Preferred citation: B. Norman and D. Söderberg. Overview of Forming Literature, 1990-2000.

In
The science of papermaking, Trans. of the XIIth Fund. Res. Symp. Oxford, 2001, (C.F. Baker, ed.),
pp 431–558, FRC, Manchester, 2018. DOI: 10.15376/frc.2001.1.431.

OVERVIEW OF FORMING
LITERATURE, 1990–2000

Bo Norman, KTH /STFI

Daniel Söderberg, STFI

ABSTRACT

Paper structure characterisation has been extended to include 2-D

formation and periodic marks as well as wavelet techniques. Local
fibre orientation anisotropy and misalignment angle analysis, in
combination with tape-splitting layering and image analysis have
improved the understanding of the relationship between paper
structure and properties as well as increased the possibilities to
trace the forming history from final paper structure.
The Crowding Factor has been introduced, to describe the ten-
dency of fibre suspensions to form fibre flocs. It gives an improved
predictability in comparison with fibre concentration, by also tak-
ing fibre slenderness into account. The dynamic characterisation
of de-flocculating and re-flocculating mechanisms is getting
increased attention, partly due to the potential of CCD-cameras
and image analysis techniques. It is the opinion of the authors
that floc stretching is a more promising way of de-flocculating
than turbulent shear. It is doubtful whether turbulence in simple
fluids can aim as a model for the flow irregularities in fibre
suspensions, due to the damping effects of fibres and flocs.
For headbox CD-profile control, dilution water injection has
been introduced. Since this makes slice lip deformations un-
needed, it has the potential to uncouple control of grammage and
misalignment angle profiles and also to improve CD-control
resolution.
The importance of headbox nozzle design for fibre orientation

12th Fundamental Research Symposium, Oxford, September 2001 431

B. Norman and D. Söderberg

anisotropy is now appreciated. A high nozzle contraction ratio

will increase anisotropy and also introduce some deflocculation
through floc stretching. Vane insertion helps to decrease fibre
orientation anisotropy and is also applied for multi-layer forming.
Although applied in tissue forming and in linerboard forming,
further improvements have to be made for printing paper
applications.
In twin-wire forming, the roll-blade principle has been accepted
by the main machine manufacturers. Since its introduction in the
STFI-Former in 1991, the blade section design with stationary
blades on one side and loadable blades on the opposite side is the
main design for printing paper applications, and also increasingly
used for different board qualities.

432 Session 3: Forming

 Overview of Forming Literature, 1990–2000

TABLE OF CONTENTS

1 Background 435
2 Paper structure characterisation 436
2.1 Power spectra 437
2.2 Specific perimeter 439
2.3 Wavelets 439
2.4 Formation 441
2.5 Medium-scale variations 446
2.6 Fibre orientation 447
3 Fibre suspensions 450
3.1 Fibre suspension modelling 450
3.2 Fibre flocculation evaluation 453
3.3 Velocity and turbulence measurements 459
4 Headboxes 462
4.1 Tapered header and CD control 462
4.2 Tube bank 465
4.3 Headbox nozzle 467
4.4 Headbox jet and streaks 480
5 Wire section designs 487
5.1 Roll-blade dewatering 487
5.2 Multi-ply board forming 496
6 Web forming 496
6.1 Random sheets 496
6.2 Self-healing effects 498
6.3 Formation improving mechanisms 501
6.4 Fibre deposition on wire 502
6.5 Wet web resistance 505
7 Dewatering processes 508
7.1 Jet impingement 508
7.2 Fourdrinier dewatering 508
7.3 Blade dewatering 510
7.4 Roll dewatering 517
7.5 Vacuum dewatering 521
8 Laboratory forming 523
8.1 Sheet forming 523
8.2 Pilot machines 528

12th Fundamental Research Symposium, Oxford, September 2001 433

B. Norman and D. Söderberg

9 Forming, structure and properties 529

9.1 Single-layer forming 530
9.2 Multi-layer forming 539
9.3 Multi-ply forming 541
10 Literature 541

434 Session 3: Forming

 Overview of Forming Literature, 1990–2000

1 BACKGROUND

The present overview can be considered as a continuation of the correspond-

ing review twelve years ago at the 9th Fundamental Research Symposium in
Cambridge, 1989 [1]. It is the intention here to give a comprehensive picture
of what R&D has taken place within the paper forming area since that time.
The developments since 1989 within the following areas will be discussed:
• Paper structure characterisation.
• Fibre suspensions.
• Headboxes.
• Wire section designs.
• Web forming.
• Dewatering processes.
• Laboratory forming.
• Forming, structure and properties.
At the 1989 Symposium, there was a separate overview dealing with paper
chemistry in forming, [2] but this will not be the case at the 2001 Symposium.
Even so, paper chemistry aspects on retention and flocculation will generally
not be included in this overview. With few exceptions, there will be no refer-
ences to forming work published before 1990. However, if a reference was
included in the 1989 overview, quotation may be made to the list of 143
literature citations included at that time [1].
Firstly, some terms, which often cause misunderstandings, are defined:
There is some confusion around the terms used to describe the process to
form a paper web as opposed to the small-scale basis weight variability in the
final product. Unfortunately, the term formation is frequently used in both
these senses. To avoid misunderstandings, it is necessary to use two distinctly
different terms. In this overview, the term forming denotes the overall process
of generating a paper web, while the term formation exclusively refers to
small-scale local basis weight variations in the final paper web.
A similar type of terminology dilemma exists regarding the term fibre
flocculation, which could mean either the state of flocculation or the process
of fibres entangling into a floc. At this time, we do however not suggest
different words for these two cases, which means that fibre flocculation will
still be an ambiguous expression. Hopefully, this will not cause too much
misunderstanding in the following text.
The term mix will be used to denote the diluted thick stock, with a fibre
concentration low enough to avoid excessive flocculation. The term “consist-
ency” will be avoided, since it alludes to the rather imprecise evaluation of
concentration traditionally performed by different kinds of “consistency”

12th Fundamental Research Symposium, Oxford, September 2001 435

B. Norman and D. Söderberg

meters – characterized by the fact that they never actually intend to directly
measure concentration.
The degree of fibre orientation anisotropy and fibre orientation misalignment
is increasingly important. Regarding terminology, fibre orientation to many
seems to specifically denote the misalignment angle. It is recommended that
anisotropy or misalignment respectively be added to avoid misunderstandings.
What used to be a wire now increasingly is denoted a forming fabric. This is
to emphasize that what was originally a two-shaft bronze material weave
is now a much more sophisticated design based on polymer threads. Cor-
respondingly, what used to be a felt is nowadays a press fabric. In this
overview the old, simpler terms will still be used to some extent. Thus, e.g.,
twin-wire, wire section, wire mark and felt mark will be the preferred terms. In
this context it does not seem quite compatible that a fabric generates a wire
mark. On the other hand, it could not generate a fabric mark, since a press
fabric might equally well have generated that.
The term roll-blade former will often be used for all designs where a roll is
followed by some kind of blade arrangement. In principle, this term could
thus also include the roll-adjustable blade design, in which the forming roll is
followed by a combination of fixed and movable blades on the two sides of
the wires.
The use of different material mixtures on different levels in a product as
now is increasingly applied to different products can be achieved using two
basically different principles. The tradition is to manufacture board in several
plies, while e.g., tissue products are increasingly formed in different layers.
The multi-layer products are formed using a single headbox, with different
component mixtures in different headbox layers, while a multi-ply product
consists of plies formed from separate headboxes.
For readers, not familiar with the basics of forming, it could be advisable
to initially study a textbook covering the area, such as Papermaking Part 1,
Stock Preparation and Wet End in the new book series Papermaking Science
and Technology [3].

2 PAPER STRUCTURE CHARACTERISATION

Variations in paper structure occur in a very wide range of scales. Different

mechanisms or different parts of the process are responsible for the initiation
of variations in the different scale ranges.
• Micro scale denotes the variations in a scale smaller than 0.1 mm. On this
scale particle size, micro flow, and colloidal interactions are important.

436 Session 3: Forming

 Overview of Forming Literature, 1990–2000

• Small scale covers the range of 0.1 to 40 mm, in which range fibre floccula-
tion and hydrodynamic conditions during forming are the main causes for
irregularity.
• Medium scale covers the range of 40 mm to 10 m and is affected mainly by
instabilities in headbox flow (partly generated in the approach flow system)
and wire section dewatering.
• Macro scale variations are those in excess of 10 m and are caused mainly by
the variability in the incoming thick stock. The variations are then mainly in
the machine direction. Such variations will not be treated in this overview.

2.1 Power spectra

To quantify the variations at different scales in a stochastic signal, the power
spectrum is often applied. Already during the 1930s, the autocorrelation and
its Fourier transform, the frequency power spectrum, were introduced, initially
to characterize the energy content within different frequency bands for elec-
trical signals. The power spectral density describes how the variance (square
of standard deviation) of a signal is distributed on different frequency ranges.
The frequency power spectrum has since been applied to characterize the
structure of turbulence for more than 60 years. The frequency power spec-
trum has also long been used to characterize variations of basis weight and
moisture at the dry end of the paper machine. An analysis of the distribution
of variance on different frequency ranges is a useful tool to trace the origin of
the variations.
Turbulence, flocculation, and formation represent stochastic variations in
local flow velocity, local fibre suspension concentration, and local basis
weight in the small-scale range, respectively. To describe such variations, it is
useful to use geometrical scale instead of frequency for the characterization.
This method is also applicable to medium and macro scale variations. Know-
ledge of geometrical size often simplifies the tracing of the origins for vari-
ations, such as wire lengths, felt lengths, and cylinder circumferences. The
frequency spectrum is therefore transformed into a wavelength spectrum. The
wavelength is calculated as the scanning speed (or flow velocity) divided by
the frequency. Since one complete sine wave can be interpreted as one nega-
tive and one positive floc, floc size will be half a wavelength. To conserve the
basic property of the power spectrum, i.e., to describe the distribution of
variance, also the frequency spectral density has to be transformed, into the
wavelength spectral density [1].
It is a common misunderstanding that the power spectrum is the Fourier
transform of the sample variations. This is incorrect, since it is instead a gen-
eralised Fourier transform, with principally different properties. In practical

12th Fundamental Research Symposium, Oxford, September 2001 437

B. Norman and D. Söderberg

terms, the difference between evaluation of a Fourier spectrum and a Power

spectrum respectively at a certain frequency could thus be described as
follows (the mathematics involved are much more complex).
• To evaluate the Fourier spectral density at one specific frequency, the
complete sample curve is correlated with an equally long sine curve and
this correlation is integrated while the sine curve is moved one wavelength
along the sample curve length.
• To evaluate the Power Spectral density, one sine wave period of a specific
frequency is correlated with the sample curve and this correlation is inte-
grated while the sine wave is moved along the whole sample curve length.
This is illustrated in Figure 1.

Figure 1 1-D sample variations (top), hypothetical waveform and wave displacement
for evaluation of Fourier spectrum (middle) and Power spectrum (bottom).

While the Fourier analysis primarily quantifies variations, which are

periodic (and thus in phase) over the whole sample length, the power spec-
trum analysis quantifies the total amount of sample variations of a certain
frequency, even if they do not occur in phase.
2-D power spectrum analysis is becoming increasingly used, often based on
optical quantification of variations using CCD-techniques. One example is
periodic wire marking, see further below.
Sometimes the absolute values of variations of a property are difficult to
evaluate due to calibration difficulties, such as in optical formation or fibre
suspension flocculation measurements, aiming to evaluate local basis weight
or local fibre concentration variations respectively. Scale information can
however still be obtained, by dividing the power spectrum area into two equal

438 Session 3: Forming

 Overview of Forming Literature, 1990–2000

parts by a vertical line at the mean wavelength. Mean floc size will then
amount to half that mean wavelength; see e.g., [4].
For practical reasons, the wavelength power spectrum is often presented
using logarithmic scales on the spectral density axis (and sometimes also
on the wavelength axis). When two spectra are compared, a certain vertical
difference between the spectra then corresponds to a specified ratio between
the spectral densities, independently of the absolute level.

2.2 Specific perimeter

One method to evaluate mean scale information from a two-dimensional field
of variations is to calculate the specific perimeter. Using image analysis, the
variation field is transformed into a two-dimensional grey-scale field, and a
cut is made at 50% of the mean grey scale intensity. At this grey scale level,
two-dimensional flocs of different sizes will be defined, and the specific per-
imeter is defined as the sum of the circumferences of all the flocs divided by
the total image area. The dimension of the specific perimeter will be inverted
length scale.

2.3 Wavelets
Wavelet methods give a more complete description of local variations in
comparison with power spectra. While the power spectrum gives an average
of variations of different scales over the whole sample area, wavelets also
specify the location of specific variations. In the power spectrum, variations
are described based on comparisons with sine wave shaped curves, while
with the wavelet technique the choice of curve shape is free (even if some
mathematical restrictions exist). Two examples are given in Figure 2.
An application of wavelet technology (Morlet) is demonstrated for the

Figure 2 Left: Mexican-hat wavelet, Right: Morlet wavelet (sinusoidal signal weighted
with a Gaussian distribution), real part (solid line) and imaginary part (dashed line).

12th Fundamental Research Symposium, Oxford, September 2001 439

B. Norman and D. Söderberg

signal sin(x2) as function of distance x, see Figure 3a. The power spectrum of
this curve is shown in Figure 3b, and demonstrates high spectral density at
short wavelengths, and a gradual decrease towards longer wavelengths.
In the image map from the wavelet analysis, Figure 3c, the wavelength is
plotted as a function of distance, and it is apparent that the wavelength
decreases towards longer distances. This information is not available in a
power spectrum.
In Figure 3d the same result is presented using contour lines.

