiber w.r.t. finish level (71)

87
Figure 23: Change in draft force polyester fiber w.r.t. finish uniformity (71)

Figure 24: Change neps generated polyester fiber w.r.t. finish level (72)

88
 Figure 25: Change in neps count for polyester fiber w.r.t. finish uniformity (73)

5.4. Comparison Between Fiber Behavior

Finish add-on percentage or finish level has a significant effect on fiber breakage for all the fibers. Both cotton and
polypropylene show a decrease in fiber breakage with an increase in finish level. The fiber breakage for polyester shows a
different trend when compared to polypropylene. Also the short fiber content of polyester shows a reverse trend when
compared to the 2.5% span length. It has to be noted that the finish levels for polyester are much lower when compared to
polypropylene, and the diameter of polyester fiber is also finer (resulting in greater surface area per unit weight of fiber).
Polyester also exhibits an interaction effect of finish level and cylinder to doffer setting on the 2.5% span length. This may
cause the 2.5% span length to show deviations for very low levels of finish add-on percentage. The short fiber content of
polyester shows that fiber breakage decreases with an increase in finish level. The trend of 2.5% span length seem to
agree with that of the short fiber content on ignoring the results of the first two levels of finish add-on percentage. Further
experiments with much higher finish add-on of polyester would help to better explain this deviation.

Finish add-on percentage was found to have significant effect on fiberweb uniformity in the case of polypropylene fibers
and no effect on fiberweb uniformity in the case of polyester fibers. Multiple comparison between the finish levels of
polypropylene fiber show that the significant change in fiberweb uniformity is only evident between the first two levels of
finish. This change is not evident in polyester due to the less range in finish add-on percentage of polyester. Still figure 20
indicates that the fiberweb uniformity increases initially for small increase in finish add-on percentage till it reaches an
optimum level and then decreases for additional increase in finish level. Figure 12 shows that at much higher range of
finish level there is no effect of finish level on fiberweb uniformity.

Finish add-on percentage was found to have a significant effect on fiberweb cohesion of both polyester and polypropylene
fibers. In both cases, the drafting forces decrease with an increase in finish add-on percentage for low levels of finish and
the polypropylene data suggests that at higher finish levels, an increase in finish level causes an increase in drafting
forces. This shows that at low levels of finish the finish performs more as a lubricant decreasing fiber to fiber friction and
thus causing a decrease in fiber cohesion. At higher levels of finish the excess finish may act as a binding agent causing
fiber cohesion to increase.

There is a significant effect of finish level on the neps generated on cotton and not so in the case of polyester. This could
be because cotton generally has a high amount of neps in the feed material than polyester. In this study, the nep count of
cotton before processing was found to be 629 per gram. And since the carding machine is better at eliminating neps than
generating them, the effect of finish level is more evident in cotton than in polyester.

89
Finish uniformity was found to have no effect on any of the response parameters for both polyester and
polypropylene. Although in some cases like fiber breakage and fiberweb uniformity, the original (more
uniform) application show a slightly better performance due to an interaction effect. However, the gain
in performance is not found to be of significant difference. Also the effect of finish uniformity greatly
depends on the extent of blending achieved while mixing two fibers with different finish levels. Though
polyester and polypropylene blends performed as good as the original finish, the blending of cotton
fibers proved unsuccessful. The cotton fibers failed to be carded when the no finish fiber was blended
with the high finish fiber. This could be because of the low levels of blending achieved in the case of
cotton fiber, as the moisture content of cotton was found to be high causing the fibers to form large
lumps. Also, the application of finish to polyester and polypropylene fibers (either by means of a kiss
roll or metered applicator) is bound to be more uniform when compared to the application of finish to
bleached cotton (by agitation in a kier). This calls for an investigation into a more efficient method of
applying a uniform finish on cotton fibers.

6. CONCLUSION

Finish add-on percentage was found to have a significant effect on fiber processability on card and the
influence of finish level depends upon the type of fiber. On the other hand, finish uniformity was found
to have no effect on fiber processability on card. Nevertheless, the degree of blending is also shown to
have an influence on the effect of finish uniformity on fiber processability.

REFERENCES

[1] ASTM D 2612 – 93a, “Standard Test Method for Fiber Cohesion in Sliver and Top in Static Tests”,
Annual Book of ASTM Standards vol 7. p683 (1997)
[2] ASTM D 4120 – 95, “Standard Test Method for Fiber Cohesion in Roving, Sliver, and Top in
Dynamic Tests”, Annual Book of ASTM Standards vol 7. p313 (1997)
[3] Kamath, Yash K., “Spin Finish Distribution - Cause and Effect”, NCRC ShortCourse on fiber
Finish, Feb, (1998)
[4] Meng, J., Seyam, A.M., Batra, S.K., "Carding Dynamics. Part I: Previous Studies of Fiber
Distribution and Movement in Carding", Text. Res. J., 69:2, 90-96 (1999)
[5] Seyam, A.M, El-Sheik, A., Meng, J., "Carding Dynamics. Part II: Development of Device for On-
line Measurement of Two-dimensional Fiber Loading on Card Elements", Text. Res. J., 69:3, 155-
161 (1999)
[6] Tantec Electrostatic Voltmeter, user manual

90
 Chapter 3

“ FINISH PERFORMANCE AS A MECHANICAL PROCESSING AID IN NEEDLEPUNCHING.”

ABSTRACT

This part of the study aims to investigate the role of fiber finish during processing on the needlepunching machine.
Certain key response parameters like dynamic needling forces and needlepunched fabric properties are analyzed for any
possible effect of finish level and its uniformity on these parameters. The experiments are performed using polyester and
polypropylene, two of the most commonly used fibers in needlepunched products, to investigate the difference response
due to fiber type. Results show a highly significant effect finish level on the response parameters namely, dynamic
needling forces and needlepunched fabric properties. Though the results of chapter 2, have shown that finish uniformity
fail to show any effect on carding processability, it was found to have a significant effect on needlepunching forces and
needled fabric properties.

91
1. INTRODUCTION

Needlepunching, a nonwoven technology of fabric formation is used in the manufacturing of several commercial products
such as home and automobile upholstery, carpeting, medical and hygiene products, industrial filters and many other high
performance applications. Structural integrity of the fabric is provided by means of fiber entanglements, caused by the
repeated punching of a board of needles through a layer of loosely held fiberweb. The resultant fabric properties are
determined by the type of fiber and fiber properties such as length, fineness, coefficient of friction and flexural rigidity.
Additionally fiberweb parameters such as basis weight, thickness and number of layers, and machine parameters such as
needling density, draft, depth of penetration of needle, needle design parameters, significantly impact fabric properties [3-
10].

Due to the nature of the feed material, finish applied on the fibers to aid its mechanical processing during the preparatory
process such as carding is carried over to the needlepunching process. Often the finish treatment is not optimized for the
needlepunching process and hence result in differences in processability depending on the finish treatment. Optimization
of the finish treatment will ensure a fiber friction low enough to permit realignment of fibers and high enough to impart
sufficient strength to the fabric after needlepunching. The fiber to metal friction should also be optimized to reduce wear
on the needles. One such study conducted by Wagner and Ahles, on kevlar aramid fibers have shown a considerable
reduction in needle breakage [16]. The study used a more static form of measurement of needling forces and the needle
breakage was recorded over a long period of time. Kevlar by nature is also a very rigid fiber compared to other fibers used
for needlepunching such as polyester and polypropylene. There has also been no study conducted on the effect of finish
uniformity on the needling process. Since the response to finish treatment differ from fiber to fiber, a study of the effect of
friction and its uniformity on the dynamic needling forces and fabric properties should provide a better understanding of
the role of fiber finish in needlepunching.

2. OBJECTIVES

The overall objective of this part of the study is to investigate the role of fiber finish with respect
to processability on the needlepunching machine. Key response parameters like dynamic
needling forces and needlepunched fabric properties are analyzed to assess the finish
performance.

3. EXPERIMENTAL APPROACH

Two fiber types namely, polyester and polypropylene are used to study the effect of finish level and finish uniformity.
Polyester of 3 denier and 38mm is provided by Wellman Inc., and polypropylene of 9 denier and 38mm is provided by
Fibervision Inc. The finish used on both fibers is of a generic kind, formulated by Goulston Tech. The finish is composed
of 90% lubricant (PEG(400) Monolaurate) and 10% antistat (Amine neutralized linear alkyl phosphate).

The effect of finish level is studied by processing different levels of finish on fibers while keeping the rest of the
processing parameters the same. As for finish uniformity, One set of fibers with a finish level A, supplied by the fiber
manufacturer, is compared with a second set of fibers with the same finish level arrived by blending fibers with different
finish level. To illustrate, suppose the fiber manufacturer provides fiber with a finish add-on percentage of 0.137%. This
set of fiber is called the original finish level. Another set of fibers is prepared by blending fibers with 0.038%, 0.11% and
0.22% finish add-on percentage in the proportion 10%, 60% and 30%, respectively. This second set of fibers is called the
blended finish level and is expected to have a finish distribution less uniform than the original finish level. A comparison
of the response parameters for the original and blended finish should provide a good indication of the effect of finish
uniformity on processability.

92
The needling parameters used for the two fibers are given in table 1. The 0.61% of finish add-on in polypropylene and
0.137% finish add-on in polyester is used to study the effect of finish uniformity. The process parameters where chosen by
first preparing a feed batt with the maximum processable basis weight and then selecting the other parameters based on
this. To form a fiberweb, the polypropylene was processed on the garnet due to its large diameter and short fiber length,
before crosslapping. In the case of polyester a revolving flat top card was used to form the fiberweb before crosslapping.
The experiments are performed on a Dilo needle loom at 400 punches/ min. The key response parameters are the dynamic
needling forces and fabric properties such as tensile strength, tear strength and energy to breaking point of the fabric in
both machine and cross machine direction. The dynamic needling forces should provide a good indication of the
resistance to penetration experienced by the needles which in turn is a good indication of the lubricating effect of the
finish (fiber to metal and fiber to fiber friction), needle wear and needle breakage. On the other hand, the tensile strength
and tear strength should provide a good indication of the effect of finish on fiber cohesion (fiber to fiber friction). Five
replicates are taken for each treatment to eliminate bias. The test method used for collecting each of the four response
parameters is described below.

Table 1: Needlepunching process parameters

Parameters Polypropylene Polyester
Finish Levels 0.11%, 0.61% 0.11%, 0.137% 0.22%
Finish Uniformity 0.61%, 0.61% (B)* 0.137%, 0.137%(B)**
2
Fabric Basis Weight 343.68g/m 307.82g/m2
2
Needling Density 58.696pn/cm 58.696pn/cm2
Needle Type GB R 33315x20x40x3.5 GB C 22215x18x36x3
Penetration Depth 8mm 8mm
* 0.61%(B) by blending 27% of 1.97% finish level and 73% of 0.11% finish level
polypropylene
** 0.137%(B) by blending 10% of 0.038% finish level, 60% of 0.111% finish level and
30% of 0.22% finish level polyester.

