Preferred citation: B. Norman. Overview of the physics of forming.

In Fundamentals of Papermaking,
Trans. of the IXth Fund. Res. Symp. Cambridge, 1989, (C.F. Baker & V. Punton, eds), pp 73–149, FRC,
Manchester, 2018. DOI: 10.15376/frc.1989.3.73.

OVERVIEW OF THE PHYSICS OF

FORMING
Professor Bo Norman
Paper Technology Department, Royal Institute of Technology and
Swedish Pulp Ff Paper Research Institute, Stockholm, Sweden

Abstract

In this overview forming covers all processes from the dilution of thick stock
into a mix, using recirculated white water in the short circulation , to the dewa-
tering in the wire section.
Grammage nonuniformity in the paper web is to a predominant degree
generated by the forming process, and especially the small scale variations
summarized in the termformation . The term massformation is recommended
when only grammage is considered and not the optical impressions thereof. The
forming process also generates the main part of the large scale variations, that is
the MD-, CD- and residual grammage variance.
Mass formation has traditionally been evaluated using beta radiography,
combined with microdensitometry or image analysis. A new technique involving
thedirect recording of electron beam transmission is under development, with
promises of faster processing, perhaps even on-line, and high geometrical reso-
lution. Characterization techniques based on the co-occurrence matrix, applica-
ble to image analysis, can be a useful complement to the traditional power spec-
tra techniques .
74

It has recently been conclusively demonstrated that in flowing fibre sus-

pensions, flocs are kept together by the bending forces of interlocked fibres . To
study the dynamic behaviour of flowing fibre suspensions, modern video tech-
niques, high speed movie pictures and image analysis are applied.
To improve grammage uniformity, the mix should be fed to the headbox
directly after the dilution of thick stock with white water No processes like
screening or cleaning, from which uncontrolled reject fibre flows are drawn
should be allowed in the short circulation . Further the consistency of each mate-
rial component in the recirculated white water should be controlled, if the con-
tent of the different components in the sheet produced is to be held constant.
First pass retention is shown to be a badly defined retention value, and
should only be used in comparisons for one paper machine when no changes of
the material content in the long circulation occurs .
In the headbox, the tapered manifold is the commonly used means for
achieving the coarse distribution across the machine width. From the cross dis-
tributor, a tube package can lead either to a stilling chamber, or directly into the
outlet nozzle . The nozzle has to be fed using maximum open area to avoid flow
instability, and the nozzle contraction must be large enough to reduce velocity
streaks and relative turbulence to an acceptable level. Mathematical models are
now being applied to the calculation of water flow patterns in headboxes .
Local slice lip adjustments, especially on headboxes with outlet nozzles of
low convergence angle, can cause considerable sideways flow on a fourdrinier
wire, and this will have a large effect on the CD grammage variations as well as
on the final sheet anisotropy profiles.
There are two basic headbox designs for stratified forming. In one of
these, thin, pointed vanes separate the different furnishes. In the other, thicker
separation walls generate air wedges, which may separate the furnishes up to the
actual starting point for dewatering. In the first case, layer mixing can start al-
ready at the headbox, while in the latter four new surfaces between air and mix
jets are created, all four being potential sources of disturbance generation.
High consistency headboxes have been developed, with which paper is
formed according to an extrusion process . To reduce mix flocculation, various
channel shapes are used inside the headboxes .
Sheet build-up generally takes place according to afiltration process,
which has an inherent self healing effect. Therefore, the mass formation of a
laboratory sheet is better than that of a random sheet. For a machine-made sheet,
the comparatively high mix consistency causes floc generation, which may re-
sult in a worse large scale mass formation than that of the random sheet.
When evaluating the mechanical and optical characteristics of a machine-
made paper sample, its properties relative to those of a laboratory sheet from the
same furnish may be expressed as the Forming Efficiency.
The Kozeny-Carman equation describing flow through porous beds can
not be used to predict filtration dewatering rates during web forming . This is be-
cause of the gradual compression of the web by the dewatering forces, the clos-
 75

ing of some pores, turbulent flow situations etc. Dewatering capacities must so
far be predicted using empirical equations, and parameters evaluated on the basis
of dewatering experiments .
The development of forming wires has led to multi-layer designs where
both the paper side and the wear side can be optimized simultaneously .
Pressure pulses in hydraulic headboxes are detrimental to fourdrinier de-
watering, since attenuation due to standing wave generation on the wire can
create large MD grammage variations.
In fourdrinier dewatering, several new dewatering elements have been de-
veloped, allowing a better control of the activity in the mix on the wire, and
thus also of the mass formation of the web formed.
In conventional twin-wire forming, the dewatering pressure is generated
by wire tension according to one of two basic principles: roll dewatering with
constant or blade dewatering with pulsating dewatering pressure . A combination
of these two principles may result in an improved combination of mass forma-
tion and retention . Recently a new method for the generation of dewatering
pressure has been demonstrated, in which the pressure along the forming zone
can be controlled freely, since it is generated by application of local forces and
thus not by wire tension .
Multi-ply products manufactured through simultaneous forming are now
used for low grammage products. The problem is to achieve acceptable layer
purity as well as layer mass formation . Controlled pressure pulse dewatering
could provide the means to reach optimum dewatering conditions .
The influence of forming conditons on product properties is a vast area
within which two subjects are discussed : the interrelationship between mass
formation and paper strength and finally fibre orientation anisotropy.
Table of contents
1 . Introduction 5
1 .1 . Background 5
2. Mass formation 8
2.1 . Formation measurement 9
2.2. Mass formation measurement 9
2.3. Characterization of mass formation. 10
3. Fibre flocculation 13
4. Mix preparation. 16
4.1 . Thick stock dilution 16
4.2. Screening, cleaning and deaeration 17
4.3. Material recirculation 18
5. Mix distribution 21
5.1 . Cross machine distribution 21
5.2. Pressure pulse damping 22
5.3. Headbox nozzle 22
5,5, Stratified headboxes 28
5.5. High consistency headboxes 29
6. Fibre deposition 31
6.1. Laboratory sheets 31
6.2. Random sheets 32
6.3. Filtration dewatering. 35
6.4. Thickening dewatering. 37
7 . Mix dewatering 39
7.1 . Forming wires 40
7.2. Fourdrinier dewatering 42
7.3. Twin-wire dewatering 47
7.4. High consistency forming 56
7.5. Multi-ply forming 56
8. Product properties 60
8.1 . Mass formation and mechanical properties 60
8.2. Fibre orientation anisotropy 62
9. Acknowledgements 65
10. References 66
78

1. Introduction
In this overview, forming is defined in a wide sense as all processes involved
in transforming the thick stock flow into a wet web . Conventional forming in-
cludes three stages as in Table 1.

Table 1 : Theforming process

The actual processes included in the forming of a paper web are extremely
complex, and many random (or with a more popular physics term: chaotic)
events are involved. The interaction between all processes involved, both in the
"forward" and in the "feed-back" sense, further adds to the complexity.
No detailed theoretical treatment of the complete sheet build-up process
has yet even been seriously attempted, and this is in sharp contrast to for in-
stance the wet pressing process, where extensive models have recently been set
up at several universities and research establishments . In some cases however,
basic models have been developed for individual parts of the forming process,
with which qualitative predictions can be made.

1.1 . Background

Peter Wrist's excellent paper Dynamics ofsheet formation on the Four-

drinier machine at the 1961 Symposium W deals with the forming process in
the light of basic knowledge . Among many other things, Wrist discussed the
basics of air cushion headboxes with perforated rolls , the dewatering mecha-
nisms of table rolls and foils as well as quantitative descriptions of dewatering
 79

capacity .
Joe Parker's now classic book from 1972, The Sheet Forming Process
(2), gives a good picture of basic forming knowledge around 1970. 60% of
Parker's 140 literature references were published after the Wrist paper, which
indicates the amount of fundamental forming work undertaken during the
1960's. Parker treated fibre networks and their reaction to turbulence energy,
dewatering resistance, material distribution in the plane as well as in the z-direc-
tion of the sheet and in a final chapter "Practical applications", headboxes and
forming methods were discussed. "Two-wire formers" were mentioned, but it
should be remembered that Parker's book was published just before twin-wire
forming made its large industrial breakthrough. Within the TAPPI Fluid
Mechanics committee, work is under way to update Parker's original book.
Ben Radvan's chapter Forming , in the "Wiggins Teape Epos"
Papermaking Science () covers mainly the same areas as Wrist and Parker did,
and adds basic work published from 1970 to 1977.
A serious reader, not well aquainted with the background described in
these contributions, should make a break here, study the three references , and
then continue with this overview .
During the last decade, the demands on machine speed, production effi-
ciency, product quality, quality evenness and environmental control have all in-
creased at an accelerating rate. It follows that, today more than ever before, not
only the basic sheet build-up process in the wire section has to be considered
when a forming section is designed.
A typical layout of the forming process is shown in Fig . 1 .

Fig, 1: Forming, from thick stock addition to a wet web .

80

As indicated in fig 1, screening and cleaning which basically belong to stock

preparation are normally placed within the forming domain . This is however an
unsatisfactory solution since it leads to a large number of internal circulation
loops with inherent concentration variations and variable fibre reject rates .
The increasing requirement for control of recirculated material through the
white water system will focus interest on the retention level . This in turn
stresses the use of process chemicals, a subject that is dealt with in a separate
review by Lindstrom (4) 1 .
Dry forming has attracted increased interest for special products, but space
does not allow its discussion in this overview .
In the following, the physical forming areas listed below will be dis-
cussed :

Sheet quality
Fibre suspensions
Mix preparation
Mix distribution
Fibre deposition
Mix dewatering
Product properties

Since these seven areas each have a large enough background to motivate an in-
dividual review, it follows that it will be impossible in this overview to go in
any detail into all the vast literature covering the whole field. The choice of lit-
erature references is therefore highly personal with some older, original refer-
ences and some more recent ones . All those which space and time did not allow
to be included this time may be just as significant.
Finally, it should be pointed out that to manufacture efficiently an end
product with given properties, the forming process cannot be studied in isola-
tion . There has to be an overall optimization including the preceding raw mate-
rial preparation as well as the downstream pressing, drying and surface treat-
ment operations .

1 Setting an all time symposium record with 103 pages and 260 references!
 81

2. Mass formation
The forming process, as defined in table 1, has a large impact on resulting sheet
properties and especially on their variability. All grammage variations, apart
from long term variations in the machine direction, are exclusively generated
during forming.
Of increasing importance for the sheet properties is also the distribution of
material and fibre orientation at different levels in the z-direction of the sheet,
which is to a high degree determined during forming . The final development of
sheet properties can however be manipulated during the pressing, drying and
surface treatment operations . .
Mix preparation and mix jet generation are the main sources of grammage
variations. Large scale grammage variance is usually subdivided into MD vari-
ance, CD variance and Residual variance and can be evaluated on-line using
beta ray absorption . However, on-line equipment today cannot record gram-
mage variations smaller in geometrical size than about 20 mm.
Small scale grammage variations, often referred to as formation, are
mainly caused by the flocculation tendency of fibre suspensions, and can be
strongly affected during forming. Formation is a very general term according to
the ISO definition : "The manner in which the fibres are distributed, disposed
and intermixed to constitute the paper" . This is a very wide definition of the
complete forming process, including the resulting sheet structure.
Wahren (5) tried to connect the term formation with small scale grammage
variations and look-through with the appearance of sheet unevenness, viewed
in transmitted light . This has not however been generally accepted, since today
"formation" is used to denote both "look-through" and "formation" as well as
"forming" . A new term is therefore necessary to specify strictly small scale
grammage variations.
Corte (6 used the term Distribution of Mass Density, DMD, which has
never been generally accepted, probably because of its being too complicated .
In this overview, the term mass formation2 will be used to denote small-
scale grammage variations. Formation can then continue to be an unspecified
characteristic, denoting for example the optical unevenness of paper . There is
however no need to use it also to denote forming, as Peter Wrist already did.

2Introduced by Douglas Wahren, STORA .

82

2.1. Formation measurement

The classical formation meter is the QNSM-meter (Z) from the 1960's, while to-
day the NUI meter ($) in North America, and the M-formation meter (9) are
usually used to quantify optical formation . All three meters record the variations
in light transmission through a paper sheet. They are valuable for the compari-
son of similar products from one paper machine, but they are not useful for
further quantitative evaluation.
Recently, equipment using image analysis techniques for recording variations in
light transmission has been developed by Papworth (IQ).
Light transmission measurements can never be used to evaluate mass for-
mation of sheets made of components with different optical properties, such as
filled sheets, or of calendered sheets where local variations in light scattering
coefficient make the correlation between light transmission and local grammage
ambiguous (11,12).

2.2. Mass formation measurement

Beta ray absorption is commonly accepted for on-line measurements of large

scale grammage variations. Corte (6) used direct measurement of beta ray ab-
sorption, with 1 mm geometrical resolution to evaluate mass distribution .
In beta radiography , the complete picture of beta ray transmission through
a test sheet is recorded on an X-ray film. As beta ray source, C-14 is useful,
although the particle energy limits maximum local grammage to around 120
g/m2 with reasonable exposure times (13). The maximum geometrical resolution
using diffuse beta radiation is of the same order as the sheet thickness .
Cresson has studied the details of exposing and developing beta radio-
graphs (14).
Local grammage variations can be evaluated from an exposed x-ray film
by optical scanning using a micro-densitometer (L5), see Fig . 2, or an image
analyser (L4,16 .
With soft x-rays , a more parallel beam of radiation can be generated. This
permits an improved geometrical resolution, allowing individual fibres to be re-
produced (17). Further advantages are the short exposure time and the unlimited
grammage range . A problem is the difficulty of obtaining a low energy x-ray
beam of acceptable cross sectional size, without too disturbing gradients in ra-
diation intensity .
A recent development which looks extremely promising is to record local
grammage via the absorption of electron beams, a method under development
by Luner and co-workers (18). This could be the ultimate way of evaluating
mass formation, with possibilities even of on-line measurement .
 83

Fig . 2: Local grammage variations along a standard newsprint

sample. Beta radiograph scanned in a microdensitometer
using 0.1 mm resolution (M.

2.3. Characterization of mass formation.

The simplest way to characterize local variations is by the standard deviation 6.

It is suggested that the dimensionless coefficient of variation, i.e. the standard
deviation divided by the mean value, be called the massformation number and
denoted F. In turbulence research, the coefficient of variation of local flow ve-
locity is called the degree of turbulence .
It should be pointed out that the geometrical resolution of the measuring
equipment directly determines the amount of variations recorded.
It should also be pointed out that mean grammage has a large effect on the
mass formation number. For a random fibre distribution there is an inverse rela-
tionship between the mass formation number and the square root of the mean
grammage. For a comparison between sheets of different grammage, the follow-
ing equation can be used for normalisation3

(l)
Fo =F w/wo

where the mass formation number F is measured at grammage w and FO is the

corresponding mass formation number at the reference grammage wo (19}.