Figure 3 a) sin(x2) signal, b) Power spectral density, c) Image map of wavelet (Morlet)
transform and d) Contour plot of wavelet transform with exact wavelength variation
(dashed line) [7].

440 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Use of wavelet techniques has been discussed by Keller et al. [5,6] and
applied by Söderberg [7] (see further below).

2.4 Formation
The formation of a paper sheet, i.e., the local grammage variations up to
about 40 mm wavelength, is determined by the fibre distribution in the plane
of the sheet and has a great influence on many sheet properties; it is therefore
desirable to be able to quantify this distribution. The geometrical resolution
of the basis weight measurement is of decisive importance for the amount of
variations recorded. The smaller the measurement area is, the more small-
scale variations can be detected and the larger the total variations recorded.
In industry, formation evaluation is often done using light transmission
methods. Besides local grammage variations, this method is however also
sensitive to other parameters, primarily variations in local light scattering
coefficient. Beta radiation is a good alternative to be used in evaluating
grammage variations.

Beta ray transmission

In the STFI method, based on beta radiography, which has been used since
the early 1970s, an X-ray film in contact with the paper sample is exposed to
transmitted beta particles from a C-14 radiation source with a size of
100 × 150 mm2 [1]. The scanning of the exposed and developed X-ray film
has later been modified, and the radiographs are now analysed using a desk-
top scanner, giving a 300 dpi resolution (which corresponds to ≈0.08 mm
resolution) grey-scale map. Based on calibration areas along the radiograph
edge, the grey-scale map can be transformed into a basis weight map. Thus a
two-dimensional map of actual sample basis weight can be generated [8].
Video-beta radiography was applied by Cresson and Luner to evaluate
paper formation [9]. They described the texture of basis weight variations
using co-occurrence matrices to evaluate parameters like floc size and shape
[10,11 and 12].
Dodson et al. have later used beta radiography with a CCD camera for
radiograph analysis [13].
The Ambertec meter [14], directly records beta ray transmission within
local areas of 1 × 1 mm2. The measuring point is moved automatically using
an x-y table, and 400 measuring points are collected on a paper sample size of
70 × 70 mm2. This method is commercially available, and widely used.
A similar principle was later applied at IPST, using a sample size of 80 × 80
mm2 [15].

12th Fundamental Research Symposium, Oxford, September 2001 441

B. Norman and D. Söderberg

Luner et al. compared four imaging techniques [16]:

• β-radiography;
• electrography;
• light transmission; and
• soft X-radiography.
Keller studied electron beam transmission imaging to quantify formation
[17].

Formation characterisation
The Formation number F is often used to denote the coefficient of variation of
local basis weight, that is, standard deviation divided by mean basis weight
and it is often expressed as a percentage [1]. One exception was the use by
Dodson of the same term to denote the variance of local grammage of a
paper sample, normalised with that of a sheet with random fibre distribution
[18]. Although physically sound, this use of the term is unpractical, since
knowledge of the corresponding random fibre distribution is usually missing.
It would require complete knowledge of dimension and coarseness values for
all the individual fibres.
Normalisation with the formation of a standard laboratory sheet made
from the same furnish has also been used be several researchers, e.g., Lloyd
[19].
Regarding normalising the formation values for differing mean grammage,
a method based on statistical means has been used [1]:

Fnorm = F√wm/wn (1)

where wn = normalization basis weight and wm = sample basis weight.

Normalization according to Equation 1 assumes that all layers in paper
samples of different basis weights have similar fibre distributions and are
independent. This will generally not be strictly true, when basis weight is
changed on a paper machine. At constant forming concentration, e.g., the
total structure will often improve with increasing basis weight due to a self-
healing effect during dewatering of additional sheet layers [1] (see further
discussion below). 60 g/m2 was chosen for normalization basis weight in the
STFI Formation method.
Equation (1) was later also promoted by Dodson [18].
A useful description of formation, should quantify the amount of local
variations, but also give some scale information. A change in forming condi-

442 Session 3: Forming

 Overview of Forming Literature, 1990–2000

tions sometimes can reduce the number of large flocs by breaking them apart,
and thus move them to a smaller floc size range. This may not always show up
in a single formation number.
In the STFI method, formation is described by its wavelength spectrum.
The information in a complete formation spectrum can be simplified by
integration within different scale ranges. In the STFI method, a small-scale
wavelength range of 0.3 to 3 mm and a large-scale range of 3 to 30 mm are
used, and together they make up the total formation number F.
In the Ambertec meter, the limited sample size sometimes requires
measurements on several samples to give accurate values of the standard
deviation. The normalized Ambertec formation value with dimension
“(g/m2)0.5” (not easy to understand) can actually be interpreted as the dimen-
sionless normalized formation number F according to Equation 1, using a
normalization basis weight of 1 g/m2. It is thus possible to compare an
Ambertec to a STFI value by multiplication with 600.5 ≈ 7.7. It should how-
ever be remembered that the Ambertec meter suppresses all variations
smaller than 1 mm, so its formation value mainly correlates with the STFI
large-scale value.
Dodson et al. used the variance against zone size to characterise the scale
distribution in basis weight variations [20]. This is defined as the integral of
the wavelength spectrum from the specified zone size to some upper limit.
This makes it hard to evaluate differences in small-scale variations between
two samples, since the influence of the dominating variance at larger zone
sizes will always dominate.
An alternative method of evaluating formation scale is by the specific per-
imeter as promoted by Jordan et al. [21].
Using wavelets to characterise formation, it is possible to describe the
distribution of floc sizes within different parts of a paper sample, not only
give the average over the whole area. This method has been applied by Keller
et al. [5].

Periodic marking
While formation measures mainly quantify the degree of fibre flocculation in
a paper sample, different kinds of periodic marking is also of interest. Such
marking is mainly generated by wire patterns, but sometimes also by the felts.
In a one-dimensional formation spectrum, wire marks will show up as peaks
in the spectrum. It will then be possible to rank samples depending on the
magnitude of such peaks.
A more complete way of describing periodic marks is to first calculate
the two-dimensional frequency power spectrum, in which periodic variations

12th Fundamental Research Symposium, Oxford, September 2001 443

B. Norman and D. Söderberg

will show up as a pattern of individual points of higher intensity than the

surrounding; see Figure 4.

Figure 4 Two-dimensional frequency spectrum (left) and corresponding geometrical

pattern for wire mark (right) [22].

These points can then be inversely transformed back into the original x-y
plane, where the geometric pattern of the generating wire and/or felt will
show up, see e.g., [22,23].

Optical measurements
The term “formation index” is reserved for optical formation meters. Such
meters have the drawback that they do not give absolute values of basis
weight variations, due to their sensitivity to variations in light scattering
coefficient.
This has been demonstrated repeatedly, recently by Duffy et al. [24]. Two
light transmission meters were compared with the Ambertec and STFI
methods (beta ray transmission). The light transmission methods gave reverse
results from the beta ray methods when e.g., grammage was changed in
laboratory forming.
Luner et al. [16] concluded that for newsprint samples, electrography gave
higher spatial resolution, shorter exposure time and wider basis weight range
than β-radiography. Light transmission gave the poorest spatial resolution
and correlation with mass. Soft X-radiography gave the highest spatial
resolution but the poorest contrast.
Bernié and Douglas [25] compared light transmission and beta ray trans-
mission and came to the conclusion that light transmission can be used
to measure formation. This was however based on absolute calibrations and

444 Session 3: Forming

 Overview of Forming Literature, 1990–2000

was applied to laboratory sheets only. For practical purposes, this is of less
interest.
Bouydain et al. used uncalibrated light transmission images in combin-
ation with wavelet analysis to study paper formation [26].
However, optical methods due to their simplicity can still be valuable, for
instance to follow changes on a specific machine at constant furnish, basis
weight and calendaring conditions. Absolute comparisons between paper
samples from different machines should be avoided.
Two of the more commonly used optical formation testers in North
America are the NUI (Non Uniformity Index) meter [27] and the M/K-meter
[28]. The NUI-meter analyses light transmitted through a rotating paper
sample, and a total variation number is given as a formation index. With the
M/K-meter, on the other hand, a high value corresponds to good formation.
Recently also the Kajaani optical formation tester has been introduced and,
like in the M/K-meter, the formation index is higher for a more even paper
sample.
If the aim of an optical formation meter is to resemble perceived look-
through formation the wavelength distribution of the light analysed should
be matched with the colour response of the human eye [29].
Lloyd [19,30] introduced the parameter “surface ply variation”, SPV, to
characterise the final mixing in paper samples manufactured from a three-
ply headbox. A light blue colour was added to the middle ply, to make it
possible to optically evaluate the degree of mixing between centre and outer
layers. 87 × 174 mm2 samples were analysed using a colour scanner at a
resolution of 300 dpi. The cyan part of the image was separated. This
image was inverted, i.e., a negative taken, because the cyan represented an
absence of the white surface layer, and it was the absence of the surface
layer that was of most interest. The resulting image was a grey-scale TIFF
image. The image was analysed with a procedure similar to that used for the
evaluation of the small-scale and large-scale STFI formation values; see
further below.

On-line measurements
Today optical CCD-techniques can be applied to both laboratory and on-line
evaluation. One example is the formation meter developed at CTP [31]. Using
a CCD camera a sample size of 120 × 120 mm2 is scanned with 0.25 mm
resolution. From this a total formation index can be evaluated as well as flocs
in the size classes 1, 2, 3, 6, 10 and 16 mm. It is also possible to evaluate shape
(MD/CD) as well as angle of orientation of flocs, see Figure 5. Means are
also provided to filter out and present periodic wire markings. In on-line

12th Fundamental Research Symposium, Oxford, September 2001 445

B. Norman and D. Söderberg

Figure 5 Presentation of results from the CTP on-line formation analyser [31].

applications a stroboscope is used for incident light, while ordinary light

sources are enough for laboratory evaluations.
A more simple on-line formation analysis can also be made using the ABB
Hyperscan System; see the following section.

2.5 Medium-scale variations

Maps of the two-dimensional basis weight distribution in the scale above the
formation range, that is, around 40 mm, up to about 10 m, give good pictures
of the medium-scale variations in basis weight and is an important tool for
tracing the origin of different basis weight defects. Equipment for handling
full size paper rolls, and methods for measuring basis weight measurements
using beta ray absorption techniques in a laboratory scale has been developed
by STORA [32]. Figure 6 shows an example of such a basis weight map.
Two-dimensional maps of light transmission are recorded on-line with the
ABB Hyper-scan system [33,34]. Even though only light transmission vari-
ations are recorded, some calibration against actual basis weight variation is
possible by comparison with the basis weight signal from the conventional
scanning beta meter at the dry end. A resolution of 1 mm2 is obtained, and it

446 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Figure 6 Grey-scale picture (original in colour) of basis weight of full width paper
web. Newsprint quality, with each level curve indicating a deviation of 1 g/m2 [32]
(Courtesy STORA, original in colour).

is even possible to on-line follow the patterns of the narrow wire shower
marks using this technique.
Honeywell-Measurex developed a method for detecting water weight in the
wire section, aiming at following the CD and MD dewatering profiles [35].
Non-scanning sensor technology was used, based on the measurement of
effective electrical properties of water between sensor elements mounted
below the wire. It is yet uncertain if it will develop into commercial equipment.

2.6 Fibre orientation

An overview of fibre orientation in paper was made by Niskanen [36].
During the last decade, the details of fibre anisotropy have received in-
creased interest, much depending on the easy recording of angular distribution

12th Fundamental Research Symposium, Oxford, September 2001 447

B. Norman and D. Söderberg

of paper elastic modulus using automated ultrasonic techniques [37,38]. Spe-

cial emphasis has been given to fibre orientation misalignment, but fibre
orientation anisotropy has also become more important.

Laboratory evaluation
Hasuike et al. [39] analysed fibre orientation anisotropy after splitting a
sample into eight layers using adhesive tape; see Figure 7. Fibre orientation

Figure 7 Principles for sample splitting into eight layers [39].

anisotropy in the layered samples was evaluated from manual counting of the
number of fibre crossings with a 5 mm-long MD and CD test line respectively,
within sample areas of 30 × 30 mm2.
They further analysed the three-dimensional structure of paper samples
[40,41]. Fibre entanglement in the thickness direction was evaluated by
embedding a sample in epoxy resin, cutting it into 4μm thick sections (2 mm
wide) and following the positions in the thickness direction of individual
fibres along its length, through 150 sections.
Hasuike et al. [42] also studied 3-D floc structure in 45 × 95 mm2 paper
samples using tape-layering techniques. Floc distribution in each layer was
evaluated optically, and the correlation between floc positions in adjacent
layers was utilised to quantify floc dimensions in the thickness direction. A
limitation of this method is that tape delamination does not strictly split the
paper samples into even layers, since some flocs tend to resist splitting.

448 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Erkkilä et al. also used a method based on tape splitting [43]. A 60 g/m2
sample was split into ca 6–8 layers, on which 30 × 30 mm2 samples were
analysed in transmitted light using a CCD-camera with a resolution of
640 × 640 pixels. An image analysis program was developed to evaluate the
orientation distribution of fibre segments in the images. The results were
presented in polar diagrams where the orientation distribution has an elliptic-
like shape. Anisotropy was evaluated from the eccentricity b/a, and mis-
alignment angle as the deviation of the main axis from MD; see Figure 8.

Figure 8 Orientation distribution and parameter definition [43].

Local fibre layer orientation was also studied, on sub areas of 3 × 3 mm2,
and presented in vector form.