3.1 Dynamic Needling Forces

Several methods of measurement of needling forces have been proposed [1,2,8]. For this study , the Nonwoven
Cooperative Research Center (NCRC) dynamic needling force measurement device is used. A critical analysis of other
needling force measurement device and the construction and working principle of the NCRC dynamic needling force
device is given by Seyam, et al., [11,12]. As the relative position of needle and needle orientation on the needle board is
known to affect needling forces [13,15], Needling forces of five needles arranged linearly along the length of the machine
and equidistant to each other are collected. A schematic diagram of the device is given in figure 1 and the relative position
of the needles on the needle board is given in figure 2. Here needle position 1 is at the feed of the machine and needle
position 5 is at the exit.

Figure 1: Schematic Diagram of the NCRC Dynamic Needling Force Measuring System

93
The signal analysis of the raw needling force data (shown in figure 3) used earlier is given by Seyam et al [14]. The signal
analysis for this experimentation has some slight modifications. In the earlier method, a fast Fourier transformation was
performed to find the most significant Fourier frequencies and use this to perform a low-level filter to remove any noise at
other frequencies. Then the force due to inertia was separated from the actual needling forces and the relevant
measurements like peak penetration force, peak striping force, penetration energy and striping energy is deduced. This
final step of deducing the forces was done for one cycle at a time for a large set of cycles for each replicate to represent the
variation in the above measurements from cycle to cycle. There is however a shift in the mean of the cyclic wave from
needle to needle based on their location and due external noise. The inertial forces may also vary from one cycle to
another, requiring extreme caution while analyzing the needle forces. The current method used attempts to characterize
the force equation by means of time series analysis.

Figure 2. Locations of the Five-Instrumented Needles in the Needle Board

The voltage readings from the instruments were transformed to force measurements using calibration equations. Then a
moving median filter was used to remove outliers and noise from the raw data. A fast Fourier transformation was
performed and the periodogram was plotted to identify the Fourier frequency with the maximum power. The scan rate
used in this experiment was 400 scans / sec. This provides 60 scans per cycle for a machine speed of 400 pn/min. Time
series analysis was used to arrive at a best-fit model for the curve with relevant harmonics of the sine and cosine function
of the Fourier frequency. The characteristic relationship is given in equation 1. Now the inertial force was separated from
the actual forces and the peak penetration force, peak striping force, penetration energy and striping energy is deduced as
per the previous method [11,12]. The prime concern here is to ensure that the derived force measurements are
representative of all possible forces for a given treatment. Since the model is fitted over a vast number of cycles and the R2
value of the curve is significantly high, the variance of the peak penetration force and peak striping force can be
represented by the square of the standard error of prediction. The variance of the penetration energy and striping energy is
given by equation 2. This reduced to equation 3, if the variance of the coefficients is independent of each other. The
variance of the force measurements can also be taken care of by increasing the number of replicates for each treatment. In
this case a replication of 5 is found to provide a good representation of the force measurements.

94
 Figure 3: Force experienced by needle during needlepunching

f ( x ) = ∑ C i Φ i ( x) (1)

b b

Var ( I ) = ∑ Cov (C i , C j )∫ Φ i ( x )dx ∫ Φ j ( x )dx  (2)
a a 
2
b 
Var ( I ) = ∑ Var (C i )∫ Φ i ( x )dx  (3)
a 
Where,
f (x ) Force Equation
Φ i ( x), Φ j ( x ) Sine / Cosine function with harmonics, i = 0,1,2,…,n , j=0,1,2,….n
Ci , C j Coefficient of sine / cosine term
Var (C i ) Variance of Coefficient
Cov (C i , C j ) Covariance of Coefficients
Var (I ) Variance of integral (penetration / striping energy)
a, b Boundary conditions for x.

This method utilizes a large number of cycles to characterize the curve and provide a good representation of the force
measurements. The method is also less time consuming and more standardized. A sample calculation is provided in
appendix IV. In this study, only the peak penetration force and penetration energy is used to compare the finish
performance. The peak striping force and striping energy is not compared, due to their small magnitude and high
variance.

95
3.2 Fabric Properties
The tensile strength and energy to breaking point of the fabric was measured on the Instron Tensile tester based on the
ASTM standard test D 5034-95. The tear strength of the needlepunched fabric was also tested on the Instron, using the
trapezoidal method as recommended in the ASTM standard test D 5733-95.

4. RESULTS AND DISCUSSION

The statistical results for all the response parameters observed are given in appendix III. The page
number of the statistical results for the corresponding parameter under discussion is given between
parentheses, following the title of the graph.

4.1 Polypropylene

4.1.1 Dynamic Needling Forces

The change in peak penetration force and penetration energy for each needle position with respect to finish level is given
in figure 4 and 5 respectively. The relative distance of needle position two, three, four and five (last row) to the needle
position one (first row) in cm is 4.3, 8.6, 12.9 and 17.2 respectively. At low needling density, such as the one used here
the needling forces increases from position one to five due to the increase in resistance by the fibers to reorientation as
they are already entangled to a certain degree by the previous needles. This behavior is explained recently [15]. There is a
significant effect of finish level on needle penetration forces of all instrumented needles. The high (0.61%) finish level
fibers displayed less resistance to needle penetration.

On comparing the needling forces with respect to finish uniformity. There is a significant effect of finish uniformity on
the needling forces (figure 6, 7). The blended finish shows less resistance to needling than the original medium finish.
This could be because the excess finish from the high finish add-on fibers of the blended finish got relocated during
carding providing good overall lubrication or this excess finish forms a coating on the needles causing a drop in needling
force once steady state is achieved.

Figure 4: Change in peak penetration force of polypropylene fibers w.r.t. finish level (75)

96
 Figure 5: Change in penetration energy of polypropylene fibers w.r.t. finish level (76)

4.1.2 Fabric Properties

The change in tensile strength, tear strength and Energy to breaking point of polypropylene needlepunched fabric is given
in figure 8, 9 and 10, respectively. There is a significant effect of finish level, direction of fabric and their interaction on
all the three fabric properties tested. The higher finish level provided better fabric performance. The high strength in the
machine direction when compared to cross machine direction is contrary to usual finding. This could due to the
reorientation of the fibers while testing on the Instron, as a result of their short fiber length and greater fiber diameter.
These two polypropylene fiber parameters cause the fiber-to-fiber friction to reduce significantly and lead to easy
reorientation of fibers during testing.

On comparing the fabric properties with respect to finish uniformity (figure 11, 12 and 13), there is a significant effect of
finish uniformity and direction on all the three fabric properties. The needled fabric with the original finish level (0.61%)
shows a higher strength, than the needled fabric with the blended medium finish (0.61%(B)).

The greater strength of the fabric in the machine direction than in cross machine direction indicates that the fiber
orientation during testing is more towards the machine direction than the cross machine direction. Usually a carded,
cross-lapped and needlepunched fabric shows greater orientation in the cross machine direction. Here the deviation could
be due to the use of the garnet to form the initial fiberweb (due to the higher diameter and smaller length of polypropylene
fiber). Also the fibers could have realigned while testing due to more freedom of movement.

97
Figure 6: Change in peak penetration force of polypropylene fibers w.r.t. finish uniformity (75)

Figure 7: Change in penetration energy of polypropylene fibers w.r.t. finish uniformity (76)

Figure 8: Change in tensile strength of polypropylene fabric w.r.t. finish level (77)

98
 Figure 9: Change in tear strength of polypropylene fabric w.r.t. finish level (78)

Figure 10: Change in energy to breaking point of polypropylene fabric w.r.t. finish level (79)

Figure 11: Change in tensile strength of polypropylene fabric w.r.t. finish Uniformity (77)

99
 Figure 12: Change in tear strength of polypropylene fabric w.r.t. finish uniformity (78)

Figure 13: Change in energy to breaking point of polypropylene fabric w.r.t. finish uniformity (80)

4.2 Polyester

4.2.1 Dynamic Needling Forces

The change in peak penetration force and penetration energy for each needle position with respect to finish level is given
in figures 14 and 15 respectively. There is a significant effect of finish level and needle position on the needling forces.
The difference in needling forces between fibers with different finish levels increases with an increase in distance of the
needle position from the feed. On comparing the needling forces with respect to finish uniformity (figure 16 & 17), the
effect of finish uniformity was not statistically significant on any of the needling forces.

100
Figure 14: Change in peak penetration force of polyester fibers w.r.t. finish level (81)

Figure 15: Change in penetration energy of polyester fibers w.r.t. finish level (82)

101
 Figure 16: Change in peak penetration force of polyester fibers w.r.t. finish uniformity (81)

Figure 17: Change in penetration energy of polyester fibers w.r.t. finish uniformity (83)

4.2.2 Fabric Properties

The change in fabric properties of polyester needlepunched fabric is shown in figure 18, 19 and 20. There is a significant
influence of finish level and direction of test on all the three fabric properties namely, tensile strength, tear strength and
tensile energy to breaking point. There is an initial drop in the strength of the fabric properties for an increase on finish
level and then the strength seems to improve for further increase in finish level.

The change in tensile strength, tear strength and energy to breaking point of polyester needlepunched fabric with respect
to finish uniformity is shown in figure 21, 22 and 23. There is once again a significant effect of finish uniformity on
tensile strength and energy to breaking point, but there was no effect of finish uniformity on the tear strength.

102
 Figure 18: Change in tensile strength of polyester fabric w.r.t. finish level (83)

Figure 19: Change in tear strength of polyester fabric w.r.t. finish level (84)

Figure 20: Change in energy to breaking point of polyester fabric w.r.t. finish level (85)

103
 Figure 21: Change in tensile strength of polyester fabric w.r.t. finish Uniformity (84)

Figure 22: Change in tear strength of polyester fabric w.r.t. finish uniformity (85)

Figure 23: Change in energy to breaking point of polyester fabric w.r.t. finish uniformity (86)

104
5. CONCLUSION

The influence of finish level on needling forces for both polypropylene and polyester suggest that, the needling forces can
be altered for optimum needling efficiency along with reduced needle wear. The response to finish level was also found to
vary with the type of fiber. Polyester fiber in spite of its smaller range of finish levels tends to show large variation in its
response to change in finish level than polypropylene fibers. It has to be noted that the polypropylene fibers used here are
of a larger diameter than polyester fibers, which might even further increase the difference in resilience between the two
fibers. The polyester data also shows that the difference in finish performance changes with respect to needle position.
The further the needle is from the feed the greater is the difference in peak penetration forces with respect to finish level.
As for finish uniformity only polypropylene shows a strong influence of finish uniformity on needling forces. A closer
look, at the blend composition of the 0.61%(B) finish level of polypropylene shows that the finish level of the high finish
component is nearly three times 0.61% (refer table 1). This excess finish could lead to a coating of the needles while
needling and hence result in a significantly better lubrication that the original 0.61%.