3-The inclusion of mathematical expressions in this publication is intended only to clarify the
contextual descriptions of the processes. They are not intended as analytical supplements. In
keeping with this function, units of the terms are unnecessary and have been omitted" (2).
84

2 .3.1 . The Power Spectrum

For a more complete characterization of grammage variations using beta radiog-

raphy, a two dimensional description is useful. Besides a variation number
some measure of the geometrical size ofthe variations is needed.
Already during the 1930's, the autocorrelation and its courier transform,
the frequency power spectrum, were introduced (2Q,) . The frequency power
spectrum describes how the variance, a2, is distributed in different frequency
ranges. The power spectrum is especially useful to characterize variables with
normal amplitude distribution, but it does not quantify the occurence of gradi-
ents.
To characterize mass formation in paper sheets as well as turbulence and
flocculation in flowing suspensions, it has proved useful to transform the fre-
quency specrum into a wavelength spectrum . The wavelength 1 is calculated
from the frequency n and the radiograph scanning speed (or flow velocity) u
using the simple transformation

1= u/n (2)

and the frequency spectral density has to be transformed into a wavelength

spectral density accordingly (21).
The advantage of the wavelength spectrum over the frequency spectrum is
the possibility of directly evaluating the geometrical scales in the variations, i.e.
flocs in paper sheets and fibre suspensions or eddies in fluid flows . It should be
remembered, however, that under turbulent conditions the transformation is not
entirely strict, since "the scanning speed u" in eq. (2) should then be replaced by
the local flow velocity.
A beta radiograph can be analysed in a microdensitometer, and the wave-
length power spectrum of mass formation recorded. By scanning in both the
machine and cross directions, the anisotropy of fibre orientation can be evalu-
ated from the difference between the spectra (22).
To condense the information of a complete power spectrum, it is possible
to present the variations within a few wavelength ranges . An STFI standard is to
express the mass formation number F for the two floc size ranges 0.3 - 3 mm
(F1) and 3-30 mm (F2). The values F1 and F2 then represent the mean floc sizes
of lmm and 10mm respectively.
As already mentioned, the geometrical resolution of the measuring equip-
ment determines the amount of variations recorded. There are approximately
equal contributions to the variance of grammage variations for each decade of
wavelength . If we assume that a measurement of mass formation covers a
wavelength range of 2.5 decades, e.g. 0.1mm - 30 mm, 40% of the variance
would be lost if the resolution were changed from 0.1mm to lmm . Because of
the quadratic relationship between variance and F-value, the latter would de-
crease by approximately 20% due to the change in resolution .
 85

With an image analyser, the two dimensional power spectrum can be eval-
uated, which is useful when floc shapes and two dimensional structures of peri-
odic variations like wire mark are to be characterized. For high resolution in the
power spectrum, the computer of a standard image analyser is not enough. A
mainframe computer is therefore needed for an accurate numerical evaluation of
the one dimensional wavelength spectrum U).

2.3.2. Mass formation scales

There are other ways of characterizing the scale of variations, such as micro
scale and macro scale (23), which are also concepts from turbulence theory.
The specific perimeter is evaluated as the total periphery offibre flocs, de-
fined as areas of above average grammage (24). Mean floc size is then inversely
proportional to the specific perimeter.
A comparison between different variation and scale values to characterize
gammage variations is given in (14) .

2.3.3. Co-occurrence matrix

In image analysis, the texture can be described by second-order statistics using

different algorithms. The spatial grey level dependence method sometimes
called the Co-occurrence Matrix 1(4, 18) has been applied to characterize mass
formation . It quantifies the probability of occurrence of a specific grey level
transition between two pixels of a given spatial separation.
In Cresson's thesis, ( 14), several parameters, see Table 2, are used to
characterize laboratory sheets and simulated sheets, see further section 6.3 .
Unfortunately, detailed comparisons with earlier results are not possible, since
power spectra were never calculated.

Table 2 : Parameters derived using the spatial grey level

dependence method (L4J.

Floc Morphology Texture maps 2:nd-order statistics

Floc probability Probability peak Energy

LWZ probability Ellipse major axis Entropy
Edge probability Ellipse minor axis Correlation
Floc size index Ellipse area Homogeneity
Floc distance index Ellipse eccentricity Contrast
Number of flocs per line Formation index

This new method of texture analysis could be be a useful complement to spectral

analysis of mass formation .
86

3. Fibre flocculation

It can be shown that if fibres are added to water under very gentle mixing condi-
tions, no fibre networks of appreciable strength are formed. However, the intro-
duction of turbulent shear into the suspension causes local fibre deformations,
which result in mechanical network forming when the turbulent shear energy
decays .
The effect of fibre properties and hydrodynamic conditions on the forming
and disruption of fibre flocs is therefore an important research area which has
been dealt with by many researchers. Summaries have been published by,
Parker (2), Norman et al (25), Kerekes (26,2J7 and Wahren (28).
Just as old as the art of papermaking itself is the knowledge that fibre
consistency during dewatering is the most critical parameter controlling the uni-
formity of the wet web . Wahren et.al. (29) showed that fibre flocs are formed
by mechanical entanglement, and that a minimum of three contact points with
surrounding fibres is required to lock a fibre in an unnatural shape. Fibre con-
sistency and fibre slenderness were shown to be the main parameters determin-
ing the degree of fibre flocculation.
An approximate relationship between average number of contact points n,
fibre consistency c and fibre slenderness L/d has been derived by Wahren and
Meyer (30):

- 16nL
cn (3)
3
2L n
d + - (n - 1)d
n n-1

Sediment consistency, which is the final consistency reached when a highly di-
luted fibre suspension is left undisturbed and allowed to settle under the influ-
ence of gravity, is the lowest consistency at which fibre flocs form mechanical
networks. Application of eq.(3) indicates that the average number of contact
points at sediment consistency falls between 3 and 4, which agrees well with the
minimum average value of n=3 for floc forming mentioned above. Usually,
thick stock has to be diluted down to or below the sediment consistency, for a
proper sheet to be formed with conventional forming systems .
Recently, Soszynski and Kerekes (J 1) performed an elegant experiment in
which they formed flocs in a rotating vessel according to the Jacquelin method
(M. At constant rotation they then diluted the suspension, and the lowest start-
ing concentration at which flocs stayed intact during dilution they named
threshold concentration . A comparison between threshold concentration and eq.
(3) is shown in Fig . 3.
 87

Fig . 3: Comparison between Threshold concentration (JI) and

Meyer-Wahren model (eq. 3) t

From fig . 3 it can also be concluded that stiffer fibres have a lower threshold
concentration .
Soszynski and Kerekes further verified the concept of flocs forming from
elastic bending of fibres . They formed flocs from nylon fibres, which were then
treated at a temperature in excess of the glass transition point of the fibre mate
rial. The stresses in the bent fibres in the flocs could relax, and compared to un-
treated flocs, the treated flocs were much more easily broken up by stirring in a
dilute suspension (33).
Riitala has applied percolation theory to evaluate sediment consistency in
relation to eq. (3) with promising results QJ4 .
Optical, "one dimensional", transmission or reflectance methods have gen-
erally been used to study local consistency fluctuations in flowing fibre suspen-
sions 2 2( 5). Recently advanced video techniques have also been applied (M.
From video films it is possible to evaluate both floc shapes and floc behaviour.
A basic study of the forming and dispersion of fibre flocs was performed
using high speed photography in a well defined Couette flow between two
moving transparent walls (J) . Two types of floc breakdown were found: one
global, stochastic process in which a fibre floc is broken down by fragmentation
or stretching and one rate-dependent local erosion process .
88

Laser Doppler Velocimetry is a common optical technique to evaluate vari-

able flow velocities, usually with the help of seeded particles . In fibre suspen-
sions, however, the technique is difficult to apply because of the large amount
of naturally occurring scattering surfaces . These cause secondary scattering ef-
fects, making a correct signal evaluation difficult, unless the fibre consistency is
very low.
Steen used a mixture of benzyl and ethyl alcohol for refractive index
matching between fibres and fluid (37 ) and studied turbulence structure in pipe
flow of fibre suspensions up to 1.2% fibre consistency ($),see Fig . 4: .

Fig . 4 : Turbulentfluid velocity as afunction ofdistance yfrom pipe wall

at Re= 8.5 103 (~L8) .
 89

4. Mix preparation .
The consistency of the thick stock delivered from the machine chest usually lies
in the range 3-4%, which is much too high for conventional sheet forming.
Therefore, the thick stock has to be initially diluted, using recirculated white
water in the short circulation loop from the wire section,.to a mix which can be
fed to the headbox. The mix should be at about the sediment consistency, with
deviations from this value for different paper grades .
In the nix, for the first time, the fibre concentration is low enough for
screening and centrifugal cleaning to be performed using conventional equip-
ment, and they are therefore generally included in the short circulation loop, see
fig . 1, although they really belong to the stock preparation domain. Screening
and cleaning both require reject loops, in which consistency and flow rate fluc-
tuations can be a serious cause ofresidual grammage variations in the final prod-
uct.
The mix usually also passes a deaerator, where air originating from the
white water is removed .
The ideal, as far as grammage stability is concerned, would be to feed the
mix directly from thick stock dilution to the headbox, and thus avoid screening,
cleaning as well as deaeration of the mix flow.
Since the retention of material in the wire secdon is below 100%, material
will be recirculated by the short circulation loop as well as by the long circula-
tion loop . The long circulation recirculates material to positions upstream in the
stock preparation process, which means that internally recirculated material is
also included in the thick stock .

4.1. Thick stock dilution

Traditionally, little thought had to be put into the process of mixing thick
stock with white water. The mixing pump was always running at full speed,
and the mix flow rate was controlled by a throttling valve.The pump was then
generally operating far from its optimum design point, and all the excess energy
was in reality used for mixing . Following the introduction of variable speed
pumps, more attention has to be paid to mixing if the residual grammage varia-
tions are to be kept under control.
Norman and Tegengren (39) added the thick stock at a considerable excess
speed centrally in a straight mixing pipe with white water surrounding it. This
design works according to the jet pump principle, with good mixing properties
at sufficient excess speed of the thick stock feed. As can be seen in Fig . S , a ve-
locity ratio greater than five is needed for good mixing .
90

Fig. 5 : Mixing of thick stock and white water at different ratios

of Thick stock velocity / .White water velocity (DJ.
Top : 1 .5 - Centre: 5.3 - Bottom: 10.8

4 .2 . Screening, cleaning and deaeration.

Unfortunately, screens, and especially cleaners, require quite substantial reject

rates . This necessitates up to two stages of screen reject recovery and five stages
of cleaning reject recovery . This in turn introduces several new loops with dif-
ferent fibre consistencies, which may also vary in time . Finally brought to-
gether, the individual flows will generate consistency fluctuations in the mix fed
to the headbox.
Today there are screens which run efficiently even at high consistencies. If
the thick stock is screened, it would thus not be necessary to introduce mix
screening . High consistency screening is made possible in one design by flu-
idizing the fibre suspension in front of the screen plate (40) and in another de-
sign by a flow situation which generates local dilution in the screen plate area
Git .
It has so far been possible to perform centrifugal cleaning efficiently only
at fibre consistencies below 1 % . Above this consistency, so much turbulent
shear has to be introduced to individualize shives, sand particles etc . from fibre
flocs, that large-scale turbulent mixing overshadows the centrifugal separation
 91

effects . A new cleaning process development is however under way, in which

the fibre network is sufficiently fluidized to separate sand particles efficiently at
3% feed consistency (42).
The shive-removing task of cleaners could probably be taken over by
properly designed screen systems, and this function could thus also be moved
from mix to thick stock.
Since most of the air in the mix flow originates from the recirculated white
water, it would be logical to deaerate the white water before adding the thick
stock see Fig . 6.

Fig 6: Short circulation design for maximum grammage stability

With a design according to the principle shown in fig. 6, all the fibres in the
carefully controlled thick stock flow would be fed directely to the headbox,
which would mean a much more stable operation than with today's conventional
design, see fig. l.
In high consistency forming, the thick stock flow is fed directly to the
headbox, which means that no short circulation loop is required at all.

4.3. Material recirculation

As mentioned aleady in the introduction, material with less than 100% wire re-
tention is recirculated in both the short and the long circulation loops . Material in
the short circulation will return directly in the dilution of the thick stock while
material in the long circulation may take a considerable time before returning as
part of the thick stock, see fig. 1.
92

Three parameters have to be specified to define a retention value.

A) material component,
B) starting position in process
C) end position in process .

The retention R is then defined as follows :

- Amount of component passing start position

R (4)
Amount of component passing end position

Any unspecified "retention" value would today be interpreted as "First Pass

Retention" (FPR), which is defined as follows :

Total consistency of white water in short circulation

FPR = 1 -
Total consistency in headbox (5)

From a comparison of eqs. (4) and (5) it is clear that FPR does not even fullfil
the basic requirements for a retention value, since consistency only and not mass
flow is considered.
Further, all components are included instead of one specific. This is less
suitable, since the retention value is then dependent on stock composition, and
not only on the process as such .
The starting position in the process is defined as the headbox. The end
position in the process is undefined however, since it depends on how white
water is collected to the short circulation, and completely neglects all material
going to the long circulation . On industrial newsprint machines, for instance, as
much as 50% of the "fines" in the machine chest may be recirculated material in
the long circulation .
The retention on twin-wire machines will often be overestimated compared
with fourdrinier machines, if FPR only is the basis for comparison . This is be-
cause on a fourdrinier machine it is possible to separate the rich white water
from the lean, and preferably bring the former to the short circulation. On twin
-wire machines this separation is sometimes not possible, which means that the
material concentrations in the short and long circulations are more equal .
Still, FPR can be of considerable value if applied to a specific system, and
if the white water system is unchanged . To give even more useful results, how-
ever, component retention such as "fines retention" or "filler retention" should
be evaluated, and the long circulation should be taken into account.
 93

In a fully "closed" white water system, all material will finally end up in the pa-
per web regardless of the retention level, but there are still several reasons for
keeping wire retention values high and the white waters as clean as possible .

Cleaner white water means less problem with slime and dirt build up,
which in turn will mean higher machine efficiency.

Cleaner white water means less circulating material, and therefore less
changes in mix composition with fluctuating retention levels . This will
mean a more constant quality of the product.

Cleaner white water means less material carry-over between machines

with connecting white water systems .

Cleaner white water means a faster response to grade changes and other
process changes, and thus a lower level of broke generation . This is
especially important for the long circulation loop (1~).
94

5. Mix distribution
The main task of the headbox is to transform the mix pipe flow into a thin jet
with an extremely precise velocity, direction and thickness across the entire wire
width. In this section, basics of flow in headboxes will be dealt with . Those in-
terested in the design features of today's headboxes are referred to an overview
presented by Waller (44).

5.1. Cross machine distribution

The tapered manifold has been the completely dominating cross-machine distri-
bution system for headboxes for over two decades, and a thorough description
thereof was given by Trufitt (45). In principle, a large number of tubes, or
holes, redirect the mix flow 90 degrees from the manifold channel into the ma-
chine direction .
An equal amount of mix flow is fed into each tube or hole, by keeping a
constant static pressure along the manifold channel. This is possible by design-
ing the channel cross-section area so that the pressure recovery through succes
sively decreasing flow velocity along the channel (Bernoulli equation) is exactly
balanced by the pressure drop along the manifold channel due to wall friction .
The greater the pressure drop across the tube/hole section the less sensitive is the
headbox flow profile to deviations from constant pressure along the channel .
To adjust the static pressure along the manifold channel, the amount of
overflow from the exit side can be varied. If the mix flow in a headbox is in-
creased too much above the original design point, it may be necessary to replace
the channel to attain a constant enough pressure across the machine width.
The separation between the individual holes/pipes leading from the mani-
fold channel must be sufficiently great so that fibre stapling on the downstream
side of the openings is avoided . Further, the flow velocity into the tubes/holes
must be comparatively large, if sufficient pressure drop is to be generated. This
leads to a design with rather low open area on the upstream side of the tube/hole
section .
The cross section of the tubes/holes is generally expanded towards the
downstream side, or their centres are brought together . This is in order to give a
large open flow area at the entrance to the next section, and thus to improve
downstream flow stability .
S yrj ala, S aarenrinne and Karvinen (46) have studied manifold flow using
three-dimensional numerical flow analysis, and compared the results with mea-
surements in an air loop.
Bubik and Christ (4J described discrete increases of the tube cross sec-
tional area, thus introducing the"step diffuser concept". This allowed a con-
trolled turbulence generation with limited eddy size. Model experiments were
performed with air, which allowed the use of hot wire anemometers for mean
 95

velocity and turbulence evaluation.