Anisotropy = b/a ≈ MD/CD (2)

Jansson [44] used another tape quality, and separated an 80 g/m2 sample
into 20 layers, which he imaged using a desktop scanner with a resolution of
600 × 600 dpi. He analysed the images of the size 40 × 40 mm2 using the
program developed by Erkkilä.
Lloyd and Chalmers used similar techniques to study the effects of sheet
structure on paper curl and cockle [45].
Scharcanski and Dodson [46] demonstrated an image gradient evaluation
technique to quantify anisotropy from a beta-radiograph or a light trans-
mission image.
Thorpe [47] used a “new method” to split a copy paper into 13 layers. The
black coloured layers were mounted in 35 mm slide frames and imaged with
a CCD camera attached to a microscope. Fibre orientating data were evalu-
ated using the Hough transform. He analysed samples of size 20 × 25 mm2,

12th Fundamental Research Symposium, Oxford, September 2001 449

B. Norman and D. Söderberg

and each picture was divided into 5 × 5 mm2 segments for local orientation
analysis. He could evaluate the occurrence of orientation streaks on the wire
side of fourdrinier formed paper.
Parker et al. [48] studied local fibre orientation at different levels in the
thickness direction using confocal laser scanning microscopy and application
of the Hough transform. The thickness resolution was ≈ 3μm and the depth
was limited to 80μm. The scanning area for each sample was limited to
650 × 650 μm2.

3 FIBRE SUSPENSIONS

It should be pointed out that the physical flocculation of papermaking fibres

is the main reason for small-scale unevenness in the final paper product. A
deeper understanding of how fibre flocs are generated and how they can be
dispersed is therefore of vital importance, if the evenness of paper products is
to be improved.
It is well established that fibre flocs are generated by mechanical fibre
entanglement [1]. Once a floc is formed, the fibres have to be disentangled.
This can be obtained by a high degree of fluid turbulence (see further in
section Headboxes). One main problem is, however, that when the fluid
motion is dissipated, fibre flocs will form again.
It was first demonstrated by Kao and Mason [49] that dispersion of flocs
by extensional flow is much more effective than by simple shear, since shear
flow mainly generates rotational movements of flocs.
Duffy and Norman [50] showed experimentally that fibre flocs can actually
be broken apart through the stretching effect in an elongational flow.
As will be discussed further below, it therefore does not seem to be possible
to avoid fibre flocculation based on turbulent shear effects, but rather by
applying floc stretching.
The behaviour of fibres and fibre flocs in flowing fibre suspensions is a
formidable two-phase problem, and its modelling is only in the initial stages.
However, experimental studies of fibre suspensions have been undertaken to
a considerable extent, and modern techniques open new possibilities in this
field.

3.1 Fibre suspension modelling

Steen [51] modelled fibre flocculation in turbulent flow, introducing the rate
of both rupture and aggregation of fibre flocs. The model was applied to pipe
flow at different Reynolds number as well as to a pipe with a step diffuser. In

450 Session 3: Forming

 Overview of Forming Literature, 1990–2000

the latter case, minimum flocculation intensity was predicted a certain

distance downstream of the step, which is close to observations reported in
literature.
Kerekes and Schell [52] introduced the concept of Crowding Factor N for
the number of fibres of length L and diameter d at a volume concentration of
cv, within a given reference volume. The reference volume was chosen as the
volume of a sphere of diameter L, i.e., the sphere created by a freely rotating
fibre of length L. The following relation-ship is then valid for the crowding
factor N:

2
2 L
N = cv
3 d 冢冣 (3)

Table 1 is a summary of the findings using the turbulence decay cell described
above. In standard hand sheet forming, the crowding factor approximately
equals unity.

Table 1 Fibre contacts at different crowding factor levels.

Crowding factor N Concentration Type of fibre contact

N<1 Dilute Rare collisions

1 < N < 60 Semi-concentrated Frequent collisions
60 < N Concentrated Continuous contact

Since the density of cellulose is approximately 1.5, it would be expected

that the volume concentration of fibres is lower than the weight concentra-
tion. However, because of the swelling of the fibre wall, the lumen volume
etc., the average volume concentration can be as high as twice the weight
concentration.
The crowding factor is a better indicator than the fibre concentration of the
flocculation tendency of a fibre suspension, and it is increasingly used to
characterise the mix conditions during forming, see Kiviranta and Dodson
[53].
Kerekes made an analysis of fibre flocculation in relation to paper forming
[54]. Besides the crowding factor, he then introduced two additional dimen-
sionless numbers: a fibre Reynolds number to reflect the ability of a fibre to
follow fluid streamlines, and a force factor to account for hydrodynamic
crowding forces in relation to fibre contact forces resisting crowding.
Basic studies of fibre network properties have been made by Andersson,

12th Fundamental Research Symposium, Oxford, September 2001 451

B. Norman and D. Söderberg

Wikström et al. [55,56] with the main aim of improving pulp processing,
which generally means concentration ranges higher than those in paper-
making. It was e.g., suggested that fibre network yield stress was a better
measure than fibre concentration for designing pulp lines [57]. Experimental
studies of fibre-fibre friction were combined with fibre force models to predict
fibre network properties [58,59].
Experiments by Bennington [60] , based on modelling of chemical reactiv-
ity indicated that in a 3% fibre suspension, 95% of the energy dissipation
takes place by direct friction between the fibres and thus only 5% in the fluid
itself.
Huhtanen [62] studied the non-Newtonian properties of fibre suspensions,
without taking fibre flocculation specifically into account. He made rheo-
logical measurements to analyse shear viscosity of pulp suspensions at mod-
erate shear rates and small deformations. He also applied a commercial code
to describe the non-Newtonian effects on pipe flow of pulp suspensions.
Klingenberg et al. have an extensive program to study flowing suspensions
of rigid and flexible fibres, modelled as chains of prolate spheroids connected
through elastic ball and socket joints [63]. Attractive forces between fibres can
form flocs, but the properties of such flocs are different from flocs formed
from elastic interlocking and inter-fiber friction. Simulations suggest that
fibre shape is important and that effective formation aids in paper-
making should reduce inter-fibre friction under normal force loadings of
1–10mN [61,64,65]. Examples of the influence of friction coefficient on the
flocculation in shear flow shown in Figure 9.
Shah et al. [66] studied the flow patterns generated by a half Δ-wing in a
water tunnel, specifically the fluid stress fields. Applying these fields on fibre
suspension models, they came to the conclusion that the local stresses were
large enough to break apart fibre flocs.

Figure 9 Simulations of shear in a 0.125 vol% suspension of helical-shaped fibres.

Static friction coefficient a: 0 and b: 20 [61].

452 Session 3: Forming

 Overview of Forming Literature, 1990–2000

3.2 Fibre flocculation evaluation

Qualitative studies of fibre flocculation in flowing fibre suspensions are
possible using different optical set-ups, but quantitative description of local
concentration variations is extremely difficult, due to secondary light-
scattering effects. Several investigations have recorded transmitted light
through a flat channel. For thicker channels or pipes, light reflection meas-
urements have a potential of better geometrical resolution, while limiting the
measurements to the surface of the suspension. Evaluations have been
carried out in turbulent flows as well as in decaying turbulence fields.
Kaji et al. [67] used a rectangular flow channel with a cross section of
60 × 3.5 mm2. Both transmission of a laser beam and imaging using a CCD
camera were used for flocculation evaluation.
Among other means they also applied fractal analysis to characterise the
geometry of floc periphery for different pulp suspensions. They found that as
floc forming proceeds, the shape of flocs tends to be simpler, while as disper-
sion proceeds, the shape of flocs tend to be more complicated
Kerekes and Schell [52] , experimentally studied the effect of different
parameters on fibre flocculation using a vertical, transparent cell of 23 × 342
mm2 cross section, and closed at the bottom; see Figure 10. A grid was pushed

Figure 10 Left: Concept for grid pulled through a vertical, narrow channel. Right:
State of flocculation for Douglas fir, 0.5% , Crowding Factor N = 75 [52].

12th Fundamental Research Symposium, Oxford, September 2001 453

B. Norman and D. Söderberg

down through the fibre suspension in the cell. After turbulence decay (app.
15s) a flocculation image was photographically recorded in transmitted light.
From this image, contrast intensity and floc scale was evaluated.
Kallmes et al. [68] demonstrated equipment for on-line analysis of mix
flocculation. Mix was extracted from the headbox into a 13 mm-wide chan-
nel. The light transmission variations of a narrow light beam directed
through the fibre suspension gave a flocculation value. A comparison was
made with the inherent flocculation tendency of the mix, as recorded after the
passage of a dispersing 90° flow elbow, as well as with a formation index
recorded with an on-line, light beam transmission meter. Preliminary test on
different paper machines showed promising correlations between flocculation
and formation values in response to e.g., changes in chemicals addition.
Zhao and Kerekes [69] used the above-mentioned cell with decaying turbu-
lence to study fibre flocculation at different suspending liquid viscosities. One
way of presenting the results is to find the “critical crowding factor”, below
which no flocculation can be detected. In pure water (viscosity 1 mPas) the
critical crowding factor for long fibre pulp is around 10, but increases linearly
to 70 when sugar is added to increase liquid viscosity to 20 mPas. This is
explained by reduced turbulent motion, decreased initial turbulence and
faster turbulence decay, and higher shear forces on the fibres.
Duffy et al. [70] made a study of pipe flow of fibres suspended in liquids
of different viscosities. The numerous flow regimes normally observed for
friction loss curves in pipe flow were not present for viscous suspensions.
Drag reduction is also observed in viscous suspensions but it onsets at higher
velocities. Increasing viscosity leads to reduced floc size in pipe flow. Overall,
it is suggested that increasing viscosity has the same effect as decreasing fibre
concentration.
Giro et al. studied the effects of adding CMC dispersants to high concen-
tration pulp slurries ( ≈ 30%) [71]. The torque requirement in a bowl mixer
was used as a measure of pulp dispersibility. It was concluded that increasing
CMC molecular weight had a positive effect, and pulp dispersion was more
closely related to the amount of CMC adsorbed than to the amount in solu-
tion. This indicates that at this high concentration level, fibre friction seems to
be the determining factor, and not fluid viscosity.
Kerekes and Schell [72] , also used the decaying turbulence cell to study the
effects of fibre length and coarseness on flocculation. Their main findings are:

• Fibre length exerts its influence on nonuniformity through the number

of contacts per fibre and floc size.
• Coarseness exerts its influence on nonuniformity through the number of
contacts per fibre, floc size and mobility.

454 Session 3: Forming

 Overview of Forming Literature, 1990–2000

• Mixtures of short and long fibres of a given length-weighted average length

gave the same nonuniformity as homogenous fractions of the same fibre
length.
They point out that the definition of the crowding factor assumes identical
fibres. Normally, a rather wide distribution of fibre lengths will be found in a
practical situation. This means that the crowding factor concept has to be
used with caution in such cases.
Beghello [4,73] used a 15 mm; wide transparent channel with turbulence
generating constrictions at the entrance; see Figure 11. A digital camera was
used and pictures were taken in reflected light of an area sized 80 × 120 mm2.
To minimize reflection from the channel bottom wall, a black surface was
placed below it. Absolute concentration variations were not possible to
evaluate. Instead the power spectrum was evaluated up to 32 mm wavelength,
and the mean floc size was calculated from the wavelength dividing the area
below the spectrum into two equal parts. The studies aimed at three main
areas:
• How the physical and mechanical aspects of suspended fibres influence the
flocculation process prior to the addition of wet end chemicals.
• How the chemical and physical environments in the fibre suspension
influence and/or contribute to the flocculation process.

Figure 11 Drawing of the turbulence generator and camera system for the floc-
measuring device (not to scale). Turbulence generator channel widths 15 mm and
length 80 mm. Lamellae length 270 mm and thickness 1 mm [73]

12th Fundamental Research Symposium, Oxford, September 2001 455

B. Norman and D. Söderberg

• How changes in the surface chemistry of suspended fibres influence the

flocculation process.
Raghem-Moayed [74,75] studied fibre flocculation in a decaying turbulence
field in a transparent flow channel with the cross section 11 × 152 mm2.
Turbulence introducing contractions were provided at the inlet. Diffuse
ingoing light was recorded in transmission over an area of 20 × 23 mm2 on the
opposite side using a CCD video camera. Calibration indicated a linear rela-
tionship between light absorbance and local concentration for hardwood
fibres in the range 0.4–0.7%. Flocculation index was evaluated as the square
of the coefficient of variation of local concentration. Due to statistical
effects, this can result in a decreasing index with increasing fibre concentra-
tion, which was also the result of the evaluations. At increased flow velocity,
flocculation index decreased. No spectral analysis was applied to evaluate floc
scale.
Wågberg had earlier made studies of chemical-induced fibre flocculation
using laser beam optics [1]. In a new investigation he applied a CCD video
camera to detect flocculation in a flow channel of cross section 3 × 170
mm2 using light transmission [76]. He determined flocculation index and
mean floc size, both values calculated from the increase in flocculation when
chemicals are added. Regarding the nature of flocculation effects for the
chemicals used, it was concluded that further tests should be made at very
short contact times to improve the understanding of the mechanisms.
Pierre [77] also used CCD-techniques to study flocculation in flowing fibre
suspensions; see Figure 12. The camera observed 60 × 60 mm2 of a trans-
parent channel with a thickness of 3–5 mm. He was particularly interested
in analysing suspension samples taken from industrial paper machines pro-
ducing filler containing fine paper, to be able to predict final paper formation.

Figure 12 Observation of flocculated suspension according to Pierre [77].