The fabric properties of both polypropylene and polyester needlepunched fabric show a significant influence of finish add-
on percentage and finish uniformity on needlepunched fabric properties. The difference in lubrication can influence the
freedom of movement of fibers with respect to one another. This in turn determine the breaking mechanism of fibers while
testing for tensile or tear strength. Greater the freedom of movement, more is the tendency of the fibers to realign and
break in groups while testing, resulting in greater effective strength. The effect of finish uniformity on fabric properties
show that the fabric with the original finish have better strength. This shows that lack of uniformity also reduces the
tendency of fibers to break in groups rather than individuals.

REFERENCE

[1] Debnath, C.R., “Needle-puching of Jute Webs”, Ind. Textile J. 89, 137 (1978)
[2] Hearle, J. W. S., Purdy, A. T., “A Technique for the Measurement of the Punching Force During Needling”, J.
Textile Inst. 63, 363 (1972)
[3] Hearle, J. W. S., & Sultan, M. A. I., “Study of Needled Fabrics, Part I: Experimental Methods and Properties”, J.
Text. Inst., 58, 251 (1967)
[4] Hearle, J. W. S., & Sultan, M. A. I., “Study of Needled Fabrics, Part II: Effect of the Needling Process”, J. Text.
Inst., 59, 103 (1968)
[5] Hearle, J. W. S., & Sultan, M. A. I., “Study of Needled Fabrics, Part III: The Influence of Fiber Type and
Dimensions.”, J. Text. Inst., 59, 137 (1968)
[6] Hearle, J. W. S., & Sultan, M. A. I., “Study of Needled Fabrics, Part IV: The effect of Strength Shrinkage and
Reinforcement.”, J. Text. Inst., 59, 161 (1968)
[7] Hearle, J. W. S., & Sultan, M. A. I., “Study of Needled Fabrics, Part V: The approach to Theoretical
understanding.”, J. Text. Inst., 59, 183 (1968)
[8] Hearle, J. W. S., & Sultan, M. A. I., “Study of Needled Fabrics, Part VI: The Measurement of Punching Force
During Needling”, J. Text. Inst., 59, 237 (1968)
[9] Hearle, J. W. S., & Sultan, M. A. I., “Study of Needled Fabrics, Part VII: The Transfer of fiber through the Web by
Needling.”, J. Text. Inst., 60, 478 (1969)
[10] Hearle, J. W. S., & Sultan, M. A. I., “Study of Needled Fabrics, Part VIII: The effect of Friction on the Processing
ad Properties of Needle-bonded Fabrics”, J. Text. Inst., 61, 83 (1971)
[11] Sarin, S., Meng, J., and Seyam, A.M., “Mechanics of Needlepunching Process and Products, Part I: Critical
Review of Previous Work on Forces Experienced by Needles During Needling of Nonwoven Fabrics”, Int.
Nonwovens J. 6 (2), 32-36 (1994)
[12] Seyam, A. M., Meng, J., and Mohamed A., “Mechanics of Needlepunching Process and Products, Part II: An On-
Line Device to Measure the Punching Forces Experienced by Individual Needles”, Int. Nonwovens J. 7 (3), 31-37
(1995)
[13] Seyam, A. M., and Sarin, S., “Effect of Needle Position and Orientation on Forces Experienced by Individual
Needles During Needle Punching”, Textile Res. J. 67, 772-776 (1997)
[14] Seyam, A. M., Mohamed, A., Kim, H., “Signal Analysis of Dynamic Forces Experienced by Individual Needles in
High Speed Needle Punching”, Test. Research J., 68(4), 296-31 (1998)

105
[15] Seyam, A.M., “Application of On-line Monitoring of Dynamic Forces Experienced by Needles During Formation
of Needled Fabrics”, Accepted for publication by Nonwovens International Journal
[16] Wagner, J.R., Ahles, L.J., “The effect of Fiber Finish and Needle Type on Needle Breakage and Compaction of
Needle Punched Kevlar Aramid Fiber Nonwovens”, Nonwovens Conference (1995)

106
 Chapter 4

a) “CONCLUSION AND FUTURE WORK”

b) CONCLUSION

The experiments conducted on carding and needlepunching machine provided a better mean of
evaluating finish performance when compared to conventional methods. Most conventional methods
relied more on translating fundamental properties of fibers and their frictional behavior under the
influence of a surface coating over the response parameters by means of a complex model of the
geometry and normal forces acting over the fibers during processing. These experiments on the other
hand rely upon direct changes in the responses to any change in the finish or processing parameters.
This helps in formulating empirical equations describing the response parameters with relation to finish
and process parameters.

The results of the experiments conducted on the carding machine and the needlepunching machine, to
evaluate finish performance shows that the finish performance is highly dependent on the type of fiber
and fiber properties. Hence cross fiber comparison in not meaningful unless the fibers are very similar in
terms of their properties and processing capabilities. Finish add-on percentage was found to greatly
influence all the response parameters in both carding and needlepunching. It is also found that a higher
finish level need not necessarily result in better performance. Many of the improvement in performance
of finish appear to level off after reaching a certain peak performance. In some cases like fiberweb
cohesion and needlepunched fabric properties the response to increase in finish level was found to
reverse in trend for further increase in finish level. Though finish uniformity was found to have little or
no influence on the carding response parameters, it was found to have a very significant effect on the
needlepunched fabric properties. Hence, the effect of finish treatment on the resultant needlepunched
fabric properties should also be taken into consideration during selection of a finish treatment for a
particular fiber.

c) SCOPE FOR FUTURE WORK

During the course of this work, it was found that the finish extraction results varied in term of accuracy
depending on the type of fibers used. This to great extent affected the accuracy of the projection of the
response parameters based on change in finish levels. A more accurate method to extract finish would
ensure a better projection of response parameters based on changes in finish level. Also, the speed of the
carding machine and the needlepunching machine used in these experiments are slow compared to
current industry standards. Since the frictional behavior of viscoelastic materials like textiles, changes
with respect to speed of operation, these results may not be applicable to very high speed operation.
Further experiments will have to be conducted to study the finish performance at such speeds. The
range of finish fibers used here was also limited. Most of the finish levels for each fiber were arrived by
means of blending three basic finish levels. To eliminate the effect of this blending operation, it would be

107
desirable to obtain the fibers with a wide range of finish add-on percentage directly from the fiber
manufacturer.

Since the drafting force (to a great extent) and the needlepunched fabric properties (to a certain extent)
rely upon the frictional behavior of fiber, a study of the correlation between the two responses should be
helpful. If a significantly high correlation is found, this could be used to predict the needlepunched fabric
properties of the fibers based on their drafting forces. Here the carding machines used for preparing the
sliver for drafting force measurements were different from the carding machines used in the preparatory
process of needlepunching. Hence no correlation study between drafting forces and needlepunched
fabric properties could be conducted.

The lack of clarity in terms of finish treatment to enhance fiber processability has lead many companies
to closely guard their methods of finish treatments to gain a competitive advantage. Hence, the finish
composition of most of the proven finish treatments is not known. This renders the effectiveness of the
generic finish treatments used under such study very debatable. So, a better interpretation of the test
results of such study on finish performance could be arrived by means of using a more proven method of
finish treatment with all information available to the research team.

108
APPENDICES

109
 Appendix I: Finish Extraction Results for Cotton

Sample # Fiber finish % Weight loss

water soluble solvent soluble
1 high 0.342 0.032
2 high 0.115 0.000
3 high 0.287 0.000
4 high 1.460 6.460
5 high 0.698 7.040
6 high 1.240 7.800
Average 0.690 3.550
1 low 0.441 6.190
2 low 0.000 6.720
3 low 0.000 7.340
4 low 0.631 0.113
5 low 1.113 0.000
6 low 1.289 0.000
Average 0.579 3.390
1 none 0.548 0.000
2 none 0.829 0.000
3 none 0.196 0.169
4 none 0.808 0.000
5 none 0.509 0.214
6 none 0.675 0.118
Average 0.594 0.084

110
 Appendix II : Statistical Results of Response Parameters of Carding Experiments Performed in
Chapter 2

The analysis of variance (ANOVA) for the parameters tested in chapter 2, are performed using the SAS library. The
model is setup in the form of a randomized block design. The ANOVA results are as given below. For orthogonal results,
the type III sum of squares are not reported as they are the same as the type I sum of squares. The page number containing
the discussion of the corresponding table and the figure number of the corresponding graph is given in brackets for each
ANOVA result.

COTTON

.2 ANOVA FOR EFFECTIVE LENGTH OF COTTON (page 24, figure 5)

(i) General Linear Models Procedure

Class Level Information

Class Levels Values

BLEND 4 h100l0 h25l75 h50l50 h75l25
SET 2 59
THRU 2 50 100
Number of observations in data set = 48

Dependent Variable: EFLEN Effective Length (mm)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 15 17.07389058 1.13825937 8.38 0.0001
Error 32 4.34407733 0.13575242
Corrected Total 47 21.41796792

R-Square C.V. Root MSE EFLEN Mean

0.797176 1.267108 0.36844595 29.07770833

Source DF Type I SS Mean Square F Value Pr > F

BLEND 3 15.98249492 5.32749831 39.24 0.0001
SET 1 0.06586008 0.06586008 0.49 0.4911
THRU 1 0.16263408 0.16263408 1.20 0.2819
BLEND*SET 3 0.26209625 0.08736542 0.64 0.5927
BLEND*THRU 3 0.33736492 0.11245497 0.83 0.4881
SET*THRU 1 0.00134408 0.00134408 0.01 0.9214
BLEND*SET*THRU 3 0.26209625 0.08736542 0.64 0.5927

111
.3 ANOVA FOR SHORT FIBER CONTENT OF COTTON (page 24, figure 6)

General Linear Models Procedure

Class Level Information

Class Levels Values

BLEND 4 h100l0 h25l75 h50l50 h75l25
SET 2 59
THRU 2 50 100
Number of observations in data set = 48

Dependent Variable: SFC Short Fiber Content (%)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 15 85.62000000 5.70800000 5.88 0.0001
Error 32 31.06666667 0.97083333
Corrected Total 47 116.68666667
R-Square C.V. Root MSE SFC Mean
0.733760 7.559914 0.98530875 13.03333333

Source DF Type I SS Mean Square F Value Pr > F

BLEND 3 77.06000000 25.68666667 26.46 0.0001
SET 1 0.03000000 0.03000000 0.03 0.8616
THRU 1 0.90750000 0.90750000 0.93 0.3409
BLEND*SET 3 3.43000000 1.14333333 1.18 0.3336
BLEND*THRU 3 1.97583333 0.65861111 0.68 0.5717
SET*THRU 1 0.80083333 0.80083333 0.82 0.3705
BLEND*SET*THRU 3 1.41583333 0.47194444 0.49 0.6943

ANOVA FOR NEPS PER GRAM OF COTTON (page 25, figure 7)

General Linear Models Procedure

Class Level Information

Class Levels Values

BLEND 4 h100l0 h25l75 h50l50 h75l25
SET 2 59
THRU 2 50 100
Number of observations in data set = 48

Dependent Variable: SQNEP, (Square root of neps per gram)

Source DF Sum of Squares Mean Square F Value Pr > F

112
Model 15 138.31296878 9.22086459 23.40 0.0001
Error 32 12.60817151 0.39400536
Corrected Total 47 150.92114029
R-Square C.V. Root MSE SQNEP Mean
0.916459 4.902100 0.62769846 12.80468571