Lin (4) has made a numerical study of turbulent flow through a circular
step diffuser.
The mix flow from a cross machine distributor is fed either directly to an
outlet nozzle, or into a stilling chamber followed by some pressure drop generat-
ing device such as perforated rolls or tube bundles . In the stilling chamber,
cross flow can occur, thus improving the cross machine evenness of the
emerging jet .

5.2 . Pressure pulse damping

In an air pad headbox, a compressed air volume above the stilling chamber can
absorb pressure variations in the approach flow.
In hydraulic headboxes, a narrow channel may lead from a stilling cham-
ber to a connected air volume, which can absorb high frequency pulsations .
Usually, however, hydraulic headboxes exhibit a very stiff design from a flow
viewpoint. If such a headbox is applied to a fourdrinier machine, special damp-
ing equipment may therefore have to be inserted in front of the headbox to avoid
excessive grammage pulsations, see section 7.2.3. In the damper, an air volume
may be located directly above the mix suspension surface or a membrane may
separate the two.

5.3. Headbox nozzle

Already in 1961, Mardon discussed the relationship between headbox design

and jet quality, and especially the degree of wake effect remaining in the
emerging jet (42). During the late 1960's and the 1970's several investigations
regarding jet quality were undertaken, mainly because of the increasing machine
speeds and the introduction of hydraulic headboxes for twin-wire formers .

5.3.1 . Headbox nozzle feed area and contraction ratio .

To keep a low degree of turbulence in the jet emerging from a headbox, a tech-
nique also used in e.g. wind tunnels is adopted. With a given absolute level of
turbulent energy in the flow entering the outlet nozzle, the degree of turbulence
in the emerging jet can be reduced in proportion to the acceleration of flow. The
larger the nozzle contraction, the larger, the reduction in degree of turbulence in
the jet.
Reiner and Wahren studied the change in the turbulence spectrum in a
headbox jet due to changes in headbox operating conditions (0). They analysed
water flow in a headbox jet using an impact probe for turbulence recording .
Considering the above mentioned relationship between turbulent energy and de-
gree of turbulence as well as the fact that the degree ofjet turbulence is indepen-
96

dent of flow velocity u at constant slice opening h, the following equation was
derived, with which a turbulence spectrum can be transformed to other running
conditions.

ui )2(
h2
E(12) = C U2 E(11) (6)
hi )3
E is the spectral density at wavelength 1 and indices 1 and 2 refer to measured
and normalized conditions respectively.
The application of eq. (6) is demonstrated in Fig . 7. It is evident from the
figure to the right that the equation holds well to normalize the turbulence spectra
at different running conditions within the "linear" range investigated.

Fig. 7: Turbulence spectral density as afunction ofwavelength (5M .

Left: Recorded Right: Normalized

Symbol 0 a a IM o
Slice opening mm 10 20 30 40 SO
Jet speed mls 3 .36 12 .24 12.02 12 .11 12.11

To reduce the wake effects in a "Converflo" headbox, the nozzle is divided into
a number of narrow channels using thin separation vanes (51). It is then possi-
ble to feed each channel from circular openings using a comparatively small
open area, since the separation vanes prevent the generation of large-scale in-
stabilities, see Fig 8 (left) .
As in the case of turbulence, velocity gradients are suppressed by the flow
acceleration in a nozzle. Sanford (S2) studied velocity profiles from Converflo
 97

channels, and the effect of slice opening is demonstrated in Fig. 8 (right) . A

smaller slice opening means larger headbox nozzle contraction ratio, and thus
also better evening out of the velocity streaks from the holes feeding the nozzle
entrance.

Fig 8: Left: Separate channels in headbox nozzle

(divider sheet outside) (51 j
Right: Flowprofile at channel outlet (52).

During the early 1970's, a "high turbulence" hydraulic headbox was designed,
with a tube bundle of nearly 90% open area feeding the headbox nozzle (5) . To
reach such a high open area, the pipes in the tube bundle were deformed into a
hexagonal shape at the downstream end, and tightly packed at the entrance to the
headbox nozzle. The same basic principle with high open area has later also
been adopted by other manufacturers.
Valmet has studied the effect onjet quality of different geometries of the
inlet section to the headbox nozzle . Turbulence spectra were measured, and the
"half energy wavelength" of turbulence was calculated (5J44 . This is the wave-
length dividing the turbulence spectrum into two equal parts, and thus one mea-
sure of the mean scale of the turbulent eddies .
Fig . 9 shows some different inlet designs and corresponding turbulence
characteristics . It can e.g. be observed that the entrance section of type E is three
98

times as high as type B, and the turbulence level correspondingly lower.

Pictures of the jets, Fig . 10, confirm the turbulence values in fig . 9.

Fig . 9: Different inlet designs to headbox nozzles and

corresponding turbulence characteristics (5 .

Fig. 10: Jetsfor type E (left) and type B (right) entrance sections .

At the outlet end, of a headbox nozzle, a sharp contraction may be applied some-
times referred to as the "Parrot's beak". Compared to a straight headbox nozzle,
a parrot's beak will generate a lower level of turbulence in the emerging jet. This
is because, at a given jet thickness, the inner part of the nozzle is more open,
and thus less turbulence is generated .
The degree of contraction of a headbox nozzle equals the degree of accel-
eration of the mix flowing through that nozzle. Flow elements will stretch dur-
ing acceleration, and their cross section will contract correspondingly . This flow
pattern causes fibre alignment, and the fibre orientation anisotropy in the jet is
 99

then a function of the degree of contraction in the nozzle.Already during the

1930's Moss and Bryant found that the fibre orientation in a headbox jet may be
far from random (&.
It should be pointed out that nozzle contraction in a conventional air pad
headbox is generally much higher than that in a modern hydraulic headbox, and
thus also the fibre orientation anisotropy in the emergingjet .
Attempts have recently been started to calculate the turbulent flow pattern
in headboxes using numerical FEM techniques. In one example, simple newto-
nian fluid properties are assumed, and only the first order of variations, i.e. the
degree of turbulence, is considered (56). However, not until it is also possible
to evaluate turbulence scales, i.e. the complete turbulence spectrum, will the
theoretical methods be useful for predicting the effect onjet quality of changes in
headbox design. Today's empirical development could then be made much more
efficient .

5.3.2 . Jet speed, angle, contraction and thickness .

To optimize the papermaking conditions, it is necessary to set the correct jet

speed in relation to the current wire speed.The jet flow velocity can be calculated
from the flow rate, the slice opening and the jet contraction ratio.
Maximum contraction and thus maximum jet speed, occurs at vena con-
tracta, some distance away from the slice opening . After this position, the jet
speed decreases due to air friction. The geometrical configuration of the outlet
nozzle, and in particular the position of the top lip in relation to the bottom lip,
determines the degree of contraction of the free jet.
Usually headbox flow rate is unknown, and the jet velocity then has to be
calculated from headbox pressure. For conventional air pad headboxes this is a
straightforward calculation using the Bernoulli energy equation . The flow ve-
locities inside the headbox are so low that friction effects can be neglected.
In hydraulic headboxes, the internal flow velocities are so high that fric-
tional effects have to be considered, and so has the velocity head at the position
of static pressure measurement inside the headbox .
The jet speed can be adjusted at constant slice opening, which will mean a
change in mix consistency . If mix consistency is to remain constant after a jet
speed change, a corresponding change also has to be made in the slice opening .
Besides setting a correct jet speed, it is necessary to adjust the jet landing
conditions (position and angle) in the wire section, see further section 7.2.1.
Detailed calculations ofjet contraction and angle have been performed by
Appel and Yu (57, 558 . Kerekes and Koller derived numerical equations to
handle the data from the complex equations in ref. Cam) .
To generate a desired grammage profile, it is further necessary to control
jet thickness across the machine width .The outlet nozzle is therefore generally
provided with a deformable slice bar, controlled by a number of slice screws
across the machine. To an increasing degree, control of local slice opening is
10 0

made automatically using special algorithms, taking into account the actual slice
opening profile as well a:. the difference between actual and target grammage
profiles (6Q).
It has long been hypothesized, and it has recently also been theoretically
shown (Ll), that a local deformation of the slice lip, apart from locally changing
jet thickness, will also cause transverse iokr. The transverse flow component
can have a large effect on the final grammage profile, but will also affect local
fibre orientation in the final product. The shape of the outlet nozzle is very im-
portant in this context, and the transverse flow component inci°eases with de-
creasing contraction angle of a nozzle, see Fig 11.

Fig 11 : Effect of nozzle contraction angle on cross direction flow (.61 .

Ideal flow conditions in a headbox can occur only with a constant pressure
along the manifold channel and a constant slice opening profile across the whole
machine width, see further section 8.2. If the slice profile is deformed from this
ideal state, transverse flow is inevitably generated .
Due to later deformation in the wire section as well as in the drying section
of the grammage profile actually delivered from the headbox, the target profile
for the headbox is never constant grammage across the whole wire width.
As an example, preferential cross machine shrinkage at the edges in the
drying section require low grammage edges to be delivered from the headbox, to
achieve even grammage profile at the -P 1 . This in turn requires distortion of the
slice at the edges, and as a consequence transverse flow is generated, causing
local variations in fibre orientation anisotropy . The prevention of uneven cross
machine shrinkage is therefore a prerequisite if anisotropy profiles as well as
grammage profile are to be even all across the paper machine .
Finally, it could be mentioned that the target fibre weight profile at the
reel-up may have to be adjusted because of moisture considerations.

5.4 . Stratified headboxes

The simplest way to form a stratified sheet is from a stratified jet. A stratified jet
can be delivered using a headbox with different mix suspensions in different
layers .
There are two main types of headbox in industrial use today for stratified
forming of two-layer or three-layer products.
In one design, the mix layers are separated by thin foils (Q, which may
or may not protrude from the slice opening, see Fig. 12. . The longer the foil,
the later the mixing between the different layers can start. In a three-layer de
sign, however, the friction conditions in the centre layer differ from those in the
outer layers, and this generates velocity differences between the layers .
The degree of mixing between the different mix layers is very sensitive to
the shape of the downstream end of the foils . They have to be tapered, since
blunt edges cause severe layer mixing.

Fig . 12 : Multiply headbox with thin separation foil (_62J

Top : Foil ends inside Bottom : Foil extends outside headbox.
102

In a second design, the layers are separated using comparatively thick walls
which protrude from the slice opening, and behind which air wedges form,
(U3). In this case, one individual jet for each layer is delivered from the head-
box, see Fig 13.
With air wedges between the different layers, these can be kept separated
until dewatering actually starts. However, air wedges mean that extra surfaces
between mix and air are created, and instabilities in these surfaces will affect
mass formation for the individual layers in a negative way .

Fig 13: Thick separation wall and air wedge in multiply headbox (0) .

5.5. High consistency headboxes

Forming at a fibre concentration as high as thick stock consistency, 3 - 4 %,

would mean large process simplifications, since the short white water circulation
would be unnecessary. This would mean reductions in both capital investment
and energy consumption .
The basic difference between high consistency forming and conventional
forming is the same as that between thickening and filtration dewatering, see
section 6. To form high quality HC-paper, a uniform network of fibres has to be
delivered from a high consistency headbox, since no radical improvements in fi-
bre distribution can be imposed during the dewatering process, in contrast to
what is possible in conventional forming .
Pioneering work in high consistency forming was performed by Reiner
and Wahren (_4) around 1970. The basic idea was to introduce dispersing tur-
bulence energy into a high consistency fibre suspension, by forcing it through a
narrow forming channel with high local pressure drops . After dispersion, the
local energy input was decreased to a level low enough for the turbulence to de-
cay . The flowing fibre suspension then transformed into a fibre network, and
 103

the generation of large scale flocculation during the turbulence decay phase was
avoided "mechanically" by the narrow dimension of the channel . Different flow
channel shapes were introduced, to induce shear of suitable levels, resulting in
an improved fibre distribution (&, see Fig. 17.

Fig 17: High consistency forming channels according to

patent drawings (65)

A further improvement was the introduction of an "eddy chamber", fed from the
individual pipes of the cross machine distributor M).The fibre suspension is
fed tangentially into the chamber, and to reach the outlet, the flow has to make a
sharp turn. This causes a large contraction of the flow into the forming channel .
A high pressure drop is generated, which is benificial for the final cross distri-
bution of fibre suspension inside the eddy chamber. Furthermore, this design
has good runnability, since it can even be started filled with a high consistency
fibre suspension.
Recently high consistency headboxes with flow channels similar to those
described in ref. (65) have also been tried in a Japaneese development project
(U.
104

6. Fibre deposition
Radvan et.al. demonstrated that fibre suspension dewatering can be either a fil-
tration or a thickening process (68). Fig 15 illustrates the principle difference
between the two.

Fig . 15 : Fibre deposition by:

Left: Filtration process Right: Thickening process (2) .
In the filtration case, the fibres are laid down individually from a dilute fibre
suspension. The result is a strongly layered sheet structure with rather poor me-
chanical properties in the z-direction . Conventional sheet forming of the four-
drinier type and of the twin-wire type are thought to be mainly filtration pro-
cesses .

6.1. Laboratory sheets

The main reasons for making laboratory sheets is to evaluate the potential of a
given pulp, and not to simulate industrial forming conditions . In laboratory
sheet forming processes, a high degree of dilution is therefore used, to avoid fi-
bre flocculation and thus to make as uniform a sheet as possible. These pro-
cesses are therefore typical filtration processes .
The British Standard Handsheet is the traditional laboratory sheet, formed
from a fibre suspension of about 0.02 % consistency, which is about 25 more
dilute than in industrial forming . Because of the high dilution, the sheet build-up
follows an ideal filtration process, resulting in a far better than random fibre
distribution, see section 6.2. The result is a laboratory sheet with mechanical
 10 5

and optical properties superior to those of machine-made paper.