456 Session 3: Forming

 Overview of Forming Literature, 1990–2000

One main objective therefore was to evaluate flocculation in the presence of

considerable amounts of highly light scattering filler particles.
Since absolute calibration of transmitted light against fibre concentration
is difficult, especially at the presence of filler, he instead evaluated floc sizes.
He evaluated flocculation as the floc area in six different size ranges from 0.7
to 13 mm but also summarised the results as a flocculation index. Effects of
refining and retention chemical addition on flocculation were studied.
Swerin applied oscillatory shear to fibre suspensions, and studied their
rheological properties [78]. A mathematical scaling was found to express the
elastic shear modulus as a function of the fibre concentration in excess of the
threshold concentration (below which no floc strength can exist). He also
studied the effects of chemical additives. It was suggested that the effect
of polymeric flocculants on the network strength is due to more fibres
being active in the network and to an increase in the bonding strength in the
fibre-fibre contacts.
What is known as MC-technology within pulp processing usually involves
fibre concentrations above 6%. Gullichsen et al. have studied floc disruption
in that concentration range for more than ten years, aiming at paper forming
processes [79,80]. Fibre floc disruption takes place in the narrow gaps
between wings on a rotor and stationary ribs; see Figure 13. Experiments to
form paper using the same geometry have been reported, but no products
with acceptable paper properties have been demonstrated [81].
Björkman [82] has performed a considerable amount of experimental
studies and interpretations of the flow of flocculated fibres. A shearing cube
demonstrated the behaviour of fibre suspensions under deformation in

Figure 13 Fluidisation chamber for high concentration forming [81].

12th Fundamental Research Symposium, Oxford, September 2001 457

B. Norman and D. Söderberg

compression and extension respectively. The main equipment developed

for his studies was a viscometer with independently computer-controlled
rotation of inner bob and outer cylinder, with a 20 mm gap width in between.
Fibre floc splitting and sizing under different shearing conditions were studied
for different fibre types and concentrations. The occurrence of Taylor vortices
was observed, as well as fibre orientation anisotropy streaks generated under
such conditions; see Figure 14. Such effects are of great importance for the
generation of fibre orientation anisotropy streaks in headbox flows.
Helmer et al. [83] studied the concentric mixing of thick stock (central
pipe) and water (surrounding pipe). By suspension sampling they measured
the concentration profile over the pipe diameter at different locations down-
stream. The way of sampling at different radial positions resulted in a con-
siderable averaging in the flow direction. The results were largely consistent
with those reported in an earlier study [1]. However, it also indicated that fibre
concentration at pipe centre could remain at a higher than average level for
a considerable distance, which was not possible to reveal in the previous,
qualitative, visual study. The pipe dimensions in the new study were compar-
able to those of the old, i.e., small compared to industrial conditions. This

Figure 14 Taylor vortices in fibre suspension of about 0.5% concentration. Left:

Schematic view with principal motions of one vortex pair. Photographs: The same
vortices photographed with photo lamp to display the motion (middle) and with flash
to display the flocs (right), channel thickness 20 mm [82].

458 Session 3: Forming

 Overview of Forming Literature, 1990–2000

means that there is still uncertainty whether or not the higher Reynolds num-
ber at realistic dimensions will generate better mixing.
Ringnér and Rasmuson [84] applied X-ray tomography to study 3D fibre
distribution under stationary conditions. Their main interest was to investi-
gate inter and intra floc concentrations in a range above 3%, to improve the
understanding of pulp treatment processes.
Kellomäki et al. [85] used a fast CCD-camera to detect light transmission
through a 30 × 40 mm2 cross sectional rectangular channel. They compen-
sated for unevenness in the background illumination and relied on the
Lambert-Beer law to calculate fibre concentration as a function on the
amount of transmitted light. They evaluated floc size at a grey scale level of
average intensity, so that 50% of the image was always interpreted as fibre
flocs. They assumed cylindrically shaped flocs, and calculated the effective
dimensionless floc volume Vf* from the equation
2
LmxLmy
V f* = , (4)
l f3

where Lmx is the mean floc size in the flow direction, Lmy is the mean floc size
in the cross direction and lf is the mean fibre length.
Karema et al. [86] studied the transient fluidisation of fibre suspensions
after a step diffuser, using an optical flocculation analysis method described
above. They evaluated the dimensionless floc volume Vf* (see Equation 4) as a
function of distance/time after the step at different levels of expansion head
loss, which was obtained by different flow velocities; see Figure 15.
Short fibre and long fibre pulps were used. The results indicate that the
growth of the dimensionless floc volume Vf* during re-flocculation follows a
common power law with respect to time for the pulp types studied. At later
times, floc size saturates at a constant value.

3.3 Velocity and turbulence measurements

Particle Image Velocimetry (PIV)
This technique has been applied to flowing fibre suspensions, mainly for flow
inside headboxes or for the free jet leaving the headbox. It will be treated in
section Headboxes.

Ultrasound
Karema et al. used ultrasound to measure fibre turbulence in the thickness
direction of a 30 mm thick, plane channel [86]. The instantaneous fibre phase

12th Fundamental Research Symposium, Oxford, September 2001 459

B. Norman and D. Söderberg

Figure 15. Dimensionless floc volume Vf* as a function of time tc elapsed from step
expansion at different head loss hf during expansion [86].

velocities were recorded with a pulsed Doppler ultrasound anemometer,

exploiting a single transducer for emitting and receiving ultrasound pulses.
The instantaneous beam-wise velocity component at a location on the path
of the ultrasound beam is obtained from the Doppler shift frequency of the
echoed pulse induced by moving fibres. Velocities at multiple locations along
the beam were achieved by detecting the signal with specific time delays
after emission. The distance travelled by the pulse was then obtained by
multiplying with the speed of sound in the fluid. An example of turbulence
measurement is shown in Figure 16.

NMR tomography
Fibre suspension flow was studied by Li et al. using NMR imaging tech-
niques. Pipe flow profiles at flow velocities up to 1 m/s were studied [87].
Various flow patterns from plug flow to complex flow were observed for
hardwood suspensions of ca 0.5% fibre concentration. Only plug flow was
observed at 0.9% for a softwood pulp suspension. Studies were also made of
pulp flow through an abrupt contraction [88], both with water flow and pulp
suspension flow. For the pulp suspension, they observed plug flow in the
regions both far upstream and far downstream of the contraction. Just

460 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Figure 16 Standard deviation of vertical velocity component of fibres across a

horizontal channel. Data collected from 500 velocity profiles at some distance
downstream of an expansion. Npulse denotes the number of pulses emitted and echoes
received for each profile. The upper and lower walls of the channel are located
approximately at y = 5 mm and at y = 35 mm [86].

upstream of the contraction, off-centred maximum velocities were observed.

The extensional flow near the sudden contraction appeared to be highly dis-
ruptive to the fibre network. Downstream, the plug in the central core of the
tube gradually re-established.
Pipe flow mechanisms with a 0.5–0.9% concentration short fibre suspen-
sions at velocities up to 3 m/s were studied in a 26 mm diameter pipe [89]. The
whole range from fully turbulent to mixed flow with a steady plug core was
observed. The volume fraction of the fluidised pulp suspension correlated
linearly to the mean bulk flow rate. They also studied the influence of reten-
tion aids addition on the flow behaviour of fibre suspensions [90]. It was
demonstrated that, at low flow rate, adding a polymeric flocculant to a dilute
hardwood fibre suspension could introduce a slip velocity between the dense
fibre aggregates and the remainder of the suspension. Introducing retention
aids at fully turbulent flow resulted in the forming of plug flow immediately
after the addition. The disruption of fibre networks flocculated by retention
aids was also monitored as a function of the shearing time at constant flow
rate.

12th Fundamental Research Symposium, Oxford, September 2001 461

B. Norman and D. Söderberg

4 HEADBOXES

A main issue in modern paper machines is to optimise paper structure with

respect to the important paper properties of a specific product, and at the
same time minimize the MD and CD variations in grammage, fibre orienta-
tion anisotropy and misalignment as well as two-sidedness. The headbox
design is critical in this perspective and it is obvious that in-depth studies of
fibre suspension flow in the headbox are needed in order to achieve higher
production rates and improved products.
In principle, the headbox consists of three parts:
• CD distribution, usually a tapered header;
• tube bank(s), to distribute the mix more evenly;
• headbox nozzle, to generate a high quality jet.
A recent overview of headbox flow considerations was given by Houvila et
al. [91], and a discussion of the future by Pantaleo [92].

4.1 Tapered header and CD control

The geometry of the tapered inlet header is primarily affecting the paper
properties on a CD scale that is the width scale of the paper machine. Histor-
ically the manifold shape has been designed to give a constant static pressure
across the machine width.
One disadvantage with the tapered header is its sensitivity to changes in
flow rate. The shape of the tapered header has together with the slice-lip
actuators been the two parameters used to control CD variations. The slice
lip has then been used to control the variations on a small scale, down
to ca 100 mm. However, when the slice lip opening varies across the width of
the machine, cross-flows (CD velocity components) are generated both inside
the headbox nozzle and in the jet, which cause fibre orientation misalignment
streaks.
A somewhat different design is the BTF distributor by Schultz [93,94]; see
Figure 17, which replaces the tapered header. The stock enters the central
distributor from the bottom. The central distributor itself is a chamber with
an air pad at the top and the stock is distributed through radial, flexible hoses
equally spaced around the periphery of the tank. The hoses are connected to
the back of the headbox, and all are of the same length in order to have an
equal pressure drop. This design allows a good CD distribution of the flow to
the headbox, regardless of the incoming flow rate.
To also overcome the problem with fibre misalignment due to slice lip
variations, Shultz presented the idea of adding dilution water at the radial

462 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Figure 17 Radial CD-distributor according to Schultz, including BTF dilution

control system.

distributor. The water is led into the fittings where the flow leaves the tank. By
this local CD addition of water, the local mix concentration can be con-
trolled. Since the basis weight is changed already inside the headbox it is no
longer necessary to use the slice lip to control the grammage profile. Hence
fibre orientation misalignment due to the induced cross-flow can be avoided
and grammage control is de-coupled from fibre orientation misalignment.
The BTF dilution control system has also been designed to be retrofitted on
existing paper machine headboxes; see Knoller et al. [95].
The concept of dilution control is today also applied in standard headbox
designs, where the stock is supplied to the headbox by a tapered header. The
first implementation was a Voith-Sulzer Module Jet headbox in Munkedal,
which was put into operation in January 1994 [96,97,98]. The dilution control
concept has been widely adapted and there are several technical solutions
used to add the extra water.
In Figure 18 (top) the addition of dilution water in the Voith-Sulzer Mod-
ule Jet headbox can be seen [97]. Mix and white water is fed to the headbox.

12th Fundamental Research Symposium, Oxford, September 2001 463

B. Norman and D. Söderberg

Figure 18 Three different concepts of dilution control. Top: Voith [97] , Bottom, left:
Beloit [99]. Bottom, right: Valmet [91].

Motorised valves control the addition of white water, which is blended with
the mix in a mixing chamber located in a first, single-row tube bank. An
orifice is located after the mixing chamber and the unit (valve, mixing cham-
ber and orifice) is designed so that the volume flow through the orifice
(mix + dilution water) is kept constant when the amount of dilution is
changed. A good blending of the additional white water and mix is obtained
by injecting the dilution water at a specific angle into the mixing chamber.
Figure 18 (bottom, left) shows the technique used by Beloit in the Concept
IV-MH headbox [99,100]. The white water is added already in the tapered

464 Session 3: Forming

 Overview of Forming Literature, 1990–2000

header. Since it is added before the entrance into the tubes the flow rate into
the tubes is only given by the pressure distribution in the tapered header.
Hence it remains practically unaffected by the addition of extra water. The
direction of the added water into the manifold assures a good mixing.
In Valmet’s Sym-Flo D headbox, there are two tube banks separated by a
stilling chamber and the additional water is injected directly into the tubes in
the first tube bank, at a step change in pipe diameter; see Figure 18 (bottom,
right). The constant pressure in the stilling chamber then assures a constant
flow rate across the headbox nozzle; see Nyberg and Malashenko [101].
Dilution control is a powerful technology to obtain good CD control of
grammage and fibre orientation misalignment, and represents a leap forward
in paper process technology. It is widely used on paper machines throughout
the world. Many investigations have been made to study the application of
dilution control systems in different applications [102,103,104,105,106,107].
Lee and Pantaleo [108] performed CFD simulations aimed at investigating
the influence of headbox flow features on fibre orientation.
In order to achieve the most uniform grammage and fibre orientation CD-
profiles, Hämäläinen and Tarvainen [109] introduced optimisation methods
coupling numerical simulations (CFD) of headbox flow and paper properties
such as grammage and fibre orientation misalignment. They illustrated the
method by two industrially applicable examples. These were the shape of the
tapered header and grammage control both by slice bar position and dilution.
In order to perform an optimisation, a cost function had to be defined. For
the case of the tapered header the cost function should capture the departure
from a uniform flow velocity CD profile. This can be i.e., the standard devi-
ation of the flow velocity in CD. For the case of slice and dilution control the
cost function could be based on grammage and/or alignment non-uniformity.
The solution to the problem was obtained by standard optimisation
techniques.

4.2 Tube bank

The main purpose of the tube bank is to produce a pressure drop that pro-
motes a more uniform CD flow profile. The pulp suspension enters the tube
bank through holes, which usually have an open area of about 10%. In mod-
ern tube banks, each flow channel has a step-diffuser design, which results in
a well predictable separation of the flow and assures a controlled and elevated
pressure drop. The step diffusers are also believed to play an important role
for floc dispersion through the strong flow gradients created by the separated
flow and associated turbulence generation [1].
Fortier [110] tested different tube designs regarding pressure drop using

12th Fundamental Research Symposium, Oxford, September 2001 465

B. Norman and D. Söderberg

water and also studied the flow conditions in the tubes with particle
visualisation.
The results can be seen in Figure 19 where the pressure drop can be seen as
a function of velocity through the tubes. The different geometries are
included in the figure. As can be seen the highest obtained pressure drop is
four times as high as the lowest.
Hauptmann et al. [111] discussed the effect of turbulence generation by
backward facing steps and wakes, and made experiments using hot-film
anemometry in water. They looked at the generation and decay of turbulence
and the turbulent length-scales. Comparisons of the results were made with
visualisations of the flow behaviour of the jet and the conditions on the
forming table.
Shariati et al. [112] numerically solved the flow through a manifold, tube
bank, contraction and slice. The tubes in the tube bank had a somewhat older
design with a tapered diffuser instead of a step diffuser. The results clearly
showed the effect of re-circulation, which was generated by flow separation in
the diffuser. Since the complete flow was solved, the effect of non-uniform
flow conditions showed a strong influence in the behaviour of the flow inside
the tubes depending on their position in the headbox.
Due to the rapid re-direction of the flow when it enters the tube from
the manifold, flow separation is also likely to occur at the inlet end on the
upstream side of the tube, immediately after the tube entrance. Since the

Figure 19 Tube pressure drop as function of inlet velocity for six different tube
designs [110].