Source DF Type I SS Mean Square F Value Pr > F

BLEND 3 20.41635222 6.80545074 17.27 0.0001
SET 1 5.72655428 5.72655428 14.53 0.0006
THRU 1 104.46381998 104.46381998 265.13 0.0001
BLEND*SET 3 2.08747993 0.69582664 1.77 0.1735
BLEND*THRU 3 3.07052012 1.02350671 2.60 0.0694
SET*THRU 1 1.07118716 1.07118716 2.72 0.1090
BLEND*SET*THRU 3 1.47705509 0.49235170 1.25 0.3081

113
 POLYPROPYLENE

ANOVA FOR 2.5% SPAN LENGTH OF POLYPROPYLENE (page 26, figure 8)

General Linear Models Procedure

Class Level Information

Class Levels Values

FLEVEL 5 0.36 0.61 0.86 1.11 1.36
SET 2 12
THRU 2 12
Number of observations in data set = 60

Dependent Variable: S25 2.5% Span length(mm)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 19 39.80933333 2.09522807 3.02 0.0016
Error 40 27.78000000 0.69450000
Corrected Total 59 67.58933333

R-Square C.V. Root MSE S25 Mean

0.588988 2.466555 0.83336667 33.78666667

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 4 31.64433333 7.91108333 11.39 0.0001
SET 1 0.03266667 0.03266667 0.05 0.8294
FLEVEL*SET 4 4.11900000 1.02975000 1.48 0.2256
THRU 1 0.11266667 0.11266667 0.16 0.6893
FLEVEL*THRU 4 1.20233333 0.30058333 0.43 0.7841
SET*THRU 1 0.17066667 0.17066667 0.25 0.6228
FLEVEL*SET*THRU 4 2.52766667 0.63191667 0.91 0.4675

General Linear Models Procedure (page 27, figure 10)

Class Level Information

Class Levels Values

FLEVEL 2 Blended Original
SET 2 12
THRU 2 12
Number of observations in data set = 24

Dependent Variable: S25 2.5% Span length(mm)

Source DF Sum of Squares Mean Square F Value Pr > F

114
Model 7 5.43833333 0.77690476 0.85 0.5614
Error 16 14.56666667 0.91041667
Corrected Total 23 20.00500000

R-Square C.V. Root MSE S25 Mean

0.271849 2.915684 0.95415757 32.72500000

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 3.37500000 3.37500000 3.71 0.0721
SET 1 0.00166667 0.00166667 0.00 0.9664
FLEVEL*SET 1 0.20166667 0.20166667 0.22 0.6442
THRU 1 0.60166667 0.60166667 0.66 0.4282
FLEVEL*THRU 1 0.00166667 0.00166667 0.00 0.9664
SET*THRU 1 0.88166667 0.88166667 0.97 0.3397
FLEVEL*SET*THRU 1 0.37500000 0.37500000 0.41 0.5301

ANOVA FOR SHORT FIBER CONTENT OF POLYPROPYLENE (page 26, figure 9)

General Linear Models Procedure

Class Level Information

Class Levels Values

FLEVEL 5 0.36 0.61 0.86 1.11 1.36
SET 2 12
THRU 2 12
Number of observations in data set = 60

Dependent Variable: SFC Short Fiber Content (%)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 19 9.12066667 0.48003509 2.59 0.0056
Error 40 7.41333333 0.18533333
Corrected Total 59 16.53400000

R-Square C.V. Root MSE SFC Mean

0.551631 25.17565 0.43050358 1.71000000

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 4 3.43733333 0.85933333 4.64 0.0036
SET 1 0.09600000 0.09600000 0.52 0.4759
FLEVEL*SET 4 1.43733333 0.35933333 1.94 0.1227
THRU 1 1.94400000 1.94400000 10.49 0.0024
FLEVEL*THRU 4 0.21266667 0.05316667 0.29 0.8848
SET*THRU 1 0.64066667 0.64066667 3.46 0.0704
FLEVEL*SET*THRU 4 1.35266667 0.33816667 1.82 0.1431

115
 General Linear Models Procedure (page 27, figure 11)
Class Level Information

Class Levels Values

FLEVEL 2 Blended Original
SET 2 12
THRU 2 12
Number of observations in data set = 24

Dependent Variable: SFC Short Fiber Content (%)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 7 2.53333333 0.36190476 1.48 0.2425
Error 16 3.90666667 0.24416667
Corrected Total 23 6.44000000

R-Square C.V. Root MSE SFC Mean

0.393375 29.06660 0.49413224 1.70000000

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 1.60166667 1.60166667 6.56 0.0209
SET 1 0.00166667 0.00166667 0.01 0.9352
FLEVEL*SET 1 0.06000000 0.06000000 0.25 0.6268
THRU 1 0.32666667 0.32666667 1.34 0.2644
FLEVEL*THRU 1 0.37500000 0.37500000 1.54 0.2331
SET*THRU 1 0.00166667 0.00166667 0.01 0.9352
FLEVEL*SET*THRU 1 0.16666667 0.16666667 0.68 0.4208

ANOVA FOR SQUARE ROOT OF CV% OF FBERWEB BASIS WEIGHT OF POLYPROPYLENE

(page 28, figure 12)

General Linear Models Procedure

Class Level Information

Class Levels Values

FLEVEL 5 0.36 0.61 0.86 1.11 1.36
SET 2 12
THRU 2 12
Number of observations in data set = 60

116
Dependent Variable: SWCV, SQRT CV of fiberweb basis weight

Source DF Sum of Squares Mean Square F Value Pr > F

Model 19 0.27104115 0.01426532 5.08 0.0001
Error 40 0.11237338 0.00280933
Corrected Total 59 0.38341453

R-Square C.V. Root MSE SWCV Mean

0.706914 1.773751 0.05300316 2.98819574

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 4 0.06210988 0.01552747 5.53 0.0012
SET 1 0.00279469 0.00279469 0.99 0.3246
FLEVEL*SET 4 0.00150722 0.00037681 0.13 0.9689
THRU 1 0.17631143 0.17631143 62.76 0.0001
FLEVEL*THRU 4 0.01795026 0.00448757 1.60 0.1938
SET*THRU 1 0.00179379 0.00179379 0.64 0.4290
FLEVEL*SET*THRU 4 0.00857387 0.00214347 0.76 0.5556

General Linear Models Procedure (page 28, figure 13)

Class Level Information

Class Levels Values

FLEVEL 2 Blended Original
SET 2 12
THRU 2 12
Number of observations in data set = 24

Dependent Variable: SWCV, SQRT CV of fiberweb basis weight

Source DF Sum of Squares Mean Square F Value Pr > F

Model 7 0.08418223 0.01202603 3.34 0.0216
Error 16 0.05754334 0.00359646
Corrected Total 23 0.14172557

R-Square C.V. Root MSE SWCV Mean

0.593981 2.016155 0.05997048 2.97449794

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 0.01466723 0.01466723 4.08 0.0605
SET 1 0.00378273 0.00378273 1.05 0.3204
FLEVEL*SET 1 0.00081445 0.00081445 0.23 0.6406
THRU 1 0.04924993 0.04924993 13.69 0.0019
FLEVEL*THRU 1 0.00016559 0.00016559 0.05 0.8328
SET*THRU 1 0.00055630 0.00055630 0.15 0.6993

117
FLEVEL*SET*THRU 1 0.01494600 0.01494600 4.16 0.0584

ANOVA FOR DRAFTING FORCE OF POLYPROPYLENE (page 29, figure 14)

General Linear Models Procedure

Class Level Information

Class Levels Values

FLEVEL 5 0.36 0.61 0.86 1.11 1.36
SET 2 12
THRU 2 12
Number of observations in data set = 55

Dependent Variable: DFORCE, Drafting Force (N)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 19 18.76093939 0.98741786 6.29 0.0001
Error 35 5.49833333 0.15709524
Corrected Total 54 24.25927273

R-Square C.V. Root MSE DFORCE Mean

0.773351 11.31260 0.39635242 3.50363636

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 4 2.03124242 0.50781061 3.23 0.0233
SET 1 0.06452263 0.06452263 0.41 0.5258
FLEVEL*SET 4 0.34150768 0.08537692 0.54 0.7049
THRU 1 15.20473529 15.20473529 96.79 0.0001
FLEVEL*THRU 4 0.72391454 0.18097863 1.15 0.3486
SET*THRU 1 0.02427028 0.02427028 0.15 0.6967
FLEVEL*SET*THRU 4 0.37074655 0.09268664 0.59 0.6721

General Linear Models Procedure (page 29, figure 15)

Class Level Information

Class Levels Values

FLEVEL 2 Blended Original
SET 2 12
THRU 2 12
Number of observations in data set = 24

Dependent Variable: DFORCE, Drafting Force (N)

118
Source DF Sum of Squares Mean Square F Value Pr > F
Model 7 5.40375000 0.77196429 2.59 0.0544
Error 16 4.76583333 0.29786458
Corrected Total 23 10.16958333

R-Square C.V. Root MSE DFORCE Mean

0.531364 16.03240 0.54576972 3.40416667

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 0.51041667 0.51041667 1.71 0.2090
SET 1 0.00041667 0.00041667 0.00 0.9706
FLEVEL*SET 1 0.07041667 0.07041667 0.24 0.6334
THRU 1 4.20042857 4.20042857 14.10 0.0017
FLEVEL*THRU 1 0.24354202 0.24354202 0.82 0.3793
SET*THRU 1 0.37284759 0.37284759 1.25 0.2797
FLEVEL*SET*THRU 1 0.00568182 0.00568182 0.02 0.8919

Source DF Type III SS Mean Square F Value Pr > F

FLEVEL 1 0.75578283 0.75578283 2.54 0.1307
SET 1 0.03457071 0.03457071 0.12 0.7378
FLEVEL*SET 1 0.00911616 0.00911616 0.03 0.8633
THRU 1 4.18305556 4.18305556 14.04 0.0018
FLEVEL*THRU 1 0.22790404 0.22790404 0.77 0.3947
SET*THRU 1 0.36972222 0.36972222 1.24 0.2817
FLEVEL*SET*THRU 1 0.00568182 0.00568182 0.02 0.8919

119
 (a) POLYESTER

.4 ANOVA FOR 2.5% SPAN LENGTH OF POLYESTER (page 30, figure 16)

General Linear Models Procedure

Class Level Information

Class Levels Values

FLEVEL 5 0.111 0.137 0.175 0.222 0.0745
SET 2 59
THRU 2 90 180
Number of observations in data set = 40

Dependent Variable: S25 2.5% Span length(mm)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 19 35.35100000 1.86057895 4.94 0.0004
Error 20 7.54000000 0.37700000
Corrected Total 39 42.89100000