On the other hand, the retention using the original British Standard
Handsheet Machine is comparatively low, due to the absence of white water re-
circulation . For mechanical pulps, or for products where fines or filler content is
important, means of recirculating the white water should therefore be considered
Q0.
The fines and filler tend to be concentrated to the wire side in laboratory
sheets, while the opposite is true for machine-made fourdrinier sheets . In twin-
-wire sheets, the distribution generally shows a symmetrical form with maxima
somewhere between the centre and the surfaces (2) .
Formette Dynamique (M) is a rotating laboratory sheet former developed
in Grenoble. It dewaters a sheet on a removable bronze wire, placed against the
inside of a perforated drum, and the dewatering pressure is generated through
centrifugal forces . One advantage of this former is the large sheet size of 250
mm x 800 mm, which allows more complete testing than the smaller, circular
British Standard Handsheets. Furthermore, Formette Dynamique is well suited
to the making ofmulti-ply sheets.
The fibre consistency in the feed to the former is as high as 0.3-0.4 % but
normal fibre flocculation is still avoided since the actual fibre deposition onto the
web takes place by centrifugation through a 10 mm thick water layer . This
makes it easy to align fibres in the direction of rotation, and thus to make ori-
ented sheets. With Formette Dynamique, it is in fact difficult to make isotropic
sheets .
When evaluating paper samples taken from industrial machines or pilot
machines, a comparison with laboratory sheets made from the same furnish is
useful.- The ratio between the properties of a machine sample and those of a lab
oratory sample is a measure of how well the potential of a specific furnish is de-
veloped on the machine, i.e. of the Forming Efficiency . When evaluating FEX
samples, Formette Dynamique laboratory sheets made from the same furnish are
used as reference for the calculation of Forming Efficiency, (71).

6.2. Random sheets

It was already suggested by Wrist (1) that dewatering on the Fourdrinier was to
some extent a self-healing process The reason for this was that areas of lower
than average local grammage also exhibit lower than average dewatering resis-
tance. Therefore, extra dewatering would take place at areas of low grammage,
and the extra fibres deposited would then result in an overall levelling of gram-
mage over the sheet area.
Early simulations of a random sheet structure by Corte and Kallmes (7)
were made by drawing black lines, representing fibres, with random position
and orientation. In this way, however, only very low grammages sheets could
be represented, since at realistic sheet grammage the random sheet would appear
almost completely black, with no floc structures. The grammage was therefore
10 6

limited to a few g/m2, see Fig 16 . The simulated random sheets were compared
with real sheets of corresponding grammage and fibre dimensions .

Fig . 16 : Sheets with grammage 2.5 g1m2 (Z21

Left : Random network of straight lines
Right : Real sheet .

Because of the low grammage, with a correspondingly open sheet structure, no

true conclusions could be drawn from these experiments regarding a self-healing
effect . The improvements for real sheets could be expected to start at a much
higher grammage, probably greater than 10 g/m2 , and to increase with gram-
mage.
Corte also calculated the mass formation for random sheets and compared
the results with measurements of beta ray transmission through real sheets (6) .
His conclusion was that the random sheet had the ultimate degree of uniformity,
at which to aim in the sheet forming process .
Norman et al calculated the mass formation wavelength spectra of random
sheets and compared them with the corresponding spectra measured on beta ra-
diographs of well formed laboratory sheets (22) . The results clearly demon
strated that random sheets were much more uneven than laboratory sheets in the
small scale wavelength range . This was thus a confirmation of the idea sug-
gested by Wrist, regarding the self-healing effect of the dewatering process on
local grammage variations .
The difference in geometrical resolution is one possible explanation of the
different results obtained by Corte and Norman . Corte used a square measuring
area of 1 mm size, and also calculated the grammage variations for a random
sheet at the same resolution. Small-scale variations, below 1 mm, were therefore
never considered. Norman, on the other hand, used a resolution of 0.1 mm both
 10 7

in the measurement and in the calculations .

In Norman's case, the large amount of random variations in the small-
-scale region overshadowed the more even random structure at large wave-
lengths, in comparison with the real sheet . In Corte's case instead, the larger
variations obtained for real than for random sheets could have been caused by
the more frequent occurence of fibre flocs in the real sheets .
To make a more subjective comparison possible, a picture of a random
sheet was prepared by Norman et al (,using a method in which fibres were
computer drawn on an oscilloscope screen . The grey level of each fibre was set
so low that up to ten fibres could be placed on top of each other without reach-
ing grey level saturation. This made it possible to study the details of a 40 g/m2
random sheet and to make comparisons with the corresponding laboratory sheet,
see Figs . 17a and b. From these pictures it was clear that there are more small-
-scale variations in the random sheet than in the real sheet.
Norman et al also analysed anisotropic random sheets . Mass formation
wavelength spectra were calculated both in the direction of main fibre orientation
(MD) and in the perpendicular direction (CD). The two spectra cross each other,
with the MD spectrum below at small wavelengths and above at large .The dif-
ference between the spectra is a measure of fibre orientation . The mass forma-
tion number is the same in both directions, however. Corresponding results
were obtained in actual measurements on paper sheets as well as on highly ori-
ented nonwoven sheets .

Fig . 17: Mass distribution in 40 glm2 sheets QJ2

Left: Computer drawn random sheet.
Right: Beta radiograph of laboratory sheet.
10 8

Cresson 14) in his thesis, which can also be recommended for its excellent lit-
terure reference coverage, besides simulating isotropic and oriented random
sheets also simulated the improving effect on mass formation of drainage level-
ling. Introducing two-dimensional fibre flocculation made it possible to simulate
the structure of laboratory sheets made with delayed drainage . Further simula-
tions on dewatering of flocculated structures are reported by Gorres, Cresson
and Luner (22), see Fig 18.

Fig . 18 : Simulation of flocculated forming (M .

A : 1 g/m2 , B: 2.5 g/m2, C: 10 g/m2 , D : 22.5 g/m2

Further work along these lines should prove useful to improve the understand-
ing of the forming process . To be of real use, however, more realistic fibre
characteristics should be included as well as three-dimensional models, shear
flow conditions and fibre network strength .
Steen will discuss the process of floc generation and breakdown in a con-
tribution at this symposium (74) .

6.3. Filtration dewatering .

Early studies of dewatering resistance were carried out by Ingmanson and co-
workers (75), according to the theories of filtration. They based their analysis
 10 9

on the Kozeny-Carman equation:

3
_dQ__1 (1-C) 1
dt K S 2 C 2 -Op

where dQ/dt is rate of drainage per unit area of the web, Ap is the pressure gra-
dient across the web, C is the volume fraction of the web occupied by solids, S
is the specific surface area of the solids per unit volume, g is the viscosity of the
fluid and K is the Kozeny Constant.
The Kozeny-Carman equation assumes laminar flow through a bunch of
parallel capillaries in an incompresssible medium. Neither of these assumptions
are true in the case of wet paper webs. The Kozeny-Carman equation in its pre
sent form can thus not be used for useful predictions . Substantial modifications
must be incorporated, before it can successfully be applied to paper webs.
Application of the equation results in values of the surface area S an order of
magnitude different from the surface area estimated using other methods (3) .
In connection with his discussions of the hydrodynamics of dewatering,
Radvan states that "the progress of technology has overtaken the need for de-
tailed understanding of the original process" . There may be reason to put a
question mark to the use of the word "need" in this context, but there is no
doubt whatsoever that the rapid development of fourdrinier as well as twin-wire
dewatering principles, which has taken place also after Radvan's statement, has
completely outpaced the corresponding theoretical understanding.
During the dewatering process in the wire section, a suspension is forced
against a compressible, porous web, backed by the surface of a woven wire.
The solid particles such as fibres, fines, filler etc. in the suspension are to a cer
tain degree trapped, initially by the wire surface and later in the pores of the ac-
cumulating wet web . The resistance to flow through the web is influenced by
several parameters, such as:

the distribution of material and voids in the z-direction and in the plane
of the web

the chemical properties of web and suspension components

the mechanical compression properties of the web

the mode of flow (laminar, turbulent) through the pore system

the time event of the pressure driving the dewatering

It follows, that the dewatering of a well defined furnish subject to a known

pressure (or force) event is extremely difficult to describe. Schopper Riegler
number or Canadian Standard Freeness level, as evaluated in standard labora-
tory equipment is not enough to predict the actual dewatering capacity on a paper
machine.
Fundamental work on dewatering has been summarized by Radvan (3) .
Several devices to study filtration dewatering resistance have been developed.
They are based on constant filtration rate, constant filtration pressure or combi-
nations of these two modes . Since these idealised conditions differ significantly
from those prevailing in a real wire section, attempts have been made to make
realistic modifications of the laboratory devices .
One example is the development by Pires, Springer and Kumar (76), who
used a Britt Jar device 77 and with some modifications designed a "Drainage
and Vacuum Retention Tester". The introduction of six electrodes made it pos-
sible to follow the drainage event, and shear rates in the fibre suspension as well
as pressure events on the downstream side of the wire comparable to those en-
countered in paper machines were introduced. They used the device in connec-
tion with modelling of fourdrinier drainage on a pilot plant machine, see further
section 7.2.5 .
For practical use, empirical equations are often used to predict the influ-
ence of process conditions on dewatering. One example is the equation of
Wahlstrom and O'Blenes (78):

G
t- wa (d p )n 8
C

where t is the time to form a web of grammage w, under a constant pressure

drop Ap. C is the suspension consistency and G, a and n are empirical con-
stants .

6.4. Thickening dewatering .

Even in a conventional forming process, all fibres are not free to move individ-
ually during dewatering. Some fibres are mechanically connected in fibre flocs,
which are successively compressed during dewatering. This event can be de-
scribed as a thickening process . Compared to filtration dewatering, the resulting
sheet structure will be more "three-dimensional", since the fibres are more en-
tangled in the z-direction. This can be expected to give improved properties in
the z-direction, together with some losses in the plane of the sheet.
Britt, Unbehand and Shridharan (80) suggest that the level of fines reten-
tion is a measure of the distribution between filtration and thickening dewater-
ing . A low retention level indicates mainly thickening dewatering, while a high
retention level is an indication of filtration dewatering. In experiments using a
Britt Dynamic Drainage Jar with controllable dewatering vacuum, they could
apply a desired degree of dewatering pulsation . They thus claim to be able to
simulate real dewatering processes by applying suitable dewatering pulses . A
uniform dewatering pressure would result in pure filtration dewatering, while
pressure pulses would generate their definition of thickening deatering.
When the filtration process is finalized on a fourdrinier wire or in a twin-
-wire section, further dewatering will cause consolidation by thickening. In the
twin wire case, when the web is squeezed between the two wires, the web will
be subject to mechanical compression. This type of dewatering follows the same
basic rules as those in wet pressing, which means that the pressure applied is
divided between hydraulic and mechanical pressure in the web . The amplitude
of pressure applicable is limited by the allowable amount of wire mark em-
bossed into the web surface.
The last part of dewatering in the wire section is performed by air suction .
Water removal during this phase is a displacement process in which water is ex-
changed for air. This process differs considerably from a filtration or a thicken
ing processes . The sheet dryness after the couch is therefore not a simple func-
tion of the flow resistance of the web .
It should be advantageous for final web dryness to keep the sheet mechan-
ically compressed during air suction. Also, as little water as possible should be
left in the wire, since much of that water may follow the web at the point of sep-
aration of wire and web .
Britt and Unbehend (IIJ found in laboratory experiments that different
stocks often behaved quite differently during initial dewatering (up to the dry
line ) and during the final vacuum dewatering . There is no general correlation
between initial and final dewatering rates, which makes it difficult to predict the
overall dewatering characteristics for a specific paper pulp.
7. Mix dewatering
When the mix suspension is dewaeerind in the wire section of a paper machine,
the actual process differs highly from the idealised conditions in laboratory sheet
formers . Since the fibre consistency in real forming may be at least an order of
magnitude higher than that in laboratory sheet forming, measures have to be un-
dertaken to improve the mass formation of the paper produced. These measures
are: drainage, oriented fluid shear and turbulence (2), see Fig . 19.

Fig . 19: Drainage, Orientedfluid shear and Turbulence (2) .

As mentioned in section 6.2, drainage has a self-healing effect on grammage

variations. Oriented shear is originally generated by a velocity difference be-
tween mix jet and wire, and may also be generated by shaking of a fourdrinier
wire and by gradients in the fluid pressure applied in twin-wire formers .
Oriented shear will move free fibres in a fibre suspension relative to one an-
other, and may also disrupt fibre flocs. Oriented shear will affect the orientation
of fibres and fibre segments, and will therefore have a strong effect on the
anisotropy of the sheet properties.
Turbulent velocity fluctuations may occur in the mix jet from the headbox,
and may also be generated in the wire section. The scale of the turbulence is vital
for its interaction with the fibres in a suspension. Only eddies up to the same
order of size as the fibre flocs will have a disrupting effect, while larger eddies
will mainly superimpose floc motion . On a fourdrinier wire, turbulence can be
generated by vertical movements of the wire as it passes over dewatering ele-
ments . In twin-wire forming, turbulence can be generated by pressure fluctua-
tions between the wires.
The degree to which these effects occur in a given drainage section and the
degree to which they are relevant to reach desired product qualities is still mainly
unknown, and much more research is needed to extend the knowledge within
the area.
In this chapter, basic process aspects in the wire section will be discussed,
starting with forming wires and followed by fourdrinier forming, twin wire
forming and multi-ply forming .
7.1. Forming wires

Helle recently presented an overview of the development of forming wires dur-

ing this century ($2)-
Earlier, forming wires were made of bronze, they were single layered and
of two shed design. Neither the material nor the design was a good choice for
maintaining a long wire running time combined with high paper quality on high
speed paper machines . When wires of polymeric materials were developed in
the 1970's, the bronze wires therefore quickly disappeared from the market.

7.1 .1 . Wire surfaces

For good wire wear resistance, it is an advantage to have as much CD yarn ma-
terial as possible facing the stationary wire support. This means that the machine
direction threads, which take the main mechanical load, are subject to less me-
chanical abrasion, which in turn increases wire life. The influence of wire de-
sign parameters on the amount of material available for wear has been described
in a theoretical model by Batty (U).
The wire side of a paper replicates the forming wire to some extent. A
predomination of paper side MD nuckles, which is the case for a single layered
wire with a predominance of CD nuckles on the bottom side, leads to increased
deflection of fibres into the wire openings, "blinding" and reduced dewatering
capacity ($_4). The preferential machine direction orientation of fibres on the wire
side of the paper is one reason for this effect . Beran (;~5 introduced a "Fiber
Support Index" to quantify the top surface quality of a wire.
The surface of the wires is also important for the generation of wire mark,
which can be analysed using optical techniques M).
It is thus obvious that a single layered forming wire cannot be optimised
for both running time and paper quality. The introduction of polymeric materials
made it possible, however, to design double layered wires, on which the top
and bottom sides are more independent. Recently the use of "shute support" on
double layered wires, which extra threads on the paper side of the wire has im-
proved the support for the fibres during dewatering . Further, multi-layered
wires have been introduced, which in principle consist of a fine top wire for pa-
per support combined with a coarse bottom wire for runnability . Multi-layered
wires are now introduced for most paper grades ($7_) .
Good mechanical wire properties are important. This is accentuated be-
cause the dewatering pressure in twin-wire formers is generated by wire ten-
sion, and a uniform tension profile across the whole machine width is essential
for a good cross machine grammage profile .
 7.1.2. Dewatering properties

The traditional criteria for evaluating the drainage efficiency of forming wires
were:
the relative vertically projected open area
the air or water flow resistance.
The relative open area concept ignored the fact that modern wires exhibit a
highly three dimensional structure, in which the openness varies considerably in
the thickness direction of the wire.
A better way of characterizing the openness is therefore to describe the
distribution of material and voids in the thickness direction of the wire (88), see
Fig 20 . The influence of MD and CD threads can then be separately quantified.