466 Session 3: Forming

 Overview of Forming Literature, 1990–2000

step-wise increase in area with its associated large-scale separation and re-
circulation will generate turbulence, it will have an influence on the dispersion
on fibre flocs. Also, the strong acceleration of the entrance flow into the tubes
can possibly contribute to the dispersion of flocs as well as to aligning fibres
in the flow direction.
However, the strong turbulence generation in the step diffuser will eventu-
ally also re-generate flocs through the strong mixing.
Karema et al. [86] presented results regarding the effect of transient fluid-
isation in channel flow due to a step increase in area, which is closely related
to the effect of step diffusers in the tube package. Their experiments are
described in section Fibre suspensions. The results show a flocculation min-
imum some distance downstream of the steps, after which the flocculation
increases; see Figure 15.
In a typical design, the manifold flow towards an individual hole is not
perfectly symmetrical, which can generate a strong swirl into the tube. With
the flow rates used in paper manufacturing this could give a considerable
effect and create strong vortices in a tube bank. Even if this can occur gener-
ically, the effect has been applied wilfully, by generating strong swirls inside
the tubes in order to promote isotropy in the final paper sheet, Aidun
[113]. This has been later patented [114] and is discussed further below; see
Figure 23.

4.3 Headbox nozzle

In the headbox nozzle, the mix flow from the tube package is accelerated to
the desired jet velocity. This will impose a pressure drop on the flow, which
will even out the CD flow profile and reduce the degree of turbulence. The
higher the nozzle contraction (inlet height divided by jet thickness), the better
the potential jet quality.

Fibre orientation effects

A surprising effect of large headbox nozzle contraction on fibre orientation
anisotropy in the paper produced was found in twin-wire roll forming
experiments with TMP in the FEX machine by Nordström [115]. Twin-wire
roll forming was used to evaluate headbox effects on paper properties, since it
introduces the least amount of fibre rearrangement in the wire section, due to
the gradual changes in dewatering pressure (however, there is not a constant
pressure level as traditionally assumed; see further below). The studies were
made at different mix-to-wire speed differences, but to avoid changes in
headbox flow characteristics, headbox flow was kept constant and only wire

12th Fundamental Research Symposium, Oxford, September 2001 467

B. Norman and D. Söderberg

speed was changed. Zero mix-to-wire speed difference was defined at the wire
speed giving minimum fibre orientation anisotropy. Due to the minimum
amount of shear introduced during these dewatering conditions, fibre orien-
tation anisotropy in the paper produced should be a good approximation to
that already existing in the jet.
At low and medium headbox nozzle contraction ratios (traditional for
hydraulic headboxes), fibre orientation anisotropy in the paper produced
behaved in the traditional way, with some increase of anisotropy at the min-
imum point, at increased contraction ratio; see Figure 20. At high nozzle
contraction, however, very high fibre orientation anisotropy values were
obtained even at zero speed difference. This indicates that fibre orientation
anisotropy was high already in the jet. The logical reason for this is the
strongly elongational nozzle flow. It was also found that paper formation was
improved with higher nozzle contraction ratio, which indicates that fibre
flocs were broken apart in the elongational flow – see further in a later
section.
As stated above, headbox jet quality improves with increasing nozzle con-
traction ratio. It would then be a natural choice to use a high nozzle contrac-
tion ratio in all headboxes. However, since most paper and board products
require a low fibre orientation anisotropy, a high contraction ratio can only
be utilised for products allowing a high MD/CD ratio, such as printing papers
containing mechanical pulp.
Ullmar [118,119] studied the fibre orientation anisotropy effects in a trans-
parent headbox nozzle. A flat channel extended the slice, to facilitate optical

Figure 20 MD/CD-ratio of tensile stiffness for three different headbox contraction

ratios during forming of TMP at different mix-to-wire speed differences [116,117].

468 Session 3: Forming

 Overview of Forming Literature, 1990–2000

measurements on the emerging jet; see Figure 21. A stroboscope and a video
camera were used to record the fibres in the jet, and image analysis was
applied to analyse fibre orientation characteristics.
Nozzle contraction ratio was varied within the range 17–50. Due to the
opaqueness of fibre suspensions, it was not possible to view individual fibres
more than a few mm inside of the jet surface. Therefore the majority of the
experiments were performed at very low fibre concentrations. Coloured
polymer fibres (comparatively straight) as well as cellulose fibres (more curly)
were used. Some of the conclusions are listed below:

• For dilute suspensions the fibre orientation anisotropy is highly depend-

ent on contraction ratio and moderately sensitivity to flow velocity
level.
• At higher fibre concentration lower anisotropy is obtained compared to
low concentration and higher flow velocities gives higher anisotropy.
• The effect of fibre dimensions is that stiffer fibres are more oriented than
flexible fibres and longer fibres are more oriented than short fibres. Also,
straight fibres are more oriented than curled fibres.

Figure 21 Transparent headbox for fibre orientation studies in the headbox jet; to
the upper right: video camera and stroboscope [118].

12th Fundamental Research Symposium, Oxford, September 2001 469

B. Norman and D. Söderberg

Table 2 shows the effects when two different contraction ratios, CR, and
two different nozzle lengths, L, are applied. It is assumed that the flow vel-
ocity at the slice is constant, which also means that the volumetric flow rate is
constant.
Given a large contraction ratio and a short nozzle the acceleration is high-
est, which would give the strongest strain on the suspension. The opposite
would be a small contraction ratio with a long nozzle, which would give a low
acceleration and weak strain. However, the longest time is spent in the elon-
gational flow field for the long nozzle with a large contraction ratio and the
shortest for the short nozzle with a small contraction ratio. Hence, if the time
spent in the contraction is of importance for the final paper sheet it is really a
two-parameter problem.
Asplund [120] used the headbox in Figure 21 to study the fibre orientation
anisotropy at different levels in the jet. Instead of a stroboscope, a ca 1 mm
thick laser light sheet was used as a light source. In Figure 22, results are
shown with a centrally inserted vane in the nozzle, ending 30 mm upstream of
the nozzle end. Fibre orientation anisotropy is significantly lower in the upper
lip boundary layer than outside of it. The thin vane produces approximately
the same decrease in anisotropy at jet centre as the boundary layer effect,
while the thicker vanes generate further disperging fluid motions.
Aidun [114] suggested a method to decrease fibre orientation anisotropy
after the tube package feeding the headbox nozzle; see Figure 23. The left
illustration shows an insert, which is mounted inside each tube in the tube
bank. This insert will generate a strong and controlled swirl. The swirls that
are generated by the different tubes are not all rotating in the same direction.
The amount of tubes generating clockwise vortices is the same as the number
that generates counter-clockwise vortices and the two types of tubes are
arranged in specially designed patterns. An example can be seen in the right
illustration where a staggered pattern is used. To obtain an overall design so
that all swirls cancel each other before leaving the headbox nozzle will be a
delicate task.

Table 2 Effect of headbox nozzle contraction geometries on flow parameters.

Contraction Contraction Flow Time in

Ratio CR Length L acceleration contraction

Large Long Intermediate Long

Large Short High Intermediate
Small Long Low Intermediate
Small Short Intermediate Short

470 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Figure 22 Fibre orientation anisotropy in the upper half of a 15 mm-thick jet. Three
alternative vanes inserted: thin, thick with pointed tip and thick with blunt tip [120].

Figure 23 Swirl generation in feed pipes (Left) and rotation directions in adjacent
feed pipes (Right) according to patent by Aidun [114].

12th Fundamental Research Symposium, Oxford, September 2001 471

B. Norman and D. Söderberg

Flow properties of nozzle flow

The flow through a contraction is basically given by the preservation of mass
flow and since the area is decreasing the flow velocity is increasing inversely
proportional to the area, i.e., an inviscid flow. The mean flow in the headbox
can, with a high degree of accuracy, be described as inviscid, i.e., neither
viscosity nor turbulence has any larger influence. However, even though it is
not important for the mean flow, the presence of turbulence is still believed to
play an essential role for individual fibres and flocs in the headbox nozzle.
In the early 1980s, Chuang studied turbulence (Laser Doppler Velocimeter,
LDV) and flocculation (0.5 μs flash and Polaroid camera) in headbox nozzles
[121]. He compared a plain nozzle, a triangular bump insertion and a double
plate insertion; see Figures 24 a–c.

Figure 24 Turbulence generators (triangular bump and double plates) in plain

headbox nozzle, and their geometrical parameters, Chuang [121].

472 Session 3: Forming

 Overview of Forming Literature, 1990–2000

His main results can be summarised as follows:

• turbulence level decreased with nozzle angle and flow velocity;
• turbulence level decreased with increasing suspension concentration;
• small-scale turbulence highest for plain nozzle and lowest with bump
insertion;
• large-scale turbulence highest with bump inserted and lowest for plain
nozzle.
Flocculation scale was largest with bump insertion and smallest with
plates insertion. Shands [122] demonstrated the effect of vanes, which have
been used in Beloit Converflo headbox nozzles for several decades. By land-
ing the jet on a fourdrinier wire and recording mix surface irregularities,
it was clear that vane presence was positive, i.e., they gave a reduction in
surface irregularities. This is due to the damping of the large-scale flow
irregularities; see Figure 25.
Farrington [123] performed simulations of the headbox nozzle flow using a
commercial CFD-package. The simulations were three-dimensional (3-D)
and used the standard k-ε turbulence model. One part of the investigation
was to compare 2-D and 3-D simulations. He showed that the result of a 2-D
simulation could differ significantly from a 3-D simulation, which is to be
expected given the nature of the numerical methods. For the 2-D case, condi-
tions have to be specified at the inlet to the headbox nozzle (exit from tube
bank) regarding velocity, turbulent kinetic energy and turbulent dissipation.
These values have to be estimated and averaged in MD for the 2-D case. For
the 3-D case, values have to be estimated at the individual tube exits. Between
the tubes conditions are given by the presence of solid surfaces. Hence

Figure 25 Mix surface roughness profiles with and without headbox nozzle vane [122].

12th Fundamental Research Symposium, Oxford, September 2001 473

B. Norman and D. Söderberg

the inlet conditions for the 3-D case are much more physically relevant.
The result of the 3-D simulations was in good agreement with experiments
regarding the behaviour in the contracting nozzle. However, by comparison
of different figures in the article regarding the behaviour of turbulent kinetic
energy, it is clear that a comparison with experimental results was not per-
formed for the end of the nozzle contraction, where large differences are to be
expected. This is discussed below.
Farrington also investigated the effect of vanes in a nozzle; see Figure 26.
The graph shows the behaviour of turbulent kinetic energy and turbulent
macro scale with and without vane. The kinetic energy behaves similarly in
the two cases throughout the nozzle until the position where the vane ends,
where a strong increase is noticed. The vane tip generates turbulence, which
however decays before nozzle outlet. The vane clearly has a restricting effect
on the turbulence macro scale.
Hua et al. [124] numerically solved the complete flow in a generic headbox
using the standard k-ε model. The result showed clear structures after the
tube bank. These gradually disappeared towards the slice. Also investigations

Figure 26 CFD calculations of flow characteristics with and without vanes (DIV/
NO DIV) in headbox nozzle; effects on turbulence scale (SCALE) and turbulent
kinetic energy (TKE) along the nozzle, X(M) [123].

474 Session 3: Forming

 Overview of Forming Literature, 1990–2000

aimed at a more detailed description of the behaviour of turbulence (water

flow) in headboxes have been carried out [125,113,112].
Parsheh and Dahlkild [126] presented numerical and experimental results
regarding the behaviour of turbulence for the case of a pure Newtonian fluid.
The experiments were performed with hot-wire anemometry and air in a
wind tunnel. Their results showed that the strong acceleration of the flow is
of major importance and that low-order turbulence models, e.g., the k-ε
model, cannot predict the turbulence in a contracting channel.
They showed that the problem originated from the underlying assumptions
in the k-ε model, which is based on an isotropic turbulence assumption, i.e.,
turbulence is the same in all directions. The flow in a contraction is, however,
accelerating and the turbulence is not at all isotropic.
This can be seen in Figure 27, which shows a graph containing both
theoretical, numerical as well as experimental results. The horizontal axis
represents the MD direction along the headbox contraction, given by the
local mean velocity divided by the value at the beginning of the contraction.
The vertical axis represents the magnitude of the fluctuation of streamwise

Figure 27 A comparison between theoretical and numerical models and

experimental results regarding the behaviour of streamwise velocity fluctuations,
normalised with inlet conditions, along a contracting nozzle; legends on the different
curves are explained in Table 3 [126].

12th Fundamental Research Symposium, Oxford, September 2001 475

B. Norman and D. Söderberg

velocity, i.e., Urms, normalised with the value at the beginning of the contrac-
tion. Table 3 explains the different plots.

Table 3 Explanations to Figure 27.