R-Square C.V. Root MSE S25 Mean

0.824206 1.789314 0.61400326 34.31500000

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 4 17.20850000 4.30212500 11.41 0.0001
SET 1 0.28900000 0.28900000 0.77 0.3917
THRU 1 0.22500000 0.22500000 0.60 0.4488
FLEVEL*SET 4 12.70350000 3.17587500 8.42 0.0004
SET*THRU 1 0.00900000 0.00900000 0.02 0.8788
FLEVEL*THRU 4 1.81250000 0.45312500 1.20 0.3408
FLEVEL*SET*THRU 4 3.10350000 0.77587500 2.06 0.1247

General Linear Models Procedure (page 31, figure 18)

Class Level Information

Class Levels Values

FLEVEL 2 Blended Original
SET 2 59
THRU 2 90 180
Number of observations in data set = 16

Dependent Variable: S25 2.5% Span length(mm)

120
Source DF Sum of Squares Mean Square F Value Pr > F
Model 7 11.17750000 1.59678571 4.60 0.0240
Error 8 2.78000000 0.34750000
Corrected Total 15 13.95750000
R-Square C.V. Root MSE S25 Mean
0.800824 1.719260 0.58949131 34.28750000

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 1.32250000 1.32250000 3.81 0.0869
SET 1 7.02250000 7.02250000 20.21 0.0020
FLEVEL*SET 1 1.32250000 1.32250000 3.81 0.0869
THRU 1 0.36000000 0.36000000 1.04 0.3386
FLEVEL*THRU 1 0.09000000 0.09000000 0.26 0.6245
SET*THRU 1 0.25000000 0.25000000 0.72 0.4210
FLEVEL*SET*THRU 1 0.81000000 0.81000000 2.33 0.1653

ANOVA FOR SHORT FIBER CONTENT OF POLYESTER

(page 31, figure 17)

General Linear Models Procedure

Class Level Information

Class Levels Values

FLEVEL 5 0.111 0.137 0.175 0.222 0.0745
SET 2 59
THRU 2 90 180

Number of observations in data set = 40

Dependent Variable: SFC, Short Fiber Content (%)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 19 14.01100000 0.73742105 2.11 0.0532
Error 20 7.00000000 0.35000000
Corrected Total 39 21.01100000

R-Square C.V. Root MSE SFC Mean

0.666841 37.80243 0.59160798 1.56500000

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 4 5.45850000 1.36462500 3.90 0.0169
SET 1 0.06400000 0.06400000 0.18 0.6735
THRU 1 0.96100000 0.96100000 2.75 0.1131
FLEVEL*SET 4 0.82850000 0.20712500 0.59 0.6725

121
SET*THRU 1 0.10000000 0.10000000 0.29 0.5989
FLEVEL*THRU 4 2.48150000 0.62037500 1.77 0.1740
FLEVEL*SET*THRU 4 4.11750000 1.02937500 2.94 0.0461

General Linear Models Procedure (page 32, figure 19)

Class Level Information

Class Levels Values

FLEVEL 2 Blended Original
SET 2 59
THRU 2 90 180
Number of observations in data set = 16

Dependent Variable: SFC Short Fiber Content (%)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 7 1.35750000 0.19392857 1.08 0.4542
Error 8 1.44000000 0.18000000
Corrected Total 15 2.79750000

R-Square C.V. Root MSE SFC Mean

0.485255 36.49583 0.42426407 1.16250000

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 0.06250000 0.06250000 0.35 0.5720
SET 1 0.06250000 0.06250000 0.35 0.5720
FLEVEL*SET 1 0.12250000 0.12250000 0.68 0.4333
THRU 1 0.12250000 0.12250000 0.68 0.4333
FLEVEL*THRU 1 0.00250000 0.00250000 0.01 0.9091
SET*THRU 1 0.56250000 0.56250000 3.12 0.1151
FLEVEL*SET*THRU 1 0.42250000 0.42250000 2.35 0.1640
ANOVA FOR SQRT OF COEFFICIENT OF VARIATION OF POLYESTER WEB
(page 33, figure 20)

General Linear Models Procedure

Class Level Information

Class Levels Values

FLEVEL 5 0.111 0.137 0.222 0.0745 0.1795
SET 2 59
THRU 2 90 180
Number of observations in data set = 60

Dependent Variable: SWCV

122
Source DF Sum of Squares Mean Square F Value Pr > F
Model 19 2.70686118 0.14246638 4.03 0.0001
Error 40 1.41318045 0.03532951
Corrected Total 59 4.12004163

R-Square C.V. Root MSE SWCV Mean

0.656999 4.173351 0.18796146 4.50385008

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 4 0.16896306 0.04224076 1.20 0.3276
SET 1 0.71090025 0.71090025 20.12 0.0001
FLEVEL*SET 4 0.10872830 0.02718207 0.77 0.5515
THRU 1 1.53187383 1.53187383 43.36 0.0001
FLEVEL*THRU 4 0.04419850 0.01104963 0.31 0.8678
SET*THRU 1 0.06017487 0.06017487 1.70 0.1993
FLEVEL*SET*THRU 4 0.08202238 0.02050559 0.58 0.6785

General Linear Models Procedure (page 33, figure 21)

Class Level Information

Class Levels Values

FLEVEL 2 Blended original
SET 2 59
THRU 2 90 180
Number of observations in data set = 24

Dependent Variable: SWCV, Sqrt of CV of fiberweb basis weight

Source DF Sum of Squares Mean Square F Value Pr > F

Model 7 0.96307559 0.13758223 4.24 0.0080
Error 16 0.51934413 0.03245901
Corrected Total 23 1.48241972

R-Square C.V. Root MSE SWCV Mean

0.649665 4.060825 0.18016384 4.43663186

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 0.00244083 0.00244083 0.08 0.7874
SET 1 0.21264505 0.21264505 6.55 0.0210
FLEVEL*SET 1 0.21259444 0.21259444 6.55 0.0210
THRU 1 0.49893896 0.49893896 15.37 0.0012
FLEVEL*THRU 1 0.00058084 0.00058084 0.02 0.8953
SET*THRU 1 0.00060930 0.00060930 0.02 0.8927
FLEVEL*SET*THRU 1 0.03526617 0.03526617 1.09 0.3127

123
 ANOVA FOR DRAFT FORCE MEASUREMENTS OF POLYESTER
(page 34, figure 22)

General Linear Models Procedure

Class Level Information

Class Levels Values

FLEVEL 5 0.111 0.137 0.222 0.0745 0.1795
SET 2 59
THRU 2 12

Number of observations in data set = 40

General Linear Models Procedure

Dependent Variable: AVG Average Draft force (N)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 19 45.16441000 2.37707421 2.25 0.0395
Error 20 21.09750000 1.05487500
Corrected Total 39 66.26191000
R-Square C.V. Root MSE AVG Mean
0.681604 19.48347 1.02707108 5.27150000

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 4 20.97733500 5.24433375 4.97 0.0060
SET 1 1.42129000 1.42129000 1.35 0.2594
THRU 1 1.30321000 1.30321000 1.24 0.2795
FLEVEL*SET 4 12.65903500 3.16475875 3.00 0.0432
SET*THRU 1 0.03721000 0.03721000 0.04 0.8529
FLEVEL*THRU 4 2.72491500 0.68122875 0.65 0.6363
FLEVEL*SET*THRU 4 6.04141500 1.51035375 1.43 0.2601

General Linear Models Procedure (page 34, figure 23)

Class Level Information

Class Levels Values

FLEVEL 2 Blended Original
SET 2 59
THRU 2 90 180

124
 Number of observations in data set = 16

Dependent Variable: AVG Average Draft force (N)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 7 15.68737500 2.24105357 0.61 0.7376
Error 8 29.52740000 3.69092500
Corrected Total 15 45.21477500

R-Square C.V. Root MSE AVG Mean

0.346952 38.41396 1.92117802 5.00125000

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 11.52602500 11.52602500 3.12 0.1152
SET 1 0.00722500 0.00722500 0.00 0.9658
FLEVEL*SET 1 1.91822500 1.91822500 0.52 0.4915
THRU 1 0.23522500 0.23522500 0.06 0.8071
FLEVEL*THRU 1 0.64802500 0.64802500 0.18 0.6862
SET*THRU 1 0.85562500 0.85562500 0.23 0.6431
FLEVEL*SET*THRU 1 0.49702500 0.49702500 0.13 0.7232

ANOVA FOR COEFF. OF RICTION MEASUREMENTS OF POLYESTER

General Linear Models Procedure

Class Level Information

Class Levels Values

FLEVEL 5 0.111 0.137 0.222 0.0745 0.1795
Number of observations in data set = 10

Dependent Variable: AVG Average Frictional force (cN)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 4 7031.47256000 1757.86814000 3.51 0.1005
Error 5 2503.76280000 500.75256000
Corrected Total 9 9535.23536000

R-Square C.V. Root MSE AVG Mean

0.737420 2.929713 22.37750120 763.81200000

ANOVA FOR SQRT OF NEPS PER GRAM OF POLYESTER WEB

(page 35, figure 24)

General Linear Models Procedure

125
 Class Level Information

Class Levels Values

FLEVEL 5 0.111 0.137 0.222 0.0745 0.1795
SET 2 59
THRU 2 90 180
Number of observations in data set = 60
Dependent Variable: SQNEP Square root of neps (per gm)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 19 9.82831387 0.51727968 3.12 0.0012
Error 40 6.62836927 0.16570923
Corrected Total 59 16.45668314
R-Square C.V. Root MSE SQNEP Mean
0.597223 3.246643 0.40707399 12.53830486

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 4 1.25572673 0.31393168 1.89 0.1303
SET 1 4.25703203 4.25703203 25.69 0.0001
FLEVEL*SET 4 0.96266490 0.24066622 1.45 0.2348
THRU 1 0.33560121 0.33560121 2.03 0.1625
FLEVEL*THRU 4 1.14276523 0.28569131 1.72 0.1637
SET*THRU 1 0.34251121 0.34251121 2.07 0.1583
FLEVEL*SET*THRU 4 1.53201256 0.38300314 2.31 0.0743

Source DF Type III SS Mean Square F Value Pr > F

FLEVEL 4 0.86783973 0.21695993 1.31 0.2830
SET 1 4.21954349 4.21954349 25.46 0.0001
FLEVEL*SET 4 0.99781027 0.24945257 1.51 0.2189
THRU 1 0.34904843 0.34904843 2.11 0.1545
FLEVEL*THRU 4 1.05783650 0.26445913 1.60 0.1942
SET*THRU 1 0.35552347 0.35552347 2.15 0.1508
FLEVEL*SET*THRU 4 1.53201256 0.38300314 2.31 0.0743

General Linear Models Procedure (page 35, figure 25)

Class Level Information

Class Levels Values

FLEVEL 2 Blended original
SET 2 59
THRU 2 90 180
Number of observations in data set = 24

Dependent Variable: SQNEP Square root of neps (per gram)

126
Source DF Sum of Squares Mean Square F Value Pr > F
Model 7 1.91942075 0.27420296 1.21 0.3535
Error 16 3.63214654 0.22700916
Corrected Total 23 5.55156729