Fig 20 Mass I void volume distribution through wires

Left: Single layer wire Right: Multi-layer wire

The wire design is also important for the level of retention of fibre material, but
the size of the wire openings makes it extremely hard to filter out fines and filler,
see Fig . 21 (88).
Johnson (89) studied the deposition of air-suspended fibres onto different
types of wires in a 32 times scaled up experiment . He found an improved reten-
tion going from a single layered to a double layered wire design . He also found
that drainage was strongly influenced by the degree of sheet support for the first
layer of fibre web .
 Fig . 21 : Size ofwire mesh "window" is 200yx 200,u ($$).

Fibre carry back is a term which refers to fibres which do not follow the wet
sheet on its removal, but which are instead left on the wire surface and thus re-
turn to the forming zone. These fibres disturb sheet dewatering, by locally in-
creasing the dewatering resistance of the forming wire, and cause a deterioration
in mass formation . In twin-wire forming, the geometrical and mechanical prop-
erties of forming wires have a marked influence on the amount of fibre carry
back (9Q).

7.2 . Fourdrinier dewatering

The history of fourdrinier dewatering was recently presented in an interesting

paper by Hansen (91). Classical fourdrinier dewatering has been described in
some detail in references (1_), (2) and (3).

7.2 .1. Initial dewatering

The setting of the jet onto the wire and its initial dewatering is extremely impor-
tant for both the dewatering capacity and the product quality. The jet should
ideally hit the wire immediately in front of a supporting forming board.If the jet
hits the wire much ahead of this board,too rapid dewatering will result in a dete-
riorated mass formation , whereas air will be trapped and disturb forming if the
jet hits the wire on top of the forming board.
 wahren has theoretically analysed the situation when ajet hits a wire. He
postulates that the ideal situation occurs when that part of the jet which remains
on the wire will move horizontally as a result of the impulse exchange when part
of the jet is deflected downwards C2)2 . The influence ofjet angle, jet thickness,
horizontal and vertical distance from headbox to forming board, front angle of
forming board and fraction of jet deflected downwards is analysed.
Osterberg applies a laboratory former to study the initial part of the dewa-
teri.ng process (22).

7.2.2. Jet velocity pulsations

Pressure pulsations in the mix feed flow generate jet velocity pulsations when
hydraulic headboxes are used. In fourdrinier forming applications, these veloc-
ity pulsations can result in serious grammage variations. Grammage variations
up to ten times as large as the jet velocity variations have been observed.
One explanation for this amplification is the generation of standing waves
on a fourdrinier wire (2 4 .Maximum amplification will occur when maximum
wave amplitude coincides with the dry line. The distance from headbox to dry
line then corresponds to one fourth of the total wavelength of the standing wave,
and the frequence is then normally in the range 10-20 Hz.
Another explanation of the amplified grammage variations could be the
forward and backward movement of the position at which the jet hits the wire

In twin wire forming, jet velocity pulsations do not generally cause seri-
ous problems. Since the bet is immediately enclosed between two wires, amplifi-
cation due to standing wave generation is avoided .

7.2.3 . Improvement of mass formation

To improve mass formation, horizontal shaking of the wire has been very use-
ful, especially at low speeds. Recently it has however also been applied at
speeds as high as 600-700 m/min.
A dandy roll, applied on the wire at a position where the average consis-
tency is aound 2.5% improves mass formation by introducing shear in the sus-
pension . A better effect would be achieved if the dandy roll were applied at even
lower consistencies, but then severe downstream spouting cannot be avoided.
This is because the dandy roll is not a dewatering element . and the suspension
consistency is therefore approximately the same on the upstream and the down-
stream sides of the roll.
A new dandy roll application has been presented by Kallmes and
Langdok, who used the roll for dewatering a). They led the fourdrinier wire
upwards on the downstream side of the dandy roll and also introduced a return
roll, see Fig 22 Having watched a demonstration run of this principle on the
IPC experimental wet end, I am still amazed at the way water clung to the inside
of the roll, allowing axial removal at "9:00 o'clock".

Fig. 22 Scheme ofMIK Dandy Roll Former (W.

Mainly because of cleaning problems, dandy rolls have not been used on high
speed machines . As far as the effect of mass formation improvement is con-
cerned, the twin wire former can be interpreted as being a development from the
dandy roll, since it can be run with low ingoing fibre consistency and still yield
a sufficiently high outlet consistency. Many fourdrinier machines, especially
those manufacturing printing grades, have been modernized to hybridformers
(see section 7 .3.3) by the introduction of "extended dandy rolls" in the form of
top wires.

7.2 .5. Dewatering devices .

Originally, all dewatering relied on gravity effects, and supporting rolls were
introduced only to keep the wire horizontal while causing a minimum offriction
drag. Today, dewatering devices are applied to fourdrinier wires with the object
both of creating dewatering effects and also of providing means for controlling
the degree of fibre flocculation in the sheet formed .

Taly1e rolls

The basic principle for table roll dewatering was understood during the 1950's,
and was described in work of Wrist (22), Taylor (98,99) and Bergstrom
At a table roll, a suction pressure peak is generated in the downstream ex-
pansion zone between wire and roll. This causes a local downward deflection of
the wire immediately after the roll, and a corresponding upward movement of
the wire then has to occur before the next dewatering element is reached. The
vertical motions of the wire generate turbulence in the suspension on the wire,
which may to some degree improve final mass formation .
Pires, Springer and Kumar developed a computer program for the evalua-
tion of drainage on a fourdrinier wire supplied with table rolls and vacuum
boxes. They used the equation for table roll drainage developed by Taylor and
determined the specific filtration resistance experimentally using their DVRT-
tester (M), see section 6.
The amplitude of the table roll suction pressure is proportional to the
square of the wire speed. There is an upper speed limit for the application of
table rolls of approximately 500 m/min, above which the vertical wire movement
is so intense that the high degree of turbulence generated disturbs the final mass
formation.
A modified table roll which has recently been introduced is the Sonic Roll
(1 Ql ). It can be described as a gear wheel of wire section width, with each tooth
broken at regular intervals along its length . During rotation, each tooth causes a
local upward lift of the wire, and the amplitude of the wire movement is deter-
mined by tooth width and separation. For very slow machines it has been found
useful to drive the Sonic Roll at an excess speed, which makes it possible to
shake the wire at its resonant frequency

Foils

Foils were introduced during the 1960's (1), and took over as the standard de-
watering element at speeds where table rolls could no longer be used. Just as in
the case of table rolls, the maximum amplitude of the suction pulse is propor-
tional to the square of wire speed - in fact it equals the dynamic pressure of the
mix on the wire . However, by choosing a very small angle between foil surface
and wire, as low as even below one degree, a suitable suction pressure can be
achieved even at high wire speed .
Usually several foil elements are used in combination, and to increase the
overall dewatering effect they are mounted in vacuum boxes.
The overall optimization of the individual foiles angles and separation dis-
tances is a complex procedure . Schmid (1 3) performed an experimental study,
and evaluated the turbulence level introduced by watching the surface of the
suspension on the wire using a strobe light. The amplitude of surface distur-
bances was graded as "activity" in numbers from 1 (smooth surface) to 10
(severe stock jump). Among other things he found that an increase in foil angle
along the wire gave an optimum combination of dewatering capacity and sheet
mass formation.
 Step f i1

One disadvantage with conventional foils is that a complete element has to be

replaced if the dewatering pulse is to be changed . This makes optimization of
sheet mass formation through suitable activity in the mix on the wire rather
complex.
The Step foil is a design in which the mix activity can be controlled by
vacuum (1~}4 . The distance between the foil element and the wire undergoes a
step increase, see Fig . 23, and the vacuum level in the foils box determines the
extent to which the wire is deflected into the step. An increased vacuum level
causes a larger vertical wire deflection, and thus create a greater activity in the
mix on the wire.
Cadieux 1(,~4) performed a study on 2 pilot fourdrinier machine using step
foils for the main part of the dewatering, and on the basis of the results he set up
a mathematical model to describe dewatering. The equation is an expansion of
the Kozeny-Carman equation, and the difference between a "static" laboratory
sheet former and dynamic fourdrinier dewatering is taken account of by a
Fourdrinier Factor. Of all the 17 independent variables tested, those found to be
most significant were fibre compressibility, the first zone of the drainage ele-
ment pressure pulse, drainage element position along the wire and headbox jet
configuration .

Fig . 23: The Step Foil (104J.

ISO-flo foil

Iso-flo foils are horizontal support blades mounted in a vacuum box. Every sec-
ond blade is slightly lowered, and thus the wire is be subject to vertical move-
120

ments when running over the box with vacuum applied. The degree of activity
on the wire is then controlled by the vertical distance between the blades .

as ade foil

A cascade foil consists of several gently sloping sections with horizontal con-
nections . A single cascade foil can have a capacity of several millimetres of de-
watering, without the application of vacuum. Contrary to ordinary foils, the cas-
cade foil does not introduce any activity in the fibre suspension on the wire. The
reason for this is that the vertical wire movements are stabilized by the introduc-
tion of the neutral zones between the dewatering zones

Flexi-Former

A dewatering device which acts from the top of the mix suspension on the wire
is the Flexi-Former (M. A flexible (or stiff) blade is pressed against the mix,
and dewatering occurs by the pressure generated in the suspension between a
stationary upper surface and the moving wire . See further description in section
7.3 .1 .

7.3. Twin-wire dewatering

Principles for twin-wire forming were patented as early as during the last cen-
tury, but not until the 1950's were machines developed which could be used for
practical paper/board making. The new developments avoided the simultaneous
mechanical guiding of both wires within the forming zone, a task impossible to
perform with the required accuracy. The secret to success was always to let one
of the two wires automatically adjust to the wire separation required by the pre-
vailing forming conditions .
David Webster's ( ) privately developed twin-wire former for paper
production and Brian Attwood's (1_Da) "Inverform" design, developed for board
production at St Anne's Board Mill, were the first two designs of practical
value.
The Webster patent for the rollformer principle was acquired by Barje
Wahlstrom of KMW, where the basic development continued and commercial
equipment was produced. The roll former concept was also studied at PPRIC
(1 9 and equipment was manufactured in co-operation with Dominion
Engineering Works (110) .
Twin-wire blade formers with stationary dewatering elements were de-
veloped by Beloit 4(11 l) and Black Clawson 112 .
In pure twin wire formers, the mix jet is delivered directly into the gap

4An earlier development by Times Inc. was taken over by Beloit .

between the two wires, whence the name gapformer
In a hybrid former, a fourdrinier section precedes the twin-wire nip .
Valmet's twin-wire formers have mostly been of the hybrid former principle
11 ), and during the early 1980's, Malashenko of Dominion Engineering sug
gested a top wire forming unit (114) to update fourdrinier machines to hybrid
twin-wire formers .
The basic advantage of twin-wire formers compared to fourdrinier ma-
chines is the closed forming zone with the possibility of two sided dewatering.
Two sided dewatering means high dewatering capacity and the possibility of
symmetrical dewatering . A closed forming zone prevents fibre movement by
surface waves, and makes it possible to introduce comparatively large amounts
of dispersing energy without the risk of free surface instabilities .
In twin-wire formers, two webs are formed, each of half the total gram-
mage. This means that a lower retention level can be expected in comparison
with fourdrinier forming, since retention increases with web grammage.
An overview of different twin-wire formers available in 1987 was pub-
lished by Thorp 1I) .

7.3.1. Roll formers

During the 1960's, paper scientists tried to describe theoretically the interrela-
tionship between twin-wire forming conditions and dewatering pressure. The
energy equation (Bernoulli equation) was applied, and the mix jet followed into
a predefined space between the forming wires. This was an unsuccessful ap-
proach, since the primary variable in reality is the fluid pressure generated by a
tensioned, curved wire, and the Bernoulli equation does not appply when outer
forces are present .
In a twin-wire roll former, a mix jet is injected into the nip between two
wires wrapping a rotating roll. The outer wire is given a tension T, and using
classical mechanics the fluid pressure p generated in the mix to support the wire
is given by the equation

p = T/R (9)

where R is the local radius of curvature of the wire.

During most of the dewatering phase, the roll radius is an acceptable ap-
proximation for R, since the separation between the outer wire and the roll sur-
face is small relative to the roll radius. Dewatering can take place through both
the outer and inner wires . The latter water is kept in the open roll surface and
released when the inner wire has been removed from the roll, see Fig . 24.
During the initial phase of dewatering, the radius of wire curvature cannot
be approximated by the roll radius, since the wire approaches the forming zone

5 1ntroduced by Antti Lehtinen, Valmet .

12 2

along a straight path, R = -. There is a gradual decrease of outer wire radius

from infinity to roll radius, and a corresponding build up of dewatering pres-
sure.

Fig . 24: Twin-wire rollformer with two sided dewatering.

The dewatering pressure has been measured in roll formers with one-sided de-
watering, with a pressure sensor mounted in the surface of a solid forming roll
11 lam). An example of a pressure curve is shown in Fig . 25:

Fig . 25: Dewatering pressure in a one-sided, twin-wire rollformer (1O.

 123

Ingemarsson developed a model for the initial phase of roll former dewatering,
118).

Jet/wire speed ratio

On a fourdrinier machine, minimum fibre orientation anisotropy is achieved at a

jet/wire speed ratio of unity, which corresponds to equal velocities of jet and
wire.
In a twin-wire gap former, ajet/wire speed ratio in excess of unity has to
be used for minimum anisotropy. This is because the jet is decelerated when
entering the pressure zone between the wires, and not until after this initial de-
celeration shouldjet speed equal wire speed for minimum fibre alignment .
The following equation for the jet/wire speed ratio required for minimum
anisotropy can be derived using the Bernoulli equation. It applies to a twin-wire
former with wire radius of curvature R, wire tension T, suspension density p,
jet velocity uj and wire velocity uW

uj
-
u = l+T
R
2
2 10)
puW

It follows from eq. (10) that the jet/wire speed ratio for minimum fibre orienta-
tion anisotropy increases with decreasing machine speed, and this is exemplified.
in Table 3

Table 3: Jet/wire speed ratiofor minimumfibre orientation

anisotropy with T=6 k1V/m and R = 0.7m .

Wire speed m/min 100 200 400 1000 2000

Jet/wire speed ratio 2 .68 1 .59 1 .18 1 .030 1 .008

On entering the forming gap, the jet decelerates and then has to undergo a rela-
tive thickness increase equal to the jet/wire speed ratio given in the ,table.
Too high an expansion is not possible while maintaining stable flow con-
ditions, and therefore a roll former cannot be run at speeds below approx . 200
m/min.

Centrifueal effects

If a fluid element of radial extension h and density p moves with velocity u along
a path with a radius of curvature R, a centrifugal pressure pc develops according
to the equation :
124

2
u
p.= ph (11)
R

It is often claimed that the dewatering pressure in a roll former is generated by

centrifugal forces . This is however entirely incorrect, and centrifugal forces in
fact only tend to decrease inwards dewatering 1(07,) . To counteract this effect,
and thus still get symmetrical dewatering, roll formers are supplied with vacuum
zones to increase inwards dewatering, see fig. 24.
The ratio between centrifugal pressure pc and wire generated pressure pw
can be described by the following equation 119

PC phu 2
(12)
PW T

If the centrifugal pressure were to exceed the wire pressure, unstable conditions
would result. This means that there is a practical upper limit for the pc/pw-ratio
of unity. From eq. (12) it is clear that these conditions are reached independently
of the local radius of wire curvature R. Corresponding values for fibre suspen-
sion radial thickness h and machine speed u are shown in Fig . 26 (119) .
Machine speeds in excess of those indicated in fig . 26 would result in un-
stable forming conditions .