Designation Method

k-ε model Numerical

Prandtl formula Theoretical
Rapid distortion theory Theoretical
Differential Reynold’s stress model Numerical
Algebraic Reynold’s stress model Numerical
“+ signs” Experimental

Boundary layer behaviour in contraction flows, e.g., effect of turbulence

level and contraction ratio, has also been investigated by Parsheh [127].

Vanes and stratification

Flexible headbox vanes are often mounted in modern paper machine head-
boxes. These are usually present to give a more isotropic paper sheet but
could also be used for stratification (layering). Layering is not uncommon for
tissue grades, and is also used in linerboard production. There are strong
incentives to also use it in printing paper applications.
The simulations performed by Farrington [123], see previous section,
included one headbox design with flexible separation vanes, and the result
was a strong effect on turbulence behaviour after the tips of these vanes. It is
advisable to handle the results, i.e., regarding the mixing behaviour, with care
since the turbulence model that was used cannot properly predict turbulence
behaviour of the accelerating flow.
Baker et al. [128] performed experiments with headboxes aimed at stratified
forming. The results clearly showed that the vanes initiated strong secondary
flows. It was argued that the strong streak-creation, which could be visualised
by individual colouring if the layers, is a result of three-dimensional flows
after the vanes.
Lloyd and Norman performed experimental pilot machine trials aimed at
stratified forming on the EuroFEX at STFI. The experiments were conducted
by forming a paper sheet with a 3-layer headbox, see Figure 28a. All three
layers consisted of the same stock, but the middle layer was coloured blue
to visualise the layer mixing. They introduced the parameter Surface Ply
Variation or SPV, see section Paper Structure Characterisation. Small-scale

476 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Figure 28 a) 3-layer headbox; effects of relative vane length and mix-to-wire speed
on SPV (MD), b) small scale and c) large scale [129].

variations and large-scale variations were evaluated. Since these are spa-
tial variations, the direction in which they are measured is of importance
(MD or CD).
The effects of relative vane length and mix-to-wire speed difference on SPV
are shown in Figures 28b (small scale) and c (large scale). Relative vane length
is the distance from the vane tip to the slice, with a negative value when the
vane ends inside the nozzle. Longer vanes seem to give lower SPV values both
for the small and the large scales. However, tape splitting of the sheets showed
that for the two longest vanes (+ 20mm and + 100mm) the flocs in all of the
layers consisted of a mixture of white and blue fibres. For the vanes that
ended inside the nozzle (primarily −95mm and −250mm) there was instead a
clear differentiation between blue and white flocs. Hence, by changing the
vane length, an optimum can be identified where the tip of the vane ends right
before the slice lip, where SPV has the lowest value, without “total” mixing.
Three headbox slice geometries in combination with different vane lengths
were also tested [30]. A parrot’s beak slice lip was compared to a linearly
converging slice lip and a plane channel slice lip. With the vanes ending inside
the nozzle, the parrot’s beak appeared to reduce layer mixing, while the
parallel channel appeared to increase it. However, the type of slice lip had
little effect on layer mixing if the vanes ended outside of the headbox; vane

12th Fundamental Research Symposium, Oxford, September 2001 477

B. Norman and D. Söderberg

tip vortices then seemed to be the dominating source of mixing. The parrot’s
beak also improved the small-scale formation.
Finally, Lloyd and Norman also studied the effect of vane shape [130]. In
one set of experiments, a 5 mm-thick vane was followed by a 0.5 mm thin-vane
of different lengths, 10–40 mm. The step height was thus 2.25 mm. In another
set of experiments, the steps were varied from 0.5 to 1.75 mm. The aim of the
steps was to break down the boundary layers formed upstream along the
vanes, and thus reduce downstream vortex generation. However, the turbu-
lence introduced by the steps increased the layer mixing, both at floc and fibre
level. This turbulence gave lower fibre orientation anisotropy, worsened
small-scale formation and Z-toughness, but improved fracture toughness of
the formed sheets.
Lloyd summarised the work on layering in a Tekn.lic. thesis [19].
Using flow simulations Parsheh and Dahlkild [131] numerically studied the
mixing behind vanes with different tip shapes and positions. The simulations
were performed with the k-ε turbulence model. The geometry was similar to
that of Lloyd and Norman [129]. In the simulations, which were two-
dimensional, the mixing was studied by adding a passive scalar into one of
the layers. The spreading of this scalar was then studied. The shortest vane
gave the lowest mixing at a relevant distance downstream in the jet. Also, hot-
wire studies of mixing behind a vane in a contraction were performed by
Parsheh [127].
The vanes in the headbox are not straight but will deform as a function of
the flow field and the bending stiffness of the vanes. Parsheh and Dahlkild
[126] modelled this problem. They developed a quasi one-dimensional model
describing the position and shape of the vane as a function of bending stiff-
ness, vane and nozzle lengths and flow conditions at the inlet. The model was
compared with direct simulations using the commercial software CFX to
model the flow and the position and shape of the vanes. This was performed
using an iterative technique.
Söderberg [7] performed a pilot machine study of the effect of contrac-
tion on headbox jet flow, a study not specifically aimed at stratified forming.
In the study the same stratified headbox as that of Lloyd and Norman
(Figure 28) was used. The length of the nozzle and vane was kept constant
and the inlet area was varied. Each layer in the headbox was given an
individual colour. The top layer was white, the middle layer was blue and
the bottom layer was red, and the headbox was mounted at the fourdrinier
position in the EuroFEX pilot machine. The headbox jet was visualised at
several MD positions after the slice and these visualisations were compared
to the paper produced, which was scanned using an ordinary desktop
scanner.

478 Session 3: Forming

 Overview of Forming Literature, 1990–2000

The influence of headbox contraction ratio can be seen in Figure 30. The
top row is scanned paper sheets where the blue colour (centre layer) has been
coded as black. The bottom row contains images showing the headbox jet at
the three different contraction ratios. It is clear that the characteristics of the
jet are reflected in the structure of the paper sheet. The low contraction case
gives a strong mixing of the different layers, as does the high contraction case.
However, the dynamics behind the mixing seem to be very different, which is
obvious from the images of the jet.
Li et al. [132] studied stratified 3-layer headbox flow by an experimental
technique where salt was added to the centre layer and the conductivity pro-
file was measured across the jet thickness. In Figure 31 the left illustration
shows the principle, where the central layer has a higher conductivity. Due to
layer mixing there is a spreading (relaxation) of the conductivity profile
downstream. The overall mixing was summarised in a single mixing par-
ameter IMH. The right part of the illustration shows some conductivity pro-
files along the jet. They investigated the effect of slice opening, jet speed and
vane tip position on the degree of mixing. Mixing was reduced at higher slice
opening and flow velocity.
Li et al. [133] also studied the mixing phenomena in a 2-layer headbox,
using the same technique. In principle, the results obtained agreed with those

Figure 30 The effect of headbox contraction ratio (CR) on jet behaviour and layer
mixing. Scanned paper sheets (top) and images from the headbox jet (bottom).

12th Fundamental Research Symposium, Oxford, September 2001 479

B. Norman and D. Söderberg

Figure 31 Conductivity analysis of a jet from a 3-layer headbox (Left) and jet
conductivity profiles at 25 to 254 mm from slice opening (Right) [132].

for the 3-layer headbox. They also performed visualisations of the mixing in a
3-layered headbox using coloured water. After a comparison with the layer-
ing in the paper produced it was concluded that the main part of mixing takes
place in headbox and jet, and not during dewatering.

4.4 Headbox jet and streaks

The origin of streaks in the final paper is the flow conditions in the headbox
and jet; see e.g., Norman [134]. MD aligned streaks can be coherent vari-
ations in grammage, formation and fibre orientation. In the case of layered
forming, it can also be streaks in layer mixing. By considering the hydro-
dynamics of Newtonian fluid flows, possible origins of the streamwise streaks
in a liquid jet can be identified.
Four streak-generating mechanisms are illustrated in Figure 32. These are:

1. Vortex stretching, which is a result of the accelerating flow, will give vor-
tices aligned in the streamwise direction. This is a well-known mechanism
and the origin of vorticity can, for example, be the upstream conditions in
the nozzle or at the exit from the tube bank, see Hauptmann et al. [111].
2. Centrifugal instabilities (vortices) in the flow, which are caused by stream-
line curvature, see e.g., Matsson and Alfredsson [135]. These could be
so-called Dean or Görtler vortices. Robertson and Mason first proposed
the existence of centrifugal instabilities in the headbox flow.
3. Streaky structures, which grow inside shear flows due to hydrodynamic
instabilities could also occur at solid surfaces in the headbox, see e.g.,
Alfredsson and Matsubara [136].

480 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Figure 32 Streak creation mechanisms, which possibly could be present in headbox

flow. a) Vortex stretching, b) centrifugal instability, c) boundary layer transition and d)
wave instability [138].

4. Instability waves, which originate from velocity profile relaxation in the

jet, Söderberg and Alfredsson [137]. Coupled to these waves is a break-up
of the jet, which generates dominating streamwise streaks. The waves are
found both for a plane water jet at low Reynolds numbers and in headbox
flow Söderberg [7].

Typically the width and/or height of streaks are of the order of the shear
layer thickness or the smallest size of the geometry perpendicular to the flow
direction. Hence, in wall-bounded flows the origin of streaks is usually closely
connected to the shear layer and thus has a size similar to its thickness. For
the case of a boundary layer flow the streaks typically will have a size com-
parable to the boundary layer thickness and for the case of fully developed
channel flows, i.e., the shear field covers the height of the channel, they will
have a size comparable to the channel height.
If the origin of the streaks is not coupled to shear fields, but instead invis-
cid mechanisms such as vortex stretching, the flow geometry itself will set the
size of the streaks. One example of this could be the flow in a nozzle (head-
box), where the final vortex diameter will be of the same size as the slice
thickness.
Söremark et al. [139] eliminated streaks on a paper machine by inserting a
small triangular bump on the bottom wall of the headbox contraction. This
small step would disrupt flow patterns (streaks) closest to the wall and due to

12th Fundamental Research Symposium, Oxford, September 2001 481

B. Norman and D. Söderberg

the small height of the bump it most probably altered the flow inside the
boundary layer. This method of streak prevention was earlier proposed by
Chuang [121] and can be seen in Figure 24 (centre), which shows a bump
mounted on the bottom nozzle wall.
Aidun and Kovacs [140] numerically simulated the flow of water inside a
headbox and identified the junctions between the converging nozzle walls and
the sidewalls as a source of secondary flow patterns, e.g., streaks.
A detailed experimental investigation of streaks generated by secondary
flow behind vanes was performed by Baker et al. [128]. They visualised the
flow patterns with Laser Induced Fluorescence and the images obtained
clearly showed MD streaks as well as CD structures/streaks.
In order to investigate the behaviour of the headbox jet at elevated speeds
(40 m/s) Schlupp et al. [141] performed visualisations of high-speed headbox
jets. These visualisations clearly showed streaks in the jet and the images
representing the jet surface were quantified regarding surface roughness
scales as a function of jet speed. This only showed a weak dependence of jet
speed. Experiments regarding streak generation by perforated plates, vanes
and slice-lip configuration were also performed; see Schlupp [142].
Lindqvist [143] made experiments concerning the creation of streaks using
water in a model headbox. The experiments were performed with Laser
Doppler Velocimetry and different visualisation methods. The visualisations
clearly showed the presence of large-scale structures in the flow, i.e., vortices,
which were generated by the tube bank. The measurements were performed
with water and showed the development of the boundary layer in the con-
verging headbox nozzle.
In order to understand how the flow from the headbox affects the jet it is
necessary to perform experiments as well as to make numerical simulations of
the jet flow. For the case of free jet flows, the simulations are a challenging
task since the location of the jet surfaces are not known, but are obtained as a
result of the simulation.
Li et al. [144] performed 2D simulations of the jet flow from a converging
nozzle.
In order to understand the behaviour of the streaks on the forming table
Aidun [145] modelled the flow from the slice and onto the forming table. This
included the flow of the jet. Also, the flow was quantified using high-speed
imaging and image processing methods, Aidun [146]. An example can be
found in Figure 33. The image is a spatio-temporal representation of the mix
on the wire. The vertical axis is the same as the CD and the horizontal axis
represents the time. The image clearly shows streaks in the mix surface.
In order to quantify flow visualisations of a headbox jet Söderberg [7] used
image processing methods, i.e., wavelets, to capture the behaviour of the jet

482 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Figure 33 Spatio-temporal representation of the jet surface. The vertical axis

represents CD and the horizontal represents the time.

and mix on the wire. These visualisations were performed with transmitted
light as well as light reflected from the surface. The transmitted light visual-
isations were also used to obtain velocity distributions in CD/MD plane
using the so-called PIV technique. These showed streaks also in the velocity
distribution. Similar methods were used by Ono et al. [147] with the differ-
ence that they used light reflected from the suspension surface.
Söderberg and Alfredsson [138] have shown that streaks are not only ori-
ginating from inside the headbox nozzle, but can be created inside the jet
itself. This can be seen in Figure 34. The figure shows the flow of a plane
water jet at low Reynolds numbers visualised with the so-called shadowgraph
method. The velocity profile inside the jet gives rise to a hydrodynamic
instability (wave instability), which can be seen directly after the nozzle exit;
Figure 34a.
In Figure 34b, the velocity of the jet has been increased and the waves are
present only closest to the nozzle exit. After this short region of waves, the jet
experiences a breakup which partly causes a disintegration of the jet and also
gives rise to MD aligned structures. These have a CD spacing larger than the
waves and much larger than the jet thickness. The difference is one order of
magnitude.
Figure 34c shows a close-up of the break-up and the streak generation
inside the jet, which is visualised using reflective particles seeded into the
water. These particles are flat and thus they tend to orient with the shear in
the flow. Before the break-up the particles cannot be seen at all. This indicates
that there is no strong orienting shear before the break-up. After the break-up
clear white streaks can be seen, oriented in the MD direction and thus
implying an onset of strong shear, i.e., streaks.