R-Square C.V. Root MSE SQNEP Mean

0.345744 3.801510 0.47645478 12.53330486

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 0.00029908 0.00029908 0.00 0.9715
SET 1 0.53820848 0.53820848 2.37 0.1432
FLEVEL*SET 1 0.00048118 0.00048118 0.00 0.9638
THRU 1 0.68820565 0.68820565 3.03 0.1008
FLEVEL*THRU 1 0.25494091 0.25494091 1.12 0.3050
SET*THRU 1 0.00237759 0.00237759 0.01 0.9198
FLEVEL*SET*THRU 1 0.43490787 0.43490787 1.92 0.1853

127
 Appendix III : Statistical Results of Response Parameters of Needlepunching Experiments
Performed in Chapter 3

The analysis of variance (ANOVA) for the parameters tested in chapter 3, are performed using the SAS library. The
model is setup in the form of a randomized block design. The ANOVA results are as given below. For orthogonal results,
the type III sum of squares are not reported as they are the same as the type I sum of squares. The page number containing
the discussion of the corresponding table and the figure number of the corresponding graph is given in brackets for each
ANOVA result.

a) POLYPROPYLENE

ANOVA FOR PEAK PENETRATION FORCE (cN) FOR POLYPROPYLENE FIBER (page
44,figure 4)

General Linear Models Procedure

Class Level Information

Class Levels Values

NPOS 5 12345
FLEVEL 2 0.11 0.61
Number of observations in data set = 50

Dependent Variable: PPF Peak Penetration Force (cN)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 9 3806807.76010791 422978.64001199 487.19 0.0001
Error 40 34728.13535879 868.20338397
Corrected Total 49 3841535.89546670

R-Square C.V. Root MSE PPF Mean

0.990960 7.204260 29.46529117 408.99815200

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 108018.77632128 108018.77632128 124.42 0.0001
NPOS 4 3307846.31536168 826961.57884042 952.50 0.0001
NPOS*FLEVEL 4 390942.66842495 97735.66710624 112.57 0.0001

General Linear Models Procedure (page 45, figure 6)

Class Level Information

Class Levels Values

NPOS 5 12345
FLEVEL 2 Blended Original
Number of observations in data set = 50

128
Dependent Variable: PPF Peak Penetration Force (cN)

Source DF Sum of Squares Mean Square F Value Pr > F

(ii) Model 9 3522925.84339261 391436.20482140 950.56

0.0001
Error 40 16471.90059348 411.79751484
Corrected Total 49 3539397.74398610

R-Square C.V. Root MSE PPF Mean

0.995346 6.088782 20.29279465 333.28169200

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 42738.99745122 42738.99745122 103.79 0.0001
NPOS 4 3467377.94035871 866844.48508968 2105.03 0.0001
NPOS*FLEVEL 4 12808.90558269 3202.22639567 7.78 0.0001

ANOVA FOR PENETRATION ENERGY FOR POLYPROPYLENE FIBER

(page 44,figure 5)

General Linear Models Procedure

Class Level Information

Class Levels Values

NPOS 5 12345
FLEVEL 2 0.11 0.61
Number of observations in data set = 50

Dependent Variable: PE Penetration Energy (mJ)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 9 198.34619054 22.03846562 256.93 0.0001
Error 40 3.43105238 0.08577631
Corrected Total 49 201.77724292

R-Square C.V. Root MSE PE Mean

0.982996 9.717568 0.29287593 3.01388080

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 4.19035038 4.19035038 48.85 0.0001
NPOS 4 165.08185231 41.27046308 481.14 0.0001
NPOS*FLEVEL 4 29.07398785 7.26849696 84.74 0.0001

General Linear Models Procedure (page 45, figure 7)

129
 (iii) Class Level Information

Class Levels Values

NPOS 5 12345
FLEVEL 2 Blended Original

(iv) Number of observations in data set = 50

Dependent Variable: PE Penetration Energy (mJ)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 9 184.12305724 20.45811747 212.13 0.0001
Error 40 3.85765341 0.09644134
Corrected Total 49 187.98071065

R-Square C.V. Root MSE PE Mean

0.979478 12.31774 0.31055005 2.52116120

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 2.06502410 2.06502410 21.41 0.0001
NPOS 4 180.54290818 45.13572704 468.01 0.0001
NPOS*FLEVEL 4 1.51512497 0.37878124 3.93 0.0088

ANOVA FOR TENSILE STRENGHT OF POLYPROPYLENE NEEDLED FABRIC

(page 46,figure 8)

General Linear Models Procedure

Class Level Information

Class Levels Values

FLEVEL 2 0.11 0.61
DIR 2 MD XD
Number of observations in data set = 16
Dependent Variable: LOAD, Tensile Strength (cN)
Source DF Sum of Squares Mean Square F Value Pr > F
Model 3 0.00232360 0.00077453 7.76 0.0038
Error 12 0.00119799 0.00009983
Corrected Total 15 0.00352159

R-Square C.V. Root MSE LOAD Mean

0.659816 20.79338 0.00999161 0.04805188

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 0.00071516 0.00071516 7.16 0.0202

130
DIR 1 0.00146402 0.00146402 14.66 0.0024
FLEVEL*DIR 1 0.00014442 0.00014442 1.45 0.2523

General Linear Models Procedure (page 47, figure 11)

Class Level Information

Class Levels Values

FLEVEL 2 Blended Original
DIR 2 MD XD
Number of observations in data set = 16

Dependent Variable: LOAD, Tensile Strength (cN)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 3 0.00416468 0.00138823 19.96 0.0001
Error 12 0.00083462 0.00006955
Corrected Total 15 0.00499930

R-Square C.V. Root MSE LOAD Mean

0.833053 19.87638 0.00833976 0.04195813
Source DF Type I SS Mean Square F Value Pr > F
FLEVEL 1 0.00261300 0.00261300 37.57 0.0001
DIR 1 0.00137882 0.00137882 19.82 0.0008
FLEVEL*DIR 1 0.00017286 0.00017286 2.49 0.1409

ANOVA FOR TEAR STRENGTH OF POLYPROPYLENE NEEDLED FABRIC

(page 46,figure 9)

General Linear Models Procedure

Class Level Information

Class Levels Values

FLEVEL 2 0.11 0.61
DIR 2 MD XD
Number of observations in data set = 16

Dependent Variable: TEAR, Tear Strength (cN)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 3 7.07546928 2.35848976 3.46 0.0510
Error 12 8.16797770 0.68066481
Corrected Total 15 15.24344698

R-Square C.V. Root MSE TEAR Mean

0.464165 25.28531 0.82502413 3.26286000

131
Source DF Type I SS Mean Square F Value Pr > F
FLEVEL 1 2.68448563 2.68448563 3.94 0.0704
DIR 1 4.14246644 4.14246644 6.09 0.0297
FLEVEL*DIR 1 0.24851721 0.24851721 0.37 0.5569

General Linear Models Procedure (page 47,figure 12)

Class Level Information

Class Levels Values

FLEVEL 2 Blended Original
DIR 2 MD XD
Number of observations in data set = 16

Dependent Variable: TEAR, Tear Strength (cN)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 3 17.84548725 5.94849575 10.89 0.0010
Error 12 6.55347517 0.54612293
Corrected Total 15 24.39896242

R-Square C.V. Root MSE TEAR Mean

0.731404 27.10123 0.73900131 2.72681875

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 14.30810059 14.30810059 26.20 0.0003
DIR 1 2.79366153 2.79366153 5.12 0.0431
FLEVEL*DIR 1 0.74372514 0.74372514 1.36 0.2659

ANOVA FOR ENERGY TO BREAKING POINT OF POLYPROPYLENE NEEDLED FABRIC

(page 46,figure 10)

General Linear Models Procedure

Class Level Information

Class Levels Values

FLEVEL 2 0.11 0.61
DIR 2 MD XD
Number of observations in data set = 16

Dependent Variable: EBREAK, Energy to Break (J)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 3 5.14766272 1.71588757 50.45 0.0001
Error 12 0.40816773 0.03401398

132
Corrected Total 15 5.55583045

R-Square C.V. Root MSE EBREAK Mean

0.926533 17.25144 0.18442879 1.06906313

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 0.36196768 0.36196768 10.64 0.0068
DIR 1 4.64306607 4.64306607 136.50 0.0001
FLEVEL*DIR 1 0.14262896 0.14262896 4.19 0.0631

General Linear Models Procedure (page 47,figure 13)

Class Level Information

Class Levels Values

FLEVEL 2 Blended Original
DIR 2 MD XD
Number of observations in data set = 16

Dependent Variable: EBREAK, Energy to Break (J)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 3 4.98314880 1.66104960 66.90 0.0001
Error 12 0.29795618 0.02482968
Corrected Total 15 5.28110498

R-Square C.V. Root MSE EBREAK Mean

0.943581 15.52058 0.15757437 1.01526062

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 0.66723984 0.66723984 26.87 0.0002
DIR 1 4.04397018 4.04397018 162.87 0.0001
FLEVEL*DIR 1 0.27193878 0.27193878 10.95 0.0062

133
 POLYESTER

ANOVA FOR PEAK PENETRATION FORCE (cN) FOR POLYESTER FIBER

(page 48, figure 14)

General Linear Models Procedure

Class Level Information

(v) Class Levels Values

NPOS 5 12345
FLEVEL 3 0.11 0.13 0.22
Number of observations in data set = 66

Dependent Variable: PPF, Peak Penetration Force (cN)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 14 3728259.90076131 266304.27862581 611.02 0.0001
Error 51 22227.62433632 435.83577130
Corrected Total 65 3750487.52509764

R-Square C.V. Root MSE PPF Mean

0.994073 5.175731 20.87668008 403.35713182

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 2 131474.36239034 65737.18119517 150.83 0.0001
NPOS 4 3496646.16150050 874161.54037513 2005.71 0.0001
NPOS*FLEVEL 8 100139.37687047 12517.42210881 28.72 0.0001

Source DF Type III SS Mean Square F Value Pr > F

FLEVEL 2 97549.47585103 48774.73792551 111.91 0.0001
NPOS 4 3394330.87673034 848582.71918259 1947.02 0.0001
NPOS*FLEVEL 8 100139.37687047 12517.42210881 28.72 0.0001

General Linear Models Procedure (page 49, figure 16)

Class Level Information

Class Levels Values

NPOS 5 12345
FLEVEL 2 Blended Original
Number of observations in data set = 45

Dependent Variable: PPF Peak Penetration Force (cN)

Source DF Sum of Squares Mean Square F Value Pr > F

134
Model 9 2494715.32096629 277190.59121848 585.35 0.0001
Error 35 16574.17496048 473.54785601
Corrected Total 44 2511289.49592677

R-Square C.V. Root MSE PPF Mean

0.993400 5.008007 21.76115475 434.52720222

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 191.07191934 191.07191934 0.40 0.5294
NPOS 4 2492486.92575386 623121.73143846 1315.86 0.0001
NPOS*FLEVEL 4 2037.32329309 509.33082327 1.08 0.3834

Source DF Type III SS Mean Square F Value Pr > F

FLEVEL 1 191.07191934 191.07191934 0.40 0.5294
NPOS 4 2455578.17366429 613894.54341607 1296.37 0.0001
NPOS*FLEVEL 4 2037.32329309 509.33082327 1.08 0.3834