Fig. 26: Maximum wire speed uz as afunction of the suspension radial
thickness h and wire tension T according to eq. (12)
 12 5

One-sided dewaterinta

Twin wire roll formers are generally equipped with "open" forming rolls, allow-
ing two sided dewatering, see fig .24. One exception is in tissue manufacture .
For cost reasons, a solid forming roll is then used, and thus only one-sided de-
watering. This is possible since neither the two-sidedness of the sheet nor the
dewatering capacity is a limiting factor .
Another device for one-sided dewatering is the Flexi-Former (106,
which has also been applied to industrial forming of tissue . During dewatering,
the fibre suspension is sheared between a stationary upper surface and the run-
ning wire. Under these extreme shear conditions, mass formation is better than
that achieved with fourdrinier dewatering. When a flexible plate is used, the de-
watering pressure is controlled by the vacuum below the wire, the frictional drag
forces along the plate, wire tension andjet/wire speed ratio. An example is given
in Fig. 27.
It is also possible to use a stiff, curved plate. Vacuum is then unneces-
sary, and the local dewatering pressure is determined by the local radius along
the backing plate . This principle has been used in high consistency forming, see
section 7.4.

Fig . 27:Pressure distribution in Flexi-Former at580 m1min (16

Jetlwire speed ratio : 1 .0
------------- 1 .2
1.0 (white water only)
12 6

7.3.2. Blade formers

As late as during the Internationl Water Removal Conference in London,1975,

the basic principle for a blade former was explained as being due to the same
mechanism as that in roll forming (1~). The dewatering pressure over a Bel
Baie forming shoe was calculated according to eq. (9), and the radius of
mounting for the blade tips was inserted instead of the radius R of wire curva-
ture. The radius was one order of magnitude larger than that of a roll former.
The hypothetically lower dewatering pressure was suggested to be the main rea-
son for the improved mass formation obtainable in a blade former.
The pulsating nature of dewatering in blade formers was however empha-
sized by Norman lM, the deflection of the wires over the individual blades
being the basis for the generation of the dewatering forces F according to the
equation :

F = 2T sin al2 (13)

where T is the wire tension and a wire deflection angle over a blade tip . It was
suggested that the pulsating nature of the dewatering pressure creates internal
shear in the fibre suspension between the wires, and thus improves the mass
formation of the sheet produced.
This basic principle has later been generally accepted. The shape of the
pressure pulse generated by the blade force F depends on several factors, such
as wire speed, wire separation and wire stiffness. No attempts to describe the
detailed pressure event mathematically has been made as yet.
Measurement of the shape of the pressure pulses has been undertaken by
Beloit, using a static pressure probe extending from within the headbox into the
fibre suspension between the wires. Examples of pressure measurements are
shown in Figs. 28 and 29, 1( 21) .
There is a general tendency that, compared with roll formers with uniform
dewatering pressure, blade formers with pulsating dewatering pressure produce
sheets with superior mass formation but with lower retention of fines and filler.
The fibre orientation anisotropy cannot be manipulated to the same degree
using the speed difference between jet and wire as variable in a blade former as
in a roll former. This is because the pressure pulses along the forming zone all
affect fibre orientation, and only the conditions in the very beginning of the
forming zone can be controlled by the jet speed . As a rule, a blade former gives
a higher minimum anisotropy than a roll former.
 127

Fig . 28: Pressure pulses, Bel Baie blades. 48 glm2, T = 4.4 kNlm
Left: 460 mlmin . Right : 920 mlmin.

Fig. 29: Pressure pulses, Bel Baie, 48 glm2, 920 mlmin, T = 7 kNlrn
Left: Standard blades Right: Vacuum shoe
12 8

Constant and pulsating dewatering

Recently combinations of roll dewatering and blade dewatering have also been
utilized.
In a blade former, the application of a forming roll during the later part of
the dewatering adds an even pressure that improves retention without any loss in
mass formation (122).
Starting with partial dewatering in a roll former, one thin web is formed
on each wire. If pressure pulses from deflector blades are subsequently applied,
shear forces can be introduced which may disperse the fibre suspension between
the webs. In a roll former, the introduction of blades after the roll has been
found to improve the mass formation, and operating experience with this design
is now documented (12 ).

Controlled pressure pulses

One drawback with all the twin-wire formers described above is that the only
on-line variable available to control dewatering conditions is wire tension . By
changing the wire tension. the amplitude of the complete pressure event can be
controlled. It is however not possible to adjust local dewatering pressure accord-
ing to specific requirements.
Recently a new method has been introduced, by which it is possible to change
the individual forces on stationary dewatering blades ), see Fig. 30.

Fig. 30.: Principle offlexible twin-wire dewatering system (L2J4 .

This means that there is freedom to set the desired dewatering pressure profile .
The new principle has been applied in boardmaking . It seems also to have a
large potential for high-speed papermaking.
 129'

7.3 .3. Hybrid forming

The hybrid former consists of a fourdrinier section for initial dewatering, and a
twin-wire section for final dewatering . For each of these parts, the principles
discussed above apply.
The twin-wire part of the hybrid former has one limitation compared with
an ordinary gap former : Since the fibre suspension is transported into the nip by
the bottom wire, the jet/wire speed ratio in the twin-wire nip is always equal to
unity. This causes some fibre alignment, and the minimum fibre orientation
anisotropy is therefore higher than that for a fourdrinier, or a roll former wet
end.

7.4. High consistency forming

As mentioned in section 5 .5 ., forming at 3-4% consistency differs from con-

ventional forming in that the fibres form a network already inside the headbox.
Dewatering in high consistency forming is, by definition, therefore a thickening
process . The resulting sheet density of a high consistency sheet is therefore
lower than that of a conventional sheet formed under filtration conditions.
It has been shown, for sheets made in a special high consistency labora-
tory sheet former, that the z-direction strength is more than doubled compared to
that of standard laboratory sheets, while the strength losses in the plane of the
sheet are only about 5% (79~.It was also found that ply bond between two high
consistency sheets was about twice as high as that for the corresponding
conventional sheets. This meant that the ply bond for high consistency sheets
was about the same as the z-direction strength of a conventional sheet.
The development of high consistency forming is mainly a matter of head-
box development . Some improvement in mass formation can however be
achieved by using dandy rolls or by dewatering in a twin-wire nip , 126 ).
Application of the Flexi-Former techniques to high consistency dewatering can
improve strength properties (127).
High consistency techniques have so far been industrially applied to form
the centre ply in board products, see further section 7.5.

7.5. Multi-ply forming

Multi-ply forming has traditionally been the dominating process for higher
grammage products, such as board, where single ply forming would involve de-
watering limitations . Multi-ply forming then also automatically offers the pos-
sibility of choosing different raw materials in the different plies.
Cyinder vats were the dominant equipment used for multi-ply board man-
ufacturing until the 1960's. Increased demands on the evenness of product
quality then opened up the field for different combinations of fourdriniers and/or
13 0

twin-wire arrangements .A recent summary of multi-ply forming concepts has

been made by White (12$) .
Lately, increased interest has also been paied to the manufacturing of low
grammage products, such as tissue and fine paper, using stratified headboxes .
In this overview, the term simultaneousforming will be used when all plies are
delivered from one headbox, while separateforming indicates the use of several
headboxes . These terms are preferred to the previously used terms multilayer
and multiply respectively, which do not seem to be clear enough.
For the forming principles of a single ply in a multi-ply product, reference
is made to the preceeding discussions within that area. In addition, ply bond
between the differeent plies has to be considered, a property which is often criti-
cally low in multi-ply products.
Ply bond is influenced by the amount of fine material at the interface be-
tween two plies, and is therefore highly influenced by the forming methods
used. As an example, the low fines content on the wire side of a fourdrinier
sheet can be critical if this side is facing another ply in a multi-ply design. The
problem is often accentuated if mechanical pulp is used as raw material in more
than one ply.
There are four basic ways of forming a two-ply product:

I Two plies are formed separately, and couched together.

II A second ply is formed on top of a first ply, and dewatering

is made away from the first ply .

III A second ply is formed on top of a first ply, and dewatering

is made through the first ply .

IV Both plies are formed simultaneously from a stratified headbox .

Generally, ply bond between the two plies will successively increase along this
list of forming principles.
As mentioned in section 7 .3, twin-wire forming at very low machine
speeds generally causes instability problems. However, the application of con-
trolled pressure pulses .in a twin-wire forming unit, permits the forming of a
high grammage ply even at speeds below 100 m/min (129) :
High consistency forming of the centre ply in a board product results in a
more bulky sheet and higher z-direction strength as well as ply bond compared
to conventional forming . Therefore, the greatest potential for high consistency
forming is as the centre ply in board products, since the positive effects then
include product properties as well as investment costs (1 Q).
 7.5.1. Simultaneous forming

In comparison with single-ply forming, multi-ply forming using a stratified

headbox offers the possibility to produce low grammage sheets with optimized
surface properties as well as centre ply characteristics. It has found a widespread
use in tissue manufacturing ( 131 , 62). Simultaneous forming has also been
applied to printing papers (63), and has been used since over ten years on a
Swedish linerboard machine .
The advantage of simultaneous compared to separate multi-ply forming of
low grammage products lies in reduced investment costs and improved
runnability . The basic problem with simultaneous forming is that the different
plies tend to mix to some degree, so that layer purity may suffer. Layer purity
can be defined from Fig . 31.

Fig 31: Fibre distribution in a rollformed FEX sheet with chemical

pulp in the surfaces and CTMP in the centre (1J2
.

Perfect layer purity would mean that the surface layers contain 100% and the
centre-ply of the sheet 0% chemical pulp respectively . In fig. 31, the centre of
the sheet contains approx . 10% chemical pulp, which can be interpreted as a
comparatively high layer purity.
In simultaneous forming it is usually not possible to completely separate
the white waters from the different furnishes . A common white water system
can mean considerable reductions in investment costs.It is then necessary to
achieve high retention levels, to minimize the deterioration of layer purity
132

through the white water system. Terland made a theoretical investigation of the
influence of the degree of white water separation, retention level and amount of
broke on layer purity in three-ply simultaneous forming 1(33) .
Terland also pointed out that even if layer purity is high, layer mass for-
mation may be unacceptable ( 132 ), which is illustrated iii Fig 32:

Fig 32 : Stratified sheet with with high layer purity, right,

but bad layer mass formation, left (_132) .

As already mentioned in section 5 .4, there are two basic types of stratified head-
boxes which can be used for simultaneous, multi-ply forming . Following either
of these headboxes, the twin wire dewatering can be performed with con;cant or
with pulsating dewatering pressure, or with a combination thereof. Probably
controlled pressure pulses will be a useful dewatering arrangement in this con-
text. The combination of headbox principle and dewatering pressure event re-
quired to achieve specified levels of layer purity and layer mass formation is an
important field for future investigations .
 133

8. Product properties .
The vast literature relating forming conditions to product properties cannot be
covered in this overview . Only two specific areas will be briefly discussed in
this section, firstly the relationship between mass formation and strength and
secondly fibre orientation anisotropy as a result of process conditions.
A recent literature review on paper quality aspects of twin-wire formed
paper in comparison to fourdrinier formed paper has been prepared by Sinkey
and Wahren (134).

8.1. Mass formation and mechanical properties .

It is well known that higher dilution of a fibre suspension before drainage re-
sults in better mass formation as well as higher paper strength. This can be ex-
plained by the self healing effect of dewatering on fibre distribution mentioned
in section 6.2.
Corte made experiments using a modified laboratory sheet former, by in-
creasing the height of the suspension tank so that the initial fibre consistency
could be decreased one order of magnitude from the standard value of 0.02% .
Also in this range he found improved product properties with increased dilution

Terland and Fellers {136) studied the effect of forming consistency on

sheet properties in twin-wire roll forming . They found a considerable influence
on tensile index and elongation to rupture, see Fig . 33, as well as burst, but
negligible effects on tensile stiffness and tear index.

Fig 33. : Influence offorming consistency in the FEX rollformer unit

on tensile index ("left) and elongation to rupture (right) (L&.

It has been claimed that multi-ply forming results in improved product properties
when compared to a single ply product . Terland and Fellers 137) showed that
134

this may be a matter of forming consistency . Single-ply and two-ply products

were formed on the FEX machine, and using the same forming consistency
there was no significant difference between the product properties . However,
two-ply forming made it possible to decrease the forming consistency in each
layer compared to single ply forming, with a resulting improvement in sheet
mass formation as well as in mechanical properties.
It has been demonstrated in several investigations that twin-wire blade
forming yields better mass formation than roll forming and thereby also better
mechanical properties, see e.g. (136). Further esults by Reiner demonstrated
that in a linerboard blade former, an improvement in mass formation as a result
of an increased blade pressure pulse had a negative effect on the burst value.
It has further been demonstrated that an improved mass formation through
eccessive retention aid addition has no effect on the strength properties of the
product 13 ).
It can thus be concluded that mass formation is not unambiguously related
to paper strength. The effect on product strength of an improved mass formation
is therefore determined by the what type of change in forming consitions gene-
rated the improvement.
Finally it should be pointed out that paper strength is not determined ex-
clusively by forming consistency, but also by other variables such as the degree
of refining of the furnish . In a practical application with given dewatering ca-
pacity, an optimization of the degree of refining and the forming consistency has
to be performed. This is demonstrated in Fig . 34, in which also the different re-
sponse to forming consistency for long fibre and short fibre pulps respectively
is demonstrated (Hallgren and Lindstrom 71) .

Fig. 34 : The effect of forming consistency and refining on the

tensile index-mass formation relationship in the FEX
roll former unit. Left: Softwood . Right: Hardwood (2J1 .
 135

8.2. Fibre orientation anisotropy

Lately, the concern about variations in fibre orientation and about fibre orien-
tation skewness has increased, partly because of increased product demands and
partly because of improved ways ofdetecting such conditions .
Anisotropic fibre orientation was noticed in the mix jet leaving the head-
box already by Moss and Bryant (_55), as mentioned in section 5.3.1 . The
higher the contraction in the headbox nozzle, the greater the fibre orientation .
The first investigation of the distribution of fibre orientation in the z-direc-
tion of a sheet was performed by Danielsen and Steenberg (129), who used po-
lar diagrams to describe the degree of orientation, see Fig 35.

Fig . 35: A polar diagram offibre orientation inpaper (12J9

Left: Top side ofpaper, Right: Wire side ofpaper.

During dewatering, longitudinal shear at the interface between wire/web and mix
suspension introduces a "combing" effect that tends to increase the fibre orienta-
tion. The orientation is stronger on the wire side of a sheet, since this side is
formed first and a relative motion between wire/web and mix gradually decrea-
ses as the dewatering process proceeds.
It is well documented that in fourdrinier forming, a difference between jet
and wire velocities affects fibre orientation as well as mass formation.When the
velocity difference deviates from zero by small amounts, mass formation im
proves. This could be caused by the breaking up of fibre flocs through shear,
but more likely be a result of the supply of extra fibres for local low grammage
areas through the relative movement of fibre suspension in the plane of the
sheet.
At larger velocity differences, fibre orientation will increase, but distur-
bing effects will cause mass formation to deteriorate .
A recent study of the influence ofjet to wire velocity difference on sheet
properties was made by Bubik and Kleppe (14Q). They demonstrated that the
degree of anisotropy is affected by the forming consistency, see Fig . 36. This
can be explained by the orientation effects in the headbox nozzle . The higher the
136

forming consistency, the smaller the slice opening, the - larger the nozzle
contraction and thus the greater the orientation effects .