12th Fundamental Research Symposium, Oxford, September 2001 483

B. Norman and D. Söderberg

Figure 34d shows a particle visualisation at the same jet thickness and
speed as in Figures 34bc. However, the nozzle is in this case formed as a slit
and the flow can be described as inviscid. The viscosity is then negligible and
boundary layers are not present.
The MD streaks seen in this image originate from within the nozzle and are
an effect of the vortex stretching mechanism, and the strong acceleration
tends to increase the strength of this vorticity and thus creates MD vortices.
This wave instability has also been shown to be present in full-scale head-
box flow, with water as well as with pulp suspensions; Söderberg [7]. In this
investigation, images taken of the jet were evaluated using a Morlet wavelet
transform; see Figure 35.
The transform is applied in MD. For one individual image (left) the wavelet
map is processed along 0.5 mm-wide MD-strips. The wavelet maps are then
averaged over all strips, which gives a spatial mean wavelet map. Further, in
this map all peaks along the wave number axis were recorded to form a peak

Figure 34 Visualisations of a 1 mm-thick plane water jet. a) Shadowgraph, channel

jet at 1.5 m/s, b) shadowgraph, channel jet at 2.0 m/s, c) particles, close-up, channel jet
at 2.0 m/s and d) particles, slit jet at 8.0 m/s.

484 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Figure 35 Left to right: Images at 5 m/s using reflected light, mean wavelet transform
and cumulative peak detection; a) water jet and b) mix jet with 0.35% softwood fibres.

event map. Finally, for both wavelet transform and peak event map averaging
was made from a sequence of 50 images.
The mean wavelet transform map (centre) gives an indication of the
dominant structures, while the cumulative peak detection map (right) shows
sub-dominant behaviour with high probability (could also be called peak
probability map).
The result, which can be seen in Figure 35, represents a jet velocity of 5m/s
and the top and bottom graphs are given by water and pulp suspension
respectively. The visualisation in this case is aimed at larger structures in the
flow; hence the image of water jet (top) does not show any presence of waves.
The top of the image is at a MD distance of 20 mm from the headbox and the
jet break-up can be seen. The white dots in the image are a result of this jet
break-up. If velocity is increased more dots will be visible.
The dots were, by visual inspection, found to be drops leaving the jet sur-
face as a consequence of the strength of the break-up. A ridge in the mean
wavelet map with a maximum at ∼4 mm can be clearly seen, which can be a

12th Fundamental Research Symposium, Oxford, September 2001 485

B. Norman and D. Söderberg

result of both the waves and the drops. It is also present in the cumulative
peak detection map.
The three graphs in Figure 35b show the flow of a fibre suspension at the
same parameters. In the image of the jet the fibre suspension (flocs) can be
clearly seen. Waves are present in this image, but are very hard to observe.
However, at higher velocity they could be clearly seen. The contour plot
shows a maximum at a wavelength of ∼20 mm, which is given by fibre floccu-
lation and in the peak map a maximum is found for wavelengths ∼4 mm,
which is a result of the surface waves.

A short note on turbulence and the flow properties of a fibre suspension

The rheology of a fibre suspension is extremely complicated. For low pipe
flow velocities, pulp flows as a plug. When the flow rate is increased the shear
at the walls causes disruption of the network close to the walls into network
fragments (flocs) and/or fibres. This state is often referred to as turbulent, but
in order to prevent confusion with turbulence in an ordinary fluid, fluidised
could be a better term. The fluidisation can be both at floc level and at fibre
level. The centre plug is gradually broken up as the pipe flow velocity is
increased, and finally the flow could be completely fluidised.
This does not necessarily mean that turbulence appears in the same manner
as for single-phase linear (Newtonian) fluids. For these, turbulence is charac-
terised by a continuous spectrum of scales in the flow, ranging from the most
energetic large scales to the smallest micro scales where energy is dissipated
into heat by viscosity. If fibres are added, non-linear effects start to appear
and the apparent viscosity will depend on the flow conditions. Due to the
fibres and flocs the available scales are fewer and the spectrum is not continu-
ous. The smallest scales are most probably of the same size as the fibres, since
the presence of the fibre will suppress scales in the flow, which are smaller
than the fibre size. Fibre flocs will also limit the scales, since the flow within a
fibre floc will probably not be turbulent (due to the small velocity difference
between the floc and fluid).
The dissipation of turbulent kinetic energy within the fibre suspension will
also differ from the case of a single-phase fluid. It has been demonstrated that
strong agitation of a fibre suspension generates network strength. Hence
energy is transferred from the larger scales into the fibre network, where some
energy will dissipate through fibre-fibre friction. The discussion of the flow in
a headbox can be performed using the word turbulence, if turbulence only
implies a non-deterministic, i.e., chaotic, behaviour of the fluctuations. Due
to the presence of fibres and fibre flocs, this turbulence will however not have
the same characteristics as turbulence in a Newtonian liquid.

486 Session 3: Forming

 Overview of Forming Literature, 1990–2000

A “true” Reynolds number Re = ud/v cannot be calculated for a fibre sus-

pension flow since viscosity is not well defined. An approximation can always
be made using a suitable apparent viscosity. Such an estimate will give a
viscosity of the fibre suspension, which easily is one order of magnitude
higher than the viscosity of water. Therefore, results obtained at low Rey-
nolds numbers using Newtonian fluids can also be relevant for flows of fibre
suspensions at higher velocities. If, for example, pure water is considered to be
the fluid in the paper manufacturing process Re ∼105 but for the model
experiments performed by e.g., Söderberg and Alfredsson [138,148] Re ∼103.
The effect of Reynolds number may still be a second order phenomenon
when compared to the presence of fibres and fibre flocs, which will have a
strong influence on the liquid behaviour.
As already mentioned in the section Fibre suspensions, the break-up of
fibre flocs has historically been considered to be promoted by turbulence. It is,
however, more likely that the best way to disperse fibre flocs is by elongational
flows. Duffy and Norman [50] proposed this and performed experiments with
different types of contractions in a cylindrical geometry. The result showed
floc break-up for TMP pulps but not for chemical pulps. Corresponding
results were obtained in the PhD thesis by Nordström [115]. Using an
increased elongational headbox nozzle flow, the formation of twin-wire roll
formed paper was significantly improved for TMP pulp, but not for chemical
pulp.

5 WIRE SECTION DESIGNS

5.1 Roll-blade dewatering
During the late 1980s, partial roll forming followed by suction shoes had been
introduced for new printing paper machines. Together with the opposite
alternative, initial suction shoe or blade dewatering followed by roll dewater-
ing, these were expected to dominate newsprint machines in the US during
the 1990s [149].
During the 1980s, Dörries had developed a new twin-wire design for board
making, with fixed blades on one side, and adjustable blades on the other
side of the wires [150]. Voith, who bought Dörries, continued to apply the
design to different board machines as the D-former. Ahlström later produced
an identical design in their Alform MB former. Although this was a clear
patent infringement, Voith did not object, although they later declared that
this has been a mistake. Two backgrounds made an influence: Firstly, it was
initially not their own design; secondly, the basic principle for dewatering
with this former was not well understood at the time. Eventually Valmet

12th Fundamental Research Symposium, Oxford, September 2001 487

B. Norman and D. Söderberg

acquired Ahlström, and their former became the “Valmet SymFormer MB”
unit.
The D-former was initially considered to generate a gradually thinning
space between the wires, through a gradually increasing force applied to the
bottom blades along the forming zone. This was supposed to make it possible
to gradually increase the dewatering pressure along the fabrics. In a presenta-
tion at the1989 Cambridge Symposium [1] it was explained for the first time
that instead of the gradually increased dewatering pressure suggested by
Dörries, a pulsating pressure was instead generated, by a zigzag movement of
the wires between top and bottom blades.
It was further suggested that the Dörries principle could have broader
applications:
The new principle has been applied to board making. It seems to have a
large potential also for high-speed papermaking. [1].
This basic principle of pulsating dewatering according to Figure 36 was
adapted in combination with roll dewatering in the STFI-Former. This was
initially presented at the PIRA Conference New technologies in multiply and

Figure 36 Concluding slide from 1989 Cambridge Symposium [1].

488 Session 3: Forming

 Overview of Forming Literature, 1990–2000

multilayer structures, one week before the start-up of the STFI-Former on the
EuroFEX pilot machine in June 1991 [151]. After some initial roll dewater-
ing, the fabrics were guided vertically upwards. Fixed blades were mounted
along the vertical tangent to the forming roll, against which the inner fabric
was run. On the opposite side, adjustable blades were mounted in a zigzag
manner; see Figure 37 [152].
In twin-wire forming, it is important to arrange the dewatering in a way
allowing automatic fabric position adjustment at changes in fabric separation
(stock quality or production changes) without significant changes in dewater-
ing pressure events. Therefore, the adjustable blades were mounted against
the movable outer fabric, and not against the inner fabric, the position of
which is initially defined by the forming roll surface.
One of the important development aims for the STFI-Former was three-
layer forming, and the idea was to form the outer layers during gentle
dewatering over the forming roll. This was followed by formation improving
pressure pulses that could break apart fibre flocs still remaining in the middle
layer.
It is important that all stages of dewatering are symmetrical, which is
the case when a suitable vacuum is applied at the roll surface and when the
blades are mounted symmetrically on both sides of the fabrics. The amount
of roll dewatering could be controlled by an adjustable cover angle along the

Figure 37 The STFI-Former with symmetrical roll dewatering followed by

symmetrical blade dewatering with adjustable blade forces F1–11.

12th Fundamental Research Symposium, Oxford, September 2001 489

B. Norman and D. Söderberg

forming roll, and the individual blade pressure pulses could be controlled by
blade force adjustments.
One principal difference between pure roll dewatering and roll-blade
dewatering is the situation when the fabrics leave the roll. In roll-blade
dewatering, contrary to what is the case in pure roll forming, there is still mix
left between the two partially formed webs when the fabrics separate from the
roll surface, and the dewatering pressure disappears. This pressure drop will
cause an increase in mix velocity [153], see Figure 38.
At the same time, a critical table roll suction pulse will be generated when
the wires leave the roll surface. This may create a local velocity pulse as
indicated in Figure 38. If this pulse is strong, and the grammage of the web
on the inside of each wire is too low, sheet damage may arise. This mechanism
limits how small cover angle on the forming roll that can be used in practice.
It should be pointed out that in reality, the mix velocity during the roll
dewatering phase will not be generated by the long, constant amplitude pres-
sure pulse indicated in Figure 38, since the pressure build-up along the roll is
much more gradual; see further below.
Influence of gravity will mean that the mix velocity along the vertical
section at an average speed of 1200 m/min will decrease by nearly 60 m/min
for 2 m increase in height level.

Figure 38 Example of the variations in speed difference between mix and wires when
viscous effects are restricted to thin boundary layers. Broken lines indicate blade pulses.
Wire speed 1180 m/min, jet speed 1200 m/min, wire tension 6 kN/m, roll radius 0.8 m.

490 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Voith developments
When the new Voith pilot machine in Heidenheim was started up in 1990
[154], it was set up for twin-wire blade forming, a technique which was pro-
moted by Voith at the time [155]. At the same time the development of the
STFI-Former was taking place, and thinking along the same lines, based on
the Dörries patent, also started at Voith. In a patent application from August
1989 [156], a roll-blade former is one of the alternative designs; see Figure 39.
However, in the patent there is no discussion of blade pressure pulses, and the
wires are described to gradually converge along the D-zone (marked II in
Figure 39).
Late in 1991, about one year after the start-up of the new pilot machine, a
rebuild was started to what eventually became the Douformer CFD unit
(“D” from Dörries) [157]; see Figure 40.
Besides the 45-degree angle against a vertical direction of the wire package
leaving the forming roll, there is one major difference from the STFI-Former.

Figure 39 Voith patent of roll blade former including adjustable blades [156].

12th Fundamental Research Symposium, Oxford, September 2001 491

B. Norman and D. Söderberg

Figure 40 Roll-blade Douformer CFD forming unit by Voith with loadable blades
against inner wire.

The fixed dewatering blades in the suction shoes are mounted against the
outer wire.
This means that the front-end position of the first shoe should be adjusted
depending on the wire separation distance, leaving the roll, if the same initial
pressure pulse should be obtained. However, the shoe position is not adjusted,
once the suction shoe has initially been installed according to instructions.
Application of the Duoformer CFD to printing papers indicated improved
formation and retention in comparison to previous designs [158,159].
The Duoformer CFD was later modified to DuoFormer TQ (Total
Quality). A main objective was to increase web dryness entering the press
section, which was obtained by increasing the wrap angle of the top (outer)
wire around the couch roll, and by the introduction of an extra suction box
after the couch roll [160,161]. A further development is the DuoFormer TQv,
where “v” indicates a vertical wire run, leaving the forming roll [162,163].
Early 2001, Moser [164] described the main Voith twin-wire forming
designs, together with a discussion of important paper properties. A total
number of 190 DouFormer D units were reported to have been delivered at
the time.
Voith also suggested a simplified version of roll-blade forming by position-
ing some adjustable blades against the outer wire on a roll former; see Figure

492 Session 3: Forming

 Overview of Forming Literature, 1990–2000

41. This design was tested on an industrial roll former producing newsprint.
A clear improvement in large-scale formation was obtained, like that with
blades placed after the roll.
However, this concept breaks a fundamental rule of twin wire forming:
“Never physically control the positions of both wires at any location along
the dewatering section”. What happened with time was a gradual increase of
grammage streaks.
This was caused by the slight friction wear of the blade surfaces, which in
turn generated local changes in roll-to-blade distance across the CD and thus
also changes in local wire separation across the machine, causing grammage
streaks. When a blade is applied against one of the wires, the opposite wire
should be able to automatically adjust its position to compensate local blade
surface deviations.