ANOVA FOR PENETRATION ENERGY FOR POLYESTER FIBER

(page 48,figure 15)

General Linear Models Procedure

Class Level Information

Class Levels Values

NPOS 5 12345
FLEVEL 3 0.11 0.13 0.22

(vi) Number of observations in data set = 66

Dependent Variable: PE Penetration Energy (mJ)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 14 247.31013931 17.66500995 174.19 0.0001
Error 51 5.17189582 0.10140972
Corrected Total 65 252.48203513

R-Square C.V. Root MSE PE Mean

0.979516 9.809645 0.31844893 3.24628394

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 2 9.81210389 4.90605194 48.38 0.0001
NPOS 4 230.75778403 57.68944601 568.87 0.0001
NPOS*FLEVEL 8 6.74025139 0.84253142 8.31 0.0001

135
 (vii) Source DF Type III SS Mean Square F Value Pr
>F
FLEVEL 2 7.46458166 3.73229083 36.80 0.0001
NPOS 4 224.79922325 56.19980581 554.19 0.0001
NPOS*FLEVEL 8 6.74025139 0.84253142 8.31 0.0001

General Linear Models Procedure (page 49, figure 17)

Class Level Information

Class Levels Values

NPOS 5 12345
FLEVEL 2 Blended Original
Number of observations in data set = 45

Dependent Variable: PE Penetration Energy (mJ)

(viii) Source DF Sum of Squares Mean Square F Value

Pr > F
Model 9 160.04194099 17.78243789 117.48 0.0001
Error 35 5.29761046 0.15136030
Corrected Total 44 165.33955146

R-Square C.V. Root MSE PE Mean

0.967959 11.35195 0.38905051 3.42716889

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 0.29689785 0.29689785 1.96 0.1702
NPOS 4 158.91691033 39.72922758 262.48 0.0001
NPOS*FLEVEL 4 0.82813282 0.20703320 1.37 0.2650

Source DF Type III SS Mean Square F Value Pr > F

FLEVEL 1 0.29689785 0.29689785 1.96 0.1702
NPOS 4 155.74128785 38.93532196 257.24 0.0001
NPOS*FLEVEL 4 0.82813282 0.20703320 1.37 0.2650

ANOVA FOR TENSILE STRENGHT OF POLYESTER NEEDLED FABRIC

(page 50,figure 18)

General Linear Models Procedure

Class Level Information

136
 Class Levels Values
FLEVEL 3 0.11 0.22 0.137
DIR 2 MD XD
Number of observations in data set = 24

Dependent Variable: LOAD, Tensile Strength (cN)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 5 0.30294914 0.06058983 412.50 0.0001
Error 18 0.00264390 0.00014688
Corrected Total 23 0.30559305

R-Square C.V. Root MSE LOAD Mean

0.991348 2.339341 0.01211954 0.51807500

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 2 0.08723890 0.04361945 296.97 0.0001
DIR 1 0.19921748 0.19921748 1356.30 0.0001
FLEVEL*DIR 2 0.01649277 0.00824638 56.14 0.0001

General Linear Models Procedure (page 51,figure 21)

Class Level Information

Class Levels Values

FLEVEL 2 Blended Original
DIR 2 MD XD
Number of observations in data set = 16

Dependent Variable: LOAD, Tensile Strength (cN)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 3 0.04472528 0.01490843 30.50 0.0001
Error 12 0.00586610 0.00048884
Corrected Total 15 0.05059138

R-Square C.V. Root MSE LOAD Mean

0.884049 5.246743 0.02210977 0.42140000

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 1 0.00607620 0.00607620 12.43 0.0042
DIR 1 0.03798601 0.03798601 77.71 0.0001
FLEVEL*DIR 1 0.00066306 0.00066306 1.36 0.2668

ANOVA FOR TEAR STRENGHT OF POLYESTER NEEDLED FABRIC

(page 50,figure 19)

137
 General Linear Models Procedure
Class Level Information

Class Levels Values

FLEVEL 3 0.11 0.22 0.137
DIR 2 MD XD
Number of observations in data set = 24
Dependent Variable: TEAR, Tear Strength (cN)
Source DF Sum of Squares Mean Square F Value Pr > F
Model 5 550.68449906 110.13689981 24.49 0.0001
Error 18 80.94271958 4.49681775
Corrected Total 23 631.62721864

R-Square C.V. Root MSE TEAR Mean

0.871850 6.191031 2.12057015 34.25229500

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 2 185.87350683 92.93675342 20.67 0.0001
DIR 1 281.79344128 281.79344128 62.67 0.0001
FLEVEL*DIR 2 83.01755094 41.50877547 9.23 0.0017

General Linear Models Procedure (page 51,figure 22)

Class Level Information

Class Levels Values

FLEVEL 2 Blended Original
DIR 2 MD XD
Number of observations in data set = 16

Dependent Variable: TEAR, Tear Strength (cN)

Source DF Sum of Squares Mean Square F Value Pr > F
Model 3 436.05213176 145.35071059 31.67 0.0001
Error 12 55.06934482 4.58911207
Corrected Total 15 491.12147658

R-Square C.V. Root MSE TEAR Mean

0.887870 6.970894 2.14222129 30.73094125
Source DF Type I SS Mean Square F Value Pr > F
FLEVEL 1 0.67032700 0.67032700 0.15 0.7090
DIR 1 420.60677569 420.60677569 91.65 0.0001
FLEVEL*DIR 1 14.77502907 14.77502907 3.22 0.0980

138
 ANOVA FOR ENERGY TO BREAKING POINT OF POLYESTER NEEDLED FABRIC (page
50,figure 20)

General Linear Models Procedure

Class Level Information

Class Levels Values

FLEVEL 3 0.11 0.22 0.137
DIR 2 MD XD
Number of observations in data set = 24

Dependent Variable: EBREAK, Energy to Break (J)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 5 262.21654629 52.44330926 57.91 0.0001
Error 18 16.30013924 0.90556329
Corrected Total 23 278.51668553

R-Square C.V. Root MSE EBREAK Mean

0.941475 12.12148 0.95161089 7.85061750

Source DF Type I SS Mean Square F Value Pr > F

FLEVEL 2 75.93100071 37.96550036 41.92 0.0001
DIR 1 166.53928705 166.53928705 183.91 0.0001
FLEVEL*DIR 2 19.74625853 9.87312927 10.90 0.0008

General Linear Models Procedure (page 51,figure 23)

Class Level Information

Class Levels Values

FLEVEL 2 Blended Original
DIR 2 MD XD
Number of observations in data set = 16

Dependent Variable: EBREAK, Energy to Break (J)

Source DF Sum of Squares Mean Square F Value Pr > F

Model 3 28.49846185 9.49948728 31.74 0.0001
Error 12 3.59191999 0.29932667
Corrected Total 15 32.09038185

R-Square C.V. Root MSE EBREAK Mean

0.888069 9.611901 0.54710755 5.69198062

Source DF Type I SS Mean Square F Value Pr > F

139
FLEVEL 1 9.46360631 9.46360631 31.62 0.0001
DIR 1 15.81384464 15.81384464 52.83 0.0001
FLEVEL*DIR 1 3.22101090 3.22101090 10.76 0.0066

140
 a) Appendix IV : Method Used for Signal Analysis of Needling Forces

2) Here, the method used for signal analysis of needling forces is explained with an example. The
needling force data shown here were collected while needlepunching polypropylene fibers with 0.61%
finish level. The raw data from the NCRC dynamic needling force measurement device is in the form
of voltage readings from the force transducers attached to the needles. These voltage readings are
converted to force measurements by means of a calibration equation. To arrive at the calibration
equation, first the strain gauges are loaded with a range of known weights and the corresponding
voltage readings are noted. These voltage readings are then regressed on the known loads and the
resulting regression equation are used for converting the voltage readings into force measurements.
The result of the regression analysis is given in table 1.

(ix) Table 1: Regression Analysis for calibration of strain gauges

Dependent Variable: FORCE

Sum of Mean
Source DF Squares Square F Value Prob>F
Model 1 5598377.3816 5598377.3816 4045657.333 0.0001
Error 5 6.91900 1.3838
C Total 6 5598384.3006
Root MSE 1.17635 R-square 1.0000 Dep Mean 0.00000
C.V. . Adj R-sq 1.0000
Parameter Estimates
Parameter Standard T for H0:

(x) Variable DF Estimate Error Parameter=0 Prob >

|T|
INTERCEP 1 0 0.44461849 0.000 1.0000
SENSOR 1 447.148800 0.22230924 2011.382 0.0001

i) F = 447.1488V where, F = Force, cN and V = Output Voltage,

Volts

A sample plot of the force measurements against time for a single needle (location 4) is given in
figure 3 of chapter 3. A moving median filter is used to smoothen the curve of the raw force
measurements and the resultant curve is shown in figure 1.

A fast Fourier transformation is performed on the data of figure 3 of chapter 3, to derive its
frequency domain and find out the significant Fourier frequencies. The result of the fast Fourier
transformation is given in figure 2. The Fourier frequencies show a significant effect till its fifth
harmonic (table 2). Now the curve is fitted with significant harmonics of sine and cosine term as
given in equation 1 in chapter 3. The model used for the needle is given in table 2. The resultant
coefficient of the sine and cosine term for each needle and its R2 value is given in table 2.

141
Figure 1: Smoothened curve of needle force measurements for location 4.

Figure 2: Frequency domain of force measurements

142
 Table 2: Time series model used for fitting curve
Model: MODEL1
Dependent Variable: F4
Sum of Mean
Source DF Squares Square F Value Prob>F
Model 10 338649401.44 33864940.144 5731.144 0.0001
Error 4985 29456028.07 5908.932
C Total 4995 368105429.52
Root MSE 76.86958 R-square 0.9200
Dep Mean -206.42943 Adj R-sq 0.9198
C.V. -37.23770

Parameter Standard T for H0:

Variable DF Estimate Error Parameter=0 Prob > |T|
INTERCEP 1 -207.035244 1.08755279 -190.368 0.0001
C1 1 3.178926 1.53775087 2.067 0.0388
S1 1 144.669606 1.53831242 94.044 0.0001
C2 1 -222.732235 1.53779530 -144.839 0.0001
S2 1 105.368328 1.53826805 68.498 0.0001
C3 1 -197.279809 1.53786104 -128.282 0.0001
S3 1 60.254336 1.53820029 39.172 0.0001
C4 1 -101.202814 1.53794143 -65.804 0.0001
S4 1 -1.439892 1.53811979 -0.936 0.3492
C5 1 -22.818667 1.53801368 -14.836 0.0001
S5 1 -20.902124 1.53802610 -13.590 0.0001