Fig . 36: Influence ofthe difference of the ofjet and wire velocity on
tensile ratio MD/CD at different forming consistencies ( 140) .

In a twin-wire roll former, the influence ofjet to wire velocity difference is simi-
lar to that of the fourdrinier machine . It should be pointed out, however, that the
relevant jet velocity is that after the initial deceleration in the twin-wire nip, see
e.g. eq. (10).
In a twin-wire blade former, the possibilities to control fibre orientation
with the jet to wire velocity difference is limited 6. The reason for this is the pul-
sating nature of the dewatering pressure. Since local shear in the fibre suspen-
sion is inevitably introduced by every pressure pulse, the local shear intduced by
the initial jet velocity is relatively seen less important than in the roll former case.
Niskanen QA1) has studied the effect on fibre orientation of velocity dif-
ferences between jet and wire, and later during this symposium his experimental
investigation and a theoretical model will be presented.
Hasuike et. al. (142 ) studied the correlation between local distribution of
fibre orientation and local grammage using X-ray techniques. They found that
fibre orientation in the heavy areas was more isotropic compared to that in the
low grammage areas, see Fig . 37 . This is natural if the heavy areas are mainly
built up from fibre flocs.

6 Unpublished FEX results .

 13 7

Fig . 37: Fibre orientation in high grammage and

low grammage areas respectively ( 142 ) .

It should be pointed out that when the anisotropic properties of a sheet are
discussed, besides the effect of headbox and dewatering also the increase in
anisotropy in the free draw after the press section and the final increase in
anisotropy by tension during drying have to be taken into consideration.

8.2 .1. Skew fibre orientation

So far, symmetric fibre orientation with respect to the machine direction has
been assumed in this overview. Skew fibre orientation often occurs, and can be
caused by a headbox jet velocity vector not in line with the machine direction or
by sideways flow generated in the wire sction .
Holik and Weisshuhn ( 143 ) demonstrated that even with an angular devia-
tion between jet and machine directions as small as a half degree, a considerable
cross flov velocity arises on a modern high speed paper machine. The sum of
this cross flow velocity component and the difference between machine speed
and jet MD-speed form a velocity vector representing the relative suspension
movement on the wire .
Holik and Weisshuhn pointed out that local variations in the angle of this
relative velocity is largest when the MD-component is low, and that the angle of
skewness will even change sign with a change from excess jet velocity to deficit
jet velocity, see Fig . 38 . They further exemplified the impact of imperfections in
headbox design on jet skewness .
13 8

Fig. 38: The relative velocity urel as afunction ofthe difference

between jet and wire speeds, uj - uw and cross flow ucross*
Left: uj - uw > 0, Centere : uj - uw = 0, Right : uj - uw > 0

As already mentioned, Westmayer (61) has calculated the effect of headbox

nozzle shape on the amount of sideways flow caused by a local change of slice
opening . A low angle headbox nozzle can be over four times as sensitive as a
headbox with vertical front wall in this respect.
To completely avoid sideways flow from a headbox, the cross machine
distribution must be perfect and the slice opening constant across the whole ma-
chine width, see further section 5.3.2 .

9. Acknowledgements
The author is indebted to Dr Anthony Bristow for the linguistic revision of the
material.
Ping, Anna-Karin, Qlof and Brian volounteered to spend the Saturday
evening and Sunday night immediately before my flight to Cambridge, prepa-
ring the 350 copies of this overview required at the symposium. This help is
gratefully acknowledged .
Finally, I would like to thank the responsible persons within FRC for not
striking out this contribution when deadlines were not met, and instead allowing
its presentation in this third volume of the proceedings .
 13 9

10. References
I "Formation and Structure ofPaper" Trans of the 2nd Fundamental
Research Symposium in Oxford, 1961. Francis Bolam, Ed. Techn Sec .
British Paper and Board Makers Assoc., London (1962) .

II "Consolidation of the paper web" Trans. of the Symposium held at

Cambridge, 1965. Francis Bolam, Ed. Techn Sec . British Paper and
Board Makers Assoc., London (1966) .

III "The Fundamental Properties of Paper Related to its Uses", Trans. of the
Symposium held at Cambridge, 1973. Francis Bolam, Ed. Techn . Sect .
BPBIF, London (1976) .

IV "Fibre - Water Interactions in Paper-Making". Trans. of the Symposium

held at Oxford, 1977 . FRC Ed. Techn.Div. BPBIF, London (1978).

V "The Role of Fundamental Research in Paper Making" . Trans. of the

Symposium held at Cambridge, 1981. James Brander, Ed. Mech.Eng.
Publ.Ltd., London (1983).

VI "Papermaking Raw Materials" . Trans . of the 8th Fundamental Research

Symposium held at Oxford, 1985. Vaughan Punton, Ed. Mech. Eng .
Publ. Ltd., London (1985).

VII "Fundamentals of Papermaking". Trans. of the 9th Fundamental

Research Symposium held at Cambridge, 1989. C. Baker and W.
Punton Ed. Mech.Eng .Publ .Ltd., London (1989) .

1 Wrist, P.: Dynamics of Sheet Formation on the Fourdrinier Machine.

In Ref I, 839.

2 Parker, J.: The Sheet-Forming Process, TAPPI STAP no 9 (1972) ..

3 Radvan, B. : Forming . Papermaking Science, Elsevier, 176 (1976) .

4 Lindstr6m, T. : Some fundamental chemical aspects on paper forming .

In Ref VII, 309.

5 Wahren, D.: Proposed definitions of some basic papermaking terms . Sv.

Papperstidn . 70 (21), 725 (1967) .
140

6 Corte, H. : Ober die Verteilung der Massendichte in Papier. Das Papier

24 (5), 261 (1970) .

7 Burkhard,G ., Wrist,P .E. and Mounce,G .R. : A Formation Meter Which

Graphically Records Paper Formation. Pulp and Paper Mag. of Canada
61 (6), T139 (1960) .

8 Daunais, R. and Garner, R. : The NUI formation tester - an evaluation,

modification and comparison with other techniques. Int . Paper Physics
Conf, 43 (1987).

9 Kallmes, O. and Ayer, J.: Light scanning system provides qualitative

formation measurement . Int. Paper Physics Conf., 209 (1987).

10 Papworth, S.: The application of image analysis to evaluate optical

properties of paper. 6th Scand. Conf. on Image Analysis, 808 (1989) .

11 Norman, B. and Wahren, D.: Mass distribution and sheet properties. In

Ref III, 7.

12 Komppa,A. and Ebeling,K . : Correlation between the Areal Mass and

Optical Densities of Paper In Ref V, 603.

13 Norman, B. and Wahren, D.: The measurement of mass distribution in

paper sheets using a beta radiographic method . Sv. Papperstidn 77 (11),
397 (1974) .
14 Cresson, T. : The Sensing, Analysis and Simulation of Paper Formation .
Thesis, Syracuse, New York (1988) .

15 Paper Structure and Properties . Anthony Bristow and Petter Kolseth

Ed., Marcel Dekker, New York (1986) .

16 Kajanto, Komppa,Ritala : How formation should be measured and

characterized. Nordic Pulp and Paper Research Journal, 4 (3) (1989) .

17 Yuhara, T., Hasuike, M. and Murakami, K. : Application of Image

Process Technique for Analysis of Sheet Structure of Paper.
Japan Tappi 61 (1), 85 (1986).

18 Tomimasu, H., Luner, P. and Hanna, R. : Rapid imaging of mass

distribution in paper by electron beams . In Ref VII, 159 .
19 Norman, B and Axell, O. : A Method for Measurement of Chalk
Distribution in Paper Sheets . In Ref. VI, 577.

20 Taylor, G.I. : The Spectrum of Turbulence . Proc . Roy . Soc. Lond., A

164, 476 (1938) .

21 Norman, B. and Wahren, D. : A comprehensive method for the

description of mass distribution in sheets and flocculation and turbulence
in suspnsions. Sv. Papperstidn . 75 (), 807 (1972) .

22 Haglund, L, Norman, B. and Wahren, D. : "Mass Distribution in

Random Sheets . Theoretical Evaluation and Comparison with Real
Sheets . Sv. Papperstidn .77 (10), 362 (1974).

23 Taylor, G.I. : Statistical theory of turbulence. Proc.Royal .Soc. London .

A151 , 421 (1935) .

24 Jordan, B .D. and Nguyen, N.G.: Specific perimeter . A graininess

parameter for formation and print-mottle textures" Paperi ja Puu 68 (6-
7): 476, 1986.

25 Duffy, G., Ek, R., M611er, K. and Norman, B .: Hydrodynamics of

papermaking fibres in water suspensions . In Ref IV, 195 .

26 Kerekes, R., Soszynski, R. and Tam Doo, P.: The Flocculation of Pulp
Fibres . In ref VI, 265 .

27 Kerekes, R. : Pulp flocculation in decaying turbulence : A literature

review.Journ . of Pulp and Paper Sci . 9 (3), TR 86 (1983).

28 Wahren, D. : Fibre Network Structures in Papermaking Operations . The

Cutting Edge Conf., Inst of Paper Chemistry, 112 (1979) .

29 Steenberg, B, Thalen, N. and Wahren, D. : Formation and properties of

fibre networks. In Ref 11, 177 .

30 Wahren, D. and Meyer, R.: On the Elastic Properties of Three-

Dimensional Fibre Networks . Sv Papperstidning, 67 (10), 432 (1964).

31 Soszynski, R . and Kerekes, R.: Elastic interlocking of nylon fibres

suspended in liquid . Part 1 . Nature of cohesion among fibres . Nordic
Pulp & Paper Research Journal, 3 (4), 172 (1988)

32 Jacquelin, G. : Prepared contribution . In Ref. II, 299 .

14 2

33 Soszynski, R. and Kerekes, R.: Elastic interlocking of nylon fibres

suspended in liquid. Part 2. Process of interlcking. Nordic Pulp & Paper
Research Journal, 3 (4), 180 (1988).

34 Riitala, R. and Huiku, M.: Scaling, percolation and network theories:

new insights into papermaking. In Ref VII, 195 .

35 Onabe F. and Sakurai, K.: Applications of flow visualisation technique

in wet end chemsitry . In ref VII, 219.

36 Wagle, D ., Lee, C. and Broadkey, R.: Further coments on a visual

study of pulp floc dispersion mechanisms. Tapi J. 71 (8), 137 (1988).

37 Steen, M.: The application of refractive index matching for two-phase

flow measurements in turbulent fiber suspensions by laser doppler
anemometry . Nordic Pulp & PaperRes. Journ. 4 (4), (1989).

38 Steen, M.: On turbulence structure in vertical pipe flow of fiber

suspensions . Nordic Pulp & Paper Res Journ. 4 (4), (1989).

39 Norman, B. and Tegengren A. : Mixing of thich stock and white water .

XXIII Eucepa Conf., 51 (1988).

40 Norman, B ., Wedin, P-O ., Grundstr6m, K-J . and Hagen, N.:

Visualization of flow in a high consistency screen . TAPPI Eng . Conf.,
585 (1986) .

41 Goldenburg, P., Fredriksson, B. and Leblanc, P. : The Beloit S-screen -

a new concept for high consistency screening. New Available
Technique, SPCI, 1063 (1987).

42 Carlsson, Bengt and Gistren, Henrik.: A new design for high

consistency cleaning. Graduate work, Royal Inst. of Techn . (1989)

43 R6ding, S., and Norman, B. : FEX the STFI experimental paper

machine - design of white water system . Tappi Eng. Conf., 449 (1985) .

44 Waller, M. : Recent developments in headboxes . Tappi J. 701 (l), 33

(1987).

45 Trufitt, A. : Design aspects of manifold type flowspreaders . TAPPI Pulp

and Paper Technology Series, 32 (1975).
 14 3

46 Syrjala, S., Saarenrinne, P. and Karvinen, R.: Fluid dynamics of

tapered manifold flow spreader. TAPPI Engineering Conference, 233
(1988).

47 Bubik, A. and Christ, A. : Step diffuser - the hydraulics of Escher Wyss

headboxes . Tappi Eng . Conf., A 149 (1976) .

48 Lin, J.: A numericla study of turbulent flow through a circular step

diffuser. TAPPI Eng . Conf., 21 (1988) .

49 Mardon, J. : Discussion. In Ref . I, 889 (1961) .

50 Reiner, L., and Wahren, D.: The Characterization of Flow from the
Headbox of a Paper Machine. S vensk Papperstidn, 74 (8), 225 (1971).

51 Parker, J. and Hergert, R. : Simultaneous convergence - a new concept

in headbox design . Tappi 51. (10), 425 (1968).

52 Sanford L.: Characteristics of flow from a two-dimensional nozzle with

multiple-tube inlet . Tappi Eng . Conf, A 159 (1976) .

53 Wolf, K. and Egelhof, D.: The W-type High-Turbulence Headbox - its

development concept and basic arrangement . Paper Techn. if (2), 77
(1974).

54 Ilmoniemi, E., Aroviita, L. and Yli-Kokko, E. : The role of headbox

turbulence generators in grammage uniformity. Appita 39 (1), 36
(1986).

55 Moss, L and Bryant,E.: Photographic study of the Motion of Fibres and

Water in Flowing Fiber Suspensions . Paper Trade J., 106 (April 14)
Tappi Section, 226 (1938).

56 Jones, G. and Ginnow, R. : Modeling headbox performance with

computational fluid mechanics . TAPPI Eng . Conf., 15 (1988).

57 Appel, D. and Yu, Y.: Free-streamline analysis of flow from nozzles,

flow through side inlets and flow past corners. Studies in Engineering
Mechanics, No 17, Center for Research in Engineering Science,
University of Kansas, Lawrence (1963) .

58 For the results in ref. (57), see alternatively : TAPPI TIS 014-2,3,4.
14 4

59 Kerekes, R. and Koller, E.: Equations for calculating headbox jet

contraction and angle of outflow . Tappi 64 (1), 95 (1981), with
correction in 64 (4), 104 (1981).

60 Karlsson, H. and Eriksson, L. : Controllability of Paper Making .

In Ref VII .

61 Westermeyer, W.N.: Modellierung der Stromung in der

Stoffauflaufduse and auf dem Langsieb im Hinblick auf das
Flachengewichts-Querprofil der Papierbahn . Das Paper 41 (11), 591
(1987)

62 Linkletter, M. : Tissue layering . PIRA Conf. : New Technologies in

Multiply & Multilayer Structures, paper 6 (1987) .

63 Stenberg, G.: A new headbox for multilayered printing paper. Tappi J.

70 (9), 49 (1987).

64 Swedish pat. 355 .615

65 Swedish pat. 385 .029

66 Swedish pat. 412.078

67 Nomura, T., Wada, K. and Shimizu, T. : High consistency sheet

forming. 1 : Research and development of headboxes .Tappi J. 72 (1),
115 (1989).

68 Radvan, B., Dodson, D. and Skold, C.G. : Detection and cause of the
layered structure of paper. In Ref II, 189 .

69 Mohlin, U.-B. : Mechanical pulp properties - the importance of fines

retention. Sv. Papperstidn . 80 (3), 84 (1977).

70 Sauret, G., Trinh, H. and Lefebvre, G.: Experiments to reproduce

multi-ply boards in the laboratory (in German) Das Papier 23 (1), 8
(1969)

71 Hallgren, H. and Lindstrom, T.: The influence of stock preparation on

paper forming efficiency . XXIII EUCEPA Conf., 35 (1988).

72 Corte, H. and Kallmes, 0.: Statistical geometry of a fibrous netwok. In

Ref I, 13.
 145

73 Gorres, J., Cresson, T . and Luner, P.: Sheet formation from flocculated
structures. Journ of Pulp and Paper Science 15 (2), J55 (1989) .

74 Steen, M.: A concept for fibre flocculation in turbulent flow.

In Ref VII, 251 .

75 Ingmanson, W.L. and Andrews, B.D.: High-Velocity Water Flow

Through Fiber Mats. Tappi 46 (3), 150 (1963).

76 Pires, E., Springer, A. and Kumar, V.:Computational model for water

drainage in fouidlnier paper machines. Tappi J. 71 (4), 133 (1988) .

77 Britt, K.: Mechanisms of Retention During Paper Formation. Tappi 56

(10), 47 (1973).

78 Wahlstr6m, B. and O'Blenes, G. : The Drainage of Pulps at Paper-

Making Rates and Consistencies Using a New Drainage Tester. Pulp and
Paper Mag of Canada, 63 (8), T405 (1962).

79 Grundstr6m, K.-J., Hartog, S., Norman, B. and Reiner, L. : Basic

principles of high consistency forming and the mechanical properties of
felted structures. Int . Paper Physics Conf, 11 8 (1975) .

80 Britt, K., Unbehend, J. and Shridharan, R.: Observations on water

removal in papermaking . Tappi J. 69 (7), 76 (1986) .

81 Britt, W. and Unbehend,J.: Water removal during paper formation .

Tappi . J. 68 (4), 105 (1985) .

82 Torbjbrn Helle: Paper forming wires over 75 years . 75th Annual

Meeting, Technical Section CPPA, B 285 (1989)

83 Batty, R.: A mathematical model to aid in the design of paper machine

forming fabrics . Journ of Pulp and Paper Sci. 9 (1), T25 .(1983) .

84 Helle, T. The influence of wire structure on sheet forming. Paper Techn

& Ind . 21 (4), 123 (1980) .

85 Beran, R.: The evaluation and selection of forming fabrics . Tappi J. 62

(4), 39 (1979) .

86 Helle, T.: Analysis of wire mark in printing paper. Journ of Pulp &
Paper Science 14 (4), 91 (1988) .
14 6

87 Lewyta, J.: Recent developments in paper machine clothing: forming

fabrics . Tappi J. 70 (2), 39 (1987) .

88 Helle, T. : On the evaluation of the forming characteristics of

papermaking wires, International Water Removal Symposium, 93
(1982).

89 Johnson, D.: Retention & drainage of forming wires . International Water

Removal Symposium, 67 (1982) .

90 Nordiskafilt . Unpublished FEX results (1986).

91 Hansen, V. : Draining and forming through the years . 75th Annual

Meeting, Technical Section CPPA, B 63 (1989) .

92 Wahren, D. : Fourdrinier jet geometry. Nordic Pulp and Paper Res.

Journ. 3 (special issue), 29 (1988).

93 Osterberg, L. : Initial dewatering . The Ekman Days (abstract only)

(1988).

94 Parker, J. and Epton J. : Wet end barring . Tappi 60 (10), 90 (1977).

95 Moen, C. : Basis weight barring. Tappi 60 (10), 116 (1977).

96 Kallmes, O. and Langdok, R. : Initial operation trials of a dandy roll

former. Tappi J. 71 (4), 74 (1988).

97 Wrist, P.: The Papermaking Process as a Filtration Problem. Pulp Paper

Mag Can. .55 (6), 115 (1954) .

98 Taylor, G.I.: Drainage at a Table Roll. Pulp Paper Mag Can 57 (3), 267
(1956).

99 Taylor, G.I. : Drainage at a Table Roll and a Foil. Pulp Paper Mag Can
59 (C), 172 (1956) .

100 Bergstrom, J. : Table Roll Drainage . Sv Papperstidn 600 (1), 1 (1957).

101 Kufferath, W., Kallmes, O. and Steffen, H.-R. : Die Sonic Roll,
Das Paper, 42 (l0A), V141 (1988).
 147

102 Kallmes, a., Marinari, G. and Perez, M.: Forming a sheet on a fabric
vibrating at or near its resonant frequency. Tappi Papermakers
Conference, 153 (1989) .

103 Schmid,W .: Stroboskopische Untersuchungen in der Blattbildungszone .

Wochenblatt fur Papierfabrikation, 109 (11/12), 371 (1981) .

104 Cadieux, S.: A mathematical drainage model for Fourdrinier paper

machines. Joum of Pulp & Paper Sci . 9 (4), TR 111 (1983) .

105 Kufferath, W., Kallmes, 0. and S teffen, R. : The Cascade-Foil .

Das Paper 41 (l0A), V136 (1987) .

106 Reiner, L.: The Flexforming concept .

Journ Pulp & Paper Sci . 10.(2), J30 (1984) .

107 Norman, B. : Principles of twin-wire forming.

Sv. Pappersteidning, 82 (11), 330 (1979) .

108 Attwood, B. and Lawrence, L.: Inverform - An account of the

experimental work during the development of a new method ofpaper
and paperboard manufacture. Paper Techn. 1 (5), T195 (1960) .

109 de Montigny, R., Pye, I. and Hedley, T.: The Papriformer (1) - the
machine and its performance . Pulp Paper Mag. Can. 68 (10), T482
(1967)

110 Atkinson, W. and Malashenko, A.: Commercial development ofthe

Papriformer. Tappi, 54 (9), 1464 (1971).

111 Gustavsson, D. and Dennis, L. : The "Bel-Baie" twin wire former.

Tappi, 53 (3), 483 (1970) .

112 Stewart, R. and Howarth, J. : Verti-Forma versus fourdrinier. Pulp

Paper 5 (13), 70 (1971) .

113 Mauranen, P. and Hujala, J.: Valmets former for paper machine rebuilds
and new paper machines. CPPA Ann. Mtg ., B 195 (1983).

114 Malashenko, A., Deperis, G. and Lindstr6m, R.: Dominion

Dynaformer. 68th CPPA Ann Meeting, A139 (1982).

115 Thorp, B.: Fundamental and commercial overview of twin wire and
multiple wire forming. Tappi Twin Wire seminar, 1 (1989)
14 8

116 Stenberg, G. : The KMW - Papriformer. KMW Sales Conference,

Karlstad (1971) .

117 Hergert, R. and Sanford, L.: Pressure measurements in the forming

zone of a twin-wire tissue machine . International Water Removal
Symposium, 47 (1982)

118 Ingemarsson, G.: Mathematical modelling to explain roll former

behaviour . International Water Removal Symposium, 73 (1982)

119 Norman, B.: On the mechanisms of dewatering in twin-wire and press

sections. Nordic Pulp & Paper Res Journ, Especial issue), 39 (1987)

120 Gustavsson, D.: Water removal from the Bel-Baie former. International
Water Removal Symposium, 1:234 (1975) .

121 Nash, J.: Twin wire forming development . Tappi Twin Wire Seminar,
79 (1989)

122 Hergert, R. and Harwood, J. : Retention, drainage and formation on high

speed formers . Tappi Papermakers Conf., 271 (1988) .

123 Odell, M.: Drainage Mechanism ofa Roll and Blade Former. Tappi Eng.
Conf. (1989).

124 Baumann, W.-D. : Duoformer D - A new approach to top wire forming .

Tappi Twin Wire Seminar, 115 (1989) .

125 Blomquist, B. : Alform-Technologies for Multilayer Board. XXIII

EUCEPA Conf., 187 (1988).

126 Nomura, T., Wada, K. and Shimizu, T.: High consistency sheet
forming. 2: Pilot plant tests . Tappi J. 72(4), 171 (1989).

127 Reiner, L.: Flexforming (abstract only), The Ekman days (1988).

128 White, C.: The concept of multiply & multilayer paper & board
products. PIRA Conf: New Technologies in Multiply & Multilayer
Structures, paper 3 (1987).

129 Meinecke, A. : The Duoformer (D) - Concept and Operating Results .

XXIII EUCEPA Conf., 97 (1988) .
 Transcription of Discussion

OVERVIEW OF THE PHYSICS OF

FORMING
B . Norman (Review Paper)

Dr . C . Dunning, James River

I wonder if you can comment on how rapidly flocs form in the

headbox . Let's presume for a moment that we have some means of
destroying any flocculation at the entrance to the headbox . How
rapidly would flocs form between the tube bank and the slice
opening?

Prof . B . Norman

I know you can find figures in the literature on this subject,

however I am reluctant to rely on them . For instance, Otto Kallmes
presented a Table with figures showing how quickly flocs will
reform at different consistency levels . In that paper he referred
to Douglas Wahren . When I discussed this with Douglas, he did not
agree with the interpretation put on his figures . His results were
very specific and generalisations are not necessarily valid .
Consistency level, degree of turbulence and geometry as a whole
are variables which must be considered in reflocculation .
Generally, there is adequate time for reflocculation . If you
destroy the flocs at an early stage they can definitely re-appear
later on . To break down fibre flocs you have to introduce
turbulent energy and when that energy decays flocs re-form . This
area will be explored further in tomorrow's papers .

Dr . H . Paulapuro, Finnish Pulp & Paper Research Institute

You did not mention retention at all in your review . We know that
retention is related to the chemistry of the papermaking process
but there are many physical effects also . Is there any specific
reason why you did not mention this?
Prof . B. Norman

In this presentation I did not have the time, it is mentioned in

the paper . It is impossible to make a good review without
mentioning paper chemistry . One has to be careful when discussing
retention as it must first be defined. In this paper what is
referred to as first pass retention is a function of the ratio
between the recirculated white water consistency and the headbox
consistency. This is the term used in 99% of mills . This is useful
only when considering one system where the white water
distribution is kept constant . . If you try to compare different
machines it will be invalid because the only thing that is
considered is : what is going through the short circulation . There
are examples of newsprint mills where half the fines in the thick
stock are recirculated via the long recirculation system . It is not
therefore correct to have a retention value which only considers
what is going round in the short recirculation system . The major
problem with the long circulation is one of time to equilibrium .
We have simulated this and were also able to do measurements on
a papermill together with IVL in Sweden . We changed the refining
in a liner mill and wanted to see how long it would take the extra
dissolved material to reach equilibrium . It took 48 hours for the
long circulation . I am merely pointing out the complexities of the
system and its bearing on retention . Details of retention will be
covered in the Chemical Section later this week .

Dr . H. Paulapuro

May I continue with another question - could I have your opinion

on the application of high consistency forming to low basis weight
grades eg . below 100 g/m2 .

Prof . B. Norman

If you design a new process it is much cheaper to choose a high

consistency solution . Many components are not required at all and
others are much smaller . HC forming will give some advantage for
the product . Thus the advantage of using this technique in a board
product comes from both capital cost reduction and product
improvement . It is hoped that HC application will continue to
lower grammages but the drive is not then as great as for board
products . I think a more relevant question is on stratified
production of low grammage papers . This is more important than
high consistency forming of them .
Dr . G .A . Baum, James River

You implied that modern high speed headboxes with a jet to wire
velocity ratio near to one could lead to instabilities in terms
of the fibre orientation distribution . Could you comment on the
contribution of automatic slice adjustments to that problem?

Prof . B . Norman

In devices with automatic slice flow adjustment you have to

consider the influence of cross flow on the skewness of
anisotropy, which has not been done so far . You cannot just
control the grammage profile to what you want .

Dr . A . Ibraham, Papyrus, USA

I would like to understand more about your comments on contraction

in the nozzle . If I understand you correctly are you in favour of
reducing turbulence by increasing the contraction coefficient?

Prof . B . Norman

A higher nozzle contraction will mean a lower degree of relative

turbulence which is an advantage in paper forming . It is
particularly advantageous to have a perfect jet when you are
considering twin wire forming .

Dr . A . Ibraham

In that case we are going to meet high L/b ratio . The jet is going
to travel further down through the wire and it is going to
flocculate further as shown by Kerekes . I am really against that
concept . High L/b and low angle is not good for formation .

Prof . B . Norman

I am referring to the contraction within the headbox nozzle not

the impact onto the wire . It is a function of flow velocity
entering the nozzle and flow velocity leaving the nozzle . What you
refer to is a separate topic .

Dr . M .B . Lyne, International Paper, USA

In your talk you were describing the sources of variation in

orientation angle across the papermachine . Would you care to offer
an explanation or a model for variations in orientation angle in
the Z direction during forming? The difference in fibre
orientation angle between the top and wire side of a sheet is an
important factor in the complex curl of paper sheets .

Prof . B . Norman

The jet to web speed difference in a given CD position will

gradually decrease as dewatering proceeds . The original cross flow
component in the same CD position will, by definition, move
sideways during dewatering, and thus successively influence other
CD positions .

Dr . M . MacGregor, KTH (Voith), Sweden

Is the non-uniform shrinkage which occurs during drying and the

need to have uniform basis weight possible to optimise, or does
it lead to compromise situation?

Prof . B . Norman

It is theoretically impossible to have both uniform orientation

and grammage in the cross direction if you have non uniform cross
machine shrinkage .There is a paper later this week which says how
you can restrain the paper to avoid shrinkage in the cross
direction during drying

Dr . M . MacGregor

If basis weight control is by flow redirection, then I agree -

which it is in most cases .

Prof . B . Norman

Unless you control local consistency in the headbox which is a

theoretical solution . When you control grammage with slice opening
the aforementioned problems are always present .

Dr . M . MacGregor

So in general the papermaker does not realise that he is giving

up many things to achieve a nice flat reel .

Dr . T . Uesaka, PPRIC, Canada

In one of your slides you mentioned the tensile ratio between MD

and CD as a function of jet speed and wire speed ratio - I
understand this relationship looks very unique when we keep the
contraction ratio constant . I am wondering if we increase the wire
speed by keeping the jet speed and wire speed difference constant
what will happen to that relationship . We have already experienced
in many high speed machines that we have a difficulty in
controlling the tensile ratio by changing the jet speed/wire speed
difference in contrast to the older Fourdrinier machines .

Prof . B . Norman

What do you mean by high speed machine? Is it a Bel-Bale?

Dr . T . Uesaka

Yes, in the case of Bel-Baie for example .

Prof . B . Norman

This is another project we are working on . In the case of the

Bel-Baie you haven't the same degree of control as you have on a
Fourdrinier or Roll-former because all the pressure pulses during
dewatering will of f ect the fibre orientation . You can only control
the jet conditions as it enters between the two wires .

Dr . R . Ritala, Finnish Pulp & Paper Research Institute

I would like to ask about fibre flocculation . The Wahren-Meyer

formula has been derived so that we have three contact points on
the average for each fibre . Why should this have any significance
as it is on average, I can understand why we need three points to
build flocs but why do we need three points on the average?

Prof . B . Norman

I refer you to Dr . Wahren .

Dr . R . Ritala

I would like to offer an alternative explanation . There is a

coincidence in that the threshold to have a connected network of
fibres (That can be shown to be at a consistency which is simply
the inverse of the fibre aspect ratio) in fact fits the data in
your Fig .3 a bit better in the region from 80 - 200 in the aspect
ratio .
Prof . B . Norman

Yesterday you mentioned this new reference from a physics paper

to me . This will be interesting to see . It would simplify things
if the sediment consistency is equal to the fibre diameter divided
by the fibre length .