Table 3: Force equations of needles and their R2 values

Needle
R2
Locatio Force Equation
value
n
F = -20.273468 +4.9533238COS(2πft) +10.180516SIN(2πft)
–29.764971COS(2π2ft) – 3.935241SIN(4π2ft)
1 –11.3860909COS(2π3ft) – 5.567867SIN(2π3ft) 0.829
– 11.054521COS(2π4ft) –5.58895SIN(2π4ft)
–4.732702COS(2π5ft) –5.062679SIN(2π5ft)
F = -77.388668 +27.823877COS(2πft) +22.093594SIN(2πft)
+86.959382COS(2π2ft) +11.06084SIN(4π2ft)
2 +27.005616COS(2π3ft) –62.990897SIN(2π3ft) 0.749
– 27.79451COS(2π4ft) –30.830269SIN(2π4ft)
–24.074971COS(2π5ft) +4.589339SIN(2π5ft)
3 F = -18.546904 +4.9533238COS(2πft) +17.666115SIN(2πft) 0.820

143
 –30.99704COS(2π2ft) – 3.177286SIN(4π2ft)
–24.675777COS(2π3ft) – 7.571287SIN(2π3ft)
– 13.788766COS(2π4ft) –7.46031SIN(2π4ft)
–6.273196COS(2π5ft) –6.685679SIN(2π5ft)
F = -207.035244 +3.178926COS(2πft) +144.669606SIN(2πft)
–222.732235COS(2π2ft) +105.368328SIN(4π2ft)
4 –197.279809COS(2π3ft) +60.254336SIN(2π3ft) 0.920
– 101.202814COS(2π4ft) –1.439892SIN(2π4ft)
–22.818667COS(2π5ft) –20.902124SIN(2π5ft)
F = -2.112506 +80.806558COS(2πft) +71.32329SIN(2πft)
–118.342193COS(2π2ft) –14.880656SIN(4π2ft)
5 –90.912827COS(2π3ft) –10.155625IN(2π3ft) 0.942
–46.400127OS(2π4ft) –24.792437SIN(2π4ft)
–15.511558COS(2π5ft) –19.338935SIN(2π5ft)
F => Force (cN), f => Fourier Frequency (0.01692), t => time (sec)

Using the above equation, the inertia forces (figure 3(b)) is removed from the actual forces
(figure 3(a)) and the resultant forces are given in figure 3(d). The Matlab script used for
separating the inertia force from the actual forces is given in table 6. From this, the peak
penetration force is derived by recording the lowest point in a cycle and the penetration energy is
derived by calculating the area under the curve during penetration. The area under the curve is
calculated by means of the trapz() function under Matlab. The Trapz() function calculates the
area by dividing the curve into very small trapezoids and adding up the total area of these
trapezoids.

The variance for the predictions using the equation in table 3 is given by the equation 2 and 3 in
chapter 3. Another method to arrive at the variance is to increase the number of replication for
each needle and calculate the peak penetration force for each replication. The results of the peak
penetration force for polypropylene with finish level 0.61% are given in table 4. It can be seen
that the peak penetration force for the replicates of each needle is an estimate of the actual peak
penetration force and calculating the variance between them should also provide an estimate of
the variance of the actual penetration force. The peak penetration energy for polypropylene with
finish level 0.61% are given in table 5.

144
 (d) Actual needling force
(a) Total needling force

(b) Inertia force

(c) Subtract Inertia force from

Total Needling force

90
 Figure 3: Derivation of Peak penetration force and penetration energy

Table 4: Peak penetration forces for polypropylene with 0.61% finish level
Needle Peak Penetration Forces (cN)
Locatio Rep 1 Rep 2 Rep 3 Rep 4 Rep 5 Mean Varianc
n e
1 177.973 238.266 170.574 175.467 189.215 177.973 238.266
2 336.918 323.918 303.861 346.431 331.647 336.918 323.918
3 179.536 228.323 161.667 211.404 193.303 179.536 228.323
4 683.200 694.215 734.733 679.299 712.514 683.200 694.215
5 830.215 894.371 760.902 917.505 912.254 830.215 894.371

Table 5: Penetration energy for polypropylene with 0.61% finish level

Needle Penetration Energy (J x 10-3)
Locatio Rep 1 Rep 2 Rep 3 Rep 4 Rep 5 Mean Variance
n
1 1.337 1.839 1.229 1.365 1.411 1.4362 0.055181
2 2.437 2.351 2.100 2.400 2.399 2.3374 0.018544
3 1.303 1.789 1.103 1.602 1.451 1.4496 0.070011
4 4.385 4.988 4.725 4.348 5.000 4.6892 0.099029
5 6.214 6.890 5.742 7.156 7.021 6.6046 0.363883

1
Table 6: Matlab Script for removing inertia forces from actual forces and calculating the peak
penetration force and penetration energy of each needle

qerf = 0.01692; % define the freq / 2phi

period = round(1/qerf); % Derive the period
Hper = round(period/2); % half period
Qper = round(period/4); % quarter period
intercept = -207.035244; % define Intercept
c1 = 3.178926; % cos and sine coefficients
s1 = 144.669606;
c2 = -222.732235;
s2 = 105.368328;
c3 = -197.279809;
s3 = 60.254336;
c4 = -101.202814;
s4 = -1.439892;
c5 = -22.818667;
s5 = -20.902124;

x = linspace(0,3*period,6*period); % x- axis for 3 time periods or cycles and

t = linspace(0,3*period,6*period); % Doubling the sample rate
y = intercept + c1*cos(2*pi*qerf*t) + s1*sin(2*pi*qerf*t)+ c2*cos(2*pi*2*qerf*t) +
s2*sin(2*pi*2*qerf*t)+ c3*cos(2*pi*3*qerf*t) + s3*sin(2*pi*3*qerf*t)+ c4*cos(2*pi*4*qerf*t) +
s4*sin(2*pi*4*qerf*t)+ c5*cos(2*pi*5*qerf*t) + s5*sin(2*pi*5*qerf*t);

% Now since the whole force equation is characterized. We have to isolate the
% inertial forces which are in a sense embedded in the force equations.
% For this one has to find where the inertia forces end and where the mixture
% of the two forces begin in a cycle.
% The first assumption here is that the inertial forces during upward motion of
% the needle is the same as that during the downward movement. This is dependent
% on the needle location and known to vary for each run.
% Second assumption is that the inertial force in the first half cycle is the
% same as in the second have but with the direction reversed.
% now the center of the equation does not coincide with the x-axis so one cannot
% use the x-axis to identify the amplitude of the inertia part.
Lowest1 = 1; % find lowest point of y in a cycle
for n = 1:(2*period)
if y(n) < y(Lowest1)
Lowest1 = n
end
end
Highest1 = Lowest1+1; % from there find 1st highest point
for n = Lowest1:(Lowest1+(period+10)) % here period is actually half cycle
if y(n) > y(Highest1)

2
 Highest1 = n
end
end
Lowest2 = Highest1+1; % from there find 2nd lowest point
for n = Highest1:(Highest1+(round(0.85*period))) % in the next half cycle
if y(n) < y(Lowest2)
Lowest2 = n
end
end

if (Lowest2 - Lowest1) > period-10

DF = round((Lowest2 - Lowest1) / 2);
Lowest2 = Highest1+1; % from there find 2nd lowest point
for n = Highest1:(Highest1+(DF)) % in the next half cycle
if y(n) < y(Lowest2)
Lowest2 = n
end
end
end
Highest2 = Lowest2+1; % from here find 2nd highest point
for n = Lowest2:(Lowest1+(2*period+10)) % in next half cycle
if y(n) > y(Highest2)
Highest2 = n
end
end
Lowest3 = Highest2+1; % from here find the 3rd lowest point
for n = Highest2:(Highest2+(period+10)) % in the next half cycle
if y(n) < y(Lowest3)
Lowest3 = n
end
end

Spoint = round((Lowest2 + (Lowest2 - Highest1) - (Highest2 - Lowest2))/2);

% Spoint refers to the starting point of the cycle
% round((Lowest2 - Highest1))/2;
%round((Lowest2 + (Lowest2 - Highest1) - (Highest2 - Lowest2))/2);

t = linspace(Spoint,(Spoint+3*period),6*period);
y = intercept + c1*cos(2*pi*qerf*t) + s1*sin(2*pi*qerf*t)+ c2*cos(2*pi*2*qerf*t) +
s2*sin(2*pi*2*qerf*t)+ c3*cos(2*pi*3*qerf*t) + s3*sin(2*pi*3*qerf*t)+ c4*cos(2*pi*4*qerf*t) +
s4*sin(2*pi*4*qerf*t)+ c5*cos(2*pi*5*qerf*t) + s5*sin(2*pi*5*qerf*t);
t = linspace(Spoint+Hper,(Spoint+Hper+3*period),6*period);
z = - (intercept + c1*cos(2*pi*qerf*t) + s1*sin(2*pi*qerf*t)+ c2*cos(2*pi*2*qerf*t) +
s2*sin(2*pi*2*qerf*t)+ c3*cos(2*pi*3*qerf*t) + s3*sin(2*pi*3*qerf*t)+ c4*cos(2*pi*4*qerf*t) +
s4*sin(2*pi*4*qerf*t)+ c5*cos(2*pi*5*qerf*t) + s5*sin(2*pi*5*qerf*t));

3
% Z is an inverse plot of y
plot(y); % define y and plot it (figure 3(a) )
pause;
Adjust = ( y(Lowest1) + y(Highest1))/2;

% adjust the plot of Z to account for the shift in mean of the cycle due to reversing of Z % from Y
u = Adjust + z;

% u is the adjusted plot

lag = 48;

% Calculated for each needle by trial and error till inertia zone is identified
% this lag usually remains same and fluctuated with a slight change in frequency of cycle

for n = 1:lag
w(n) = y(n);
end
for n = lag+1:(period+lag)
w(n) = u(n);
end
for n = (period+1+lag):(2*period+lag)
w(n) = y(n);
end
for n = (2*period+1+lag):(3*period+lag)
w(n) = u(n);
end
for n = (3*period+1+lag):(4*period+lag)
w(n) = y(n);
end
for n = (4*period+1+lag):(5*period+lag)
w(n) = u(n);
end
for n = (5*period+1+lag):(6*period)
w(n) = y(n);
end
Median = (max(w)+ min(w))/2
M = Median+(0*x);
plot(x,y,x,w,x,M); % this plot is as shown in figure 3 ( c )
pause;
u = y - w;
plot(u); % Actual needling force as shown in figure 3 (d)
pause;
Zcross = 0; % Determine the number of cross over points of the graph over the x-axis

4
for n = 1:150
p(n) = u(n);
end
ZC = 1;
for n = 2:149
if (p(n) ~= 0) & (((p(n-1) <= 0) & (p(n+1) >= 0)) | ((p(n-1) >= 0) & (p(n+1) <= 0)))
Zcross(ZC) = n;
ZC = ZC + 1;
end
end
Zcross % Display cross over points
min(p) % Display peak penetration force
max(p) % Display peak stripping force

% To calculate penetration energy or stripping energy

% from Zcross and plot of resultant force equation , determine the penetration zone and % striping
zone. Now for this zone alone find the area under the curve

h = 0;
a = 1;
b = 45;
for n =a:b
h(n-a+1) = p(n);
end
trapz(h)/2400; % this gives the area under the curve and penetration energy in Jx10-3.