Figure 41 Adjustable blades mounted against forming roll outer wire. Voith
patent [165].

Valmet developments
Valmet considered the application of the Sym-Former MB unit with adjust-
able blades, after some initial roll forming, and a patent was applied for in
May 1989 [166]; see Figure 42.
The blade section was designed according to the Alform MB principle for
Fourdrinier machine rebuilds, which Valmet had inherited in the acquirement
of Ahlström. In the patent there is, however, no disclosure of pressure pulses
generated by the adjustable blades; wire convergence in the zone between the
stationary and the adjustable blades is described as “mainly linear”.

12th Fundamental Research Symposium, Oxford, September 2001 493

B. Norman and D. Söderberg

Figure 42 Roll-blade former according to Valmet patent [166]. Item “127” includes
the loadable blades against inner wire. This is basically the SpeedFormer MB design.

However, serious developments on the pilot machine in Jyväskylä did not

start until 1993, and in 1994 the first Speed Former MB unit was installed as a
rebuild of a fine paper fourdrinier machine [167]. In comparison with
fourdrinier operation, the new forming unit gave better formation, two-
sidedness and sheet bending stiffness.
New installations were designed with a vertical fabric arrangement in the
MB unit; see Figure 42. Like in the corresponding Voith design, the suction
shoes were mounted against the outer fabric, with the inherent requirement
of adjustment of the front position of the first suction shoe according to the
prevailing dewatering situation.
In the latest Valmet forming section design, the “OptiFormer”, the loadable
blade unit is located as close to the forming roll as possible; see Figure 43.
This arrangement is claimed to minimize the time for reflocculation in the
centre layer of the web, and thus maximize the formation potential.
Pilot plant tests have shown the OptiFormer arrangement to give better
sheet formation than SpeedFormer MB. One major difference between
the earlier SpeedFormer MB and the new OptiFormer is, however, that in the
latter case, the loadable blades act against the outer wire, while they in the

494 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Figure 43 Forming roll followed by suction shoe against the inner wire and loadable
blades against the outer wire in the Valmet Optiformer [168].

former case were acting against the inner wire. Thus the fixed blades, like in
the STFI-Former (Fig 37), are now acting against the inner wire, the entering
position of which is always defined by the roll periphery, independently of the
separation between the two wires. Therefore no adjustment of the position of
the first fixed blade is needed, when running conditions are changed. This is
most probably the main reason why better formation is obtained on the
OptiFormer.

Beloit developments
The traditional design of the Beloit Bel Baie machines included initial
dewatering over a forming shoe consisting of nine blades. In 1995 Bel Baie IV
design was presented, in which the initial dewatering takes place over three
wide (ca 150 mm) blades with large radius of curvature [169]. This introduces
a gentler initial dewatering in comparison with the sharp pulses over the
traditionally narrow (ca 15 mm) Bel Baie blades. It can be seen as a replace-
ment of initial roll forming, with the drawback of only one-sided dewatering
as well as considerable friction against the contacting wire.
A top wire unit for rebuild of fourdrinier printing paper machines, consist-
ing of curved, inverted suction boxes on top and pneumatically loadable
blades below the wires was described in 1995 [170].

12th Fundamental Research Symposium, Oxford, September 2001 495

B. Norman and D. Söderberg

A further development of the Bel Baie IV unit is reported in 1997 [171]. An

initial for Finally, in 1998, the Bel Baie RCB (Roll and Counter Blades) was
developed [172], a roll-blade former of similar design as those by STFI, Voith
and Valmet. According to Crouse, the roll-blade former at the Rockton pilot
plant was successful in manufacturing all qualities from printing paper to
linerboard [173].
However, this introduction was too late, and no order was taken before the
company went out of business early 2000. During the 1980s three new news-
print machines were installed in Sweden, all three Bel Baie blade former
designs. The new printing paper machines installed in Sweden during the
1990s were of the roll-blade type. The inability of Beloit to deliver a com-
petitive forming section during the 1990s was possibly one reason for their
unsuccessful business operations during the last decade.
Mitsubishi, a prior Beloit licensee, has introduced a new blade holder, with
on-line control of the blade angle against the wires [174]. Some angular open-
ing between blade surface and contacting wire on the upstream side is shown
to have a positive effect on paper formation.

5.2 Multi-ply board forming

Board products have traditionally been formed from separate plies, succes-
sively couched together [175]. The ply forming can be performed using
separate fourdrinier wires.
Pressure formers are still used in different configurations, and are often
installed at machine rebuilds [176].
Twin-wire forming of the individual plies is applied increasingly. The
Dörries principle described above has widespread use as the Voith D-former,
as well as the corresponding Valmet Symformer MB design. Beloit has
developed the Inverform technique [1] by the introduction of adjustable
counter blades facing the inverted suction shoe [170].
Recently, also, separate twin-wire roll-blade forming units for board plies
have been applied [177,178]; see Figure 44. Several of these units can be
applied to add individually formed plies on top of the base ply. In principle
the same former design is also applied to form the base ply [179].

6 WEB FORMING
6.1 Random sheets
The sheet formation number FRandom (coefficient of variation) for a sample
with identical, randomly distributed fibres is determined by the square root

496 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Figure 44 Top-ply roll-blade board forming design; the Voith DuoformerTop [177].

of the inverse mean number of fibres at a point in the plane of a paper sheet;
see Equation 5:

冪
wFibre
FRandom = (5)
wSheet

where wfibre is fibre and wsheet is sheet grammage respectively [180].

This equation follows from a Poisson distribution of the number of fibres
at any specified point in the sheet plane.
Several conclusions can be drawn from this simple equation:
• Formation number is independent of degree of fibre orientation
anisotropy.
• Formation number is independent of fibre length or width.
However, if the complete formation wavelength spectrum is considered,
longer fibres will mean a shift towards longer wavelengths and a more
oriented sheet will give a shift towards longer wavelengths for the spectrum in
the orientation direction and the opposite in the perpendicular direction.
In practice, formation is not evaluated from zero to infinite wavelength,
and especially the lower limit is considerable, relative to e.g., fibre width.
Therefore, when influence is found on formation levels of random sheets by
fibre length or fibre orientation, this is often due to geometrical limitations of
the evaluation method [20,181,182,183,184].

12th Fundamental Research Symposium, Oxford, September 2001 497

B. Norman and D. Söderberg

Sometimes the effect of fibre orientation on paper formation is evaluated.

According to the above, no such general relationship exists. However, some
process changes will affect both formation and orientation, but not in a
unique way.
Dodson et al. attempted to model paper formation as a result of a random
distribution of flocs in the fibre suspension, entering the wire section [185].
Random distributions of stars [186] and discs [187,188,189] have been sug-
gested, but it has been pointed out that the results indicate unrealistic floc
grammage levels [190].

6.2 Self-healing effects

It was suggested by Wrist during the 1960s that fibre distribution in a real
paper could be more even than that of a random sheet [1]. The reason would
be that there is an inherent self-healing effect in the dewatering process. If a
hole is present in the web on the wire during the dewatering phase, the local
dewatering resistance will be low, and excess fibre suspension will be
dewatered at that position. Thus extra fibres will be deposited, and the overall
basis weight will be evened out. Norman et al. later studied this by analytic
evaluation of the formation spectrum of a random sheet and comparison
with real sheets using beta radiography evaluation of well-formed laboratory
sheets [1]. The results confirmed that the real sheet is more even than the
random sheet in the small and medium floc size range. At larger floc sizes,
the real sheet is, however, more uneven, due to fibre flocculation effects, by
definition not present in the random sheet.
Sampson et al. introduced the term “hydrodynamic smoothing s”, where s
varies from 0 to 1 with increasing self-healing effects [191]. For laboratory
handsheets formed at a range of grammage and crowding factor, s was
found to be positive in all cases, confirming that the self-healing effect exists.
Formation values at low grammage with 1 mm2 resolution were high in com-
parison with random fibre distribution, indicating a locally uneven drainage
resistance in the forming wire with its stabilising backing.
Norman et al. [192] confirmed that the real sheet is more even than the
random sheet at wavelengths smaller than about 10 mm using up-to-date
radiography analysis equipment [8], and comparing with computer simulated
random sheets.
Normalised formation numbers for handsheets formed at constant con-
centration was shown to improve with grammage up to ca 60g/m2, above
which forming concentration had to be increased due to the limiting height of
the forming vat; see Figure 45. This improvement would be a result of the
self-healing effect.

498 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Figure 45 Normalised formation number as a function of grammage for laboratory

sheets; increased forming concentration above arrow [192].

Tensile strength index increased up to 60 g/m2; see Figure 46. Although

some of the strength increase is caused by a decreasing relative influence of
the less bonded sheet surfaces, the self-healing effect is thought to be a main
effect.

Figure 46 Tensile strength index for single- and double-layered sheets. Laboratory
sheets of long fibre pulp; increased forming concentration above arrow [192].

Self-healing was further demonstrated by the comparison of one solid

60 g/m2 sheet and one 60 g/m2 sheet made from two 30 g/m2 sheets couched
together. The solid sheet was significantly stronger. This indicates that paper
made with one-sided dewatering, due to the larger self-healing effect has a

12th Fundamental Research Symposium, Oxford, September 2001 499

B. Norman and D. Söderberg

higher strength potential than the corresponding twin-wire formed sheet,

which consists of two separately dewatered halves.
It was demonstrated that the normalised formation number of Formette
Dynamique formed laboratory sheets improves with increased grammage. In
this case, forming concentration will not increase at high grammages. It was
further verified that the formation numbers are equal in MD and CD for
anisotropic laboratory sheets, as predicted by equation (5); see Figure 47.
Lucisano and Norman [193] tried to avoid the self-healing effects in stand-
ard handsheet forming by extreme reduction of the dewatering velocity (4 h
dewatering time). To avoid fibre flocculation and fibre settlement from grav-
ity, the water was made viscous and the fibres naturally buoyant by the add-
ition of sugar (60% by weight). 20 g/m2sheets were formed, and compared to
the corresponding sheets with normal drainage in pure water. To avoid peri-
odic, wire generated drainage resistance effects, a flat, porous polymer plate
was replacing the wire. The formation wavelength spectrum of the quasi-
random sheet was close to that of a random sheet. The crossover point was
at 0.5 mm, with the quasi-random sheet slightly lower at short and higher
at large wavelengths. It was found that the “quasi-random” sheets had
significantly lower tensile properties than the conventional sheet.
The formation improving self-healing effect was further demonstrated by
free gravity draining of the suspension with sugar solution, whereby the
shorter drainage time, 20 minutes in comparison to 4 hours, generated
a ca 50% lower formation spectral density over the whole wavelength range.

Figure 47 Normalised formation numbers measured in MD (filled symbols) and CD

(unfilled symbols) as a function of grammage for anisotropic Formette Dynamique
laboratory sheets of short- and long-fibre chemical pulps and TMP respectively [192].

500 Session 3: Forming

 Overview of Forming Literature, 1990–2000

6.3 Formation improving mechanisms

It is of vital importance that the fibre distribution in the web is as uniform as
possible so that the best formation is attained. Traditionally, according to
Parker [1], there were three main ways of improving the fibre distribution on
the wire:
• dewatering;
• oriented shear; and
• turbulence.
It is now suggested that four effects are important, to improve sheet
formation; see Figure 48.

Figure 48 Principles for improving the evenness of fibre distribution in web:

“turbulence” elongational flow, dewatering and oriented shear [3].

1. Turbulence in the traditional sense can probably not be present in a fibre

suspension, as discussed in section Headboxes. To have turbulence pres-
ent also during the dewatering phase would have a dominating, negative
effect on the evenness of fibre deposition. It is therefore motivated to
move turbulence effects to the headbox, and make sure that they have
decayed when the jet reaches the wire/s.
2. Elongational flow is developed by the acceleration in the headbox nozzle,
and an improvement in paper formation can be obtained using a high
nozzle contraction ratio (see Figures 71–72). Possibly, formation improv-
ing extensional flow could also be generated by the flow in the headbox
tube bank and by the pressure pulses in twin-wire blade forming.
3. Dewatering has a self-healing effect on local variations in grammage, as
described above.

12th Fundamental Research Symposium, Oxford, September 2001 501

B. Norman and D. Söderberg

4. Oriented shear has a well-established positive effect on formation,

although there is also an example where this is not the case (see Figure
72). To avoid floc rotation in a shear field, it might be important that the
flocs are fixed to the surface of the forming web by the drainage forces, to
improve the floc-breaking efficiency of a shear field. “Activity” on a
Fourdrinier wire could be considered as shear, rather than “turbulence”
effects.

6.4 Fibre deposition on wire

The geometry of the forming fabric is of vital importance for the surface
structure of the sheet formed, and thus also for the dewatering capacity.
An overview of forming fabrics is given by Kilpeläinen et al. [194].
Johnson [195] studied the effect of jet impingement in a twin-wire blade
former. It is pointed out that the wire on which the jet is landed has a domin-
ating effect on the wire mark in the final paper product.
Fejér discusses the influence of fabric design on formation and paper struc-
ture with Fourdrinier and twin-wire forming [196].
Adanur [197] studied the effects of wire design on sheet properties in
laboratory forming. Drainage index, fibre support index and number of yarns
(MD × CD) were used as fabric parameters. Considerable influence on paper
properties was reported, but to what extent this was caused by different
retention levels is unknown (pulp type not reported).
Hampson [198] points out that fabric surface aperture size has to be chosen
with consideration of the fibre length distribution. High fabric density at the
top surface avoids channel blocking in the interior of the fabric and sheet
sealing. This promotes high dewatering rates during the early phases.
Kilpeläinen et al. [199] compared different fabric designs in twin-wire form-
ing of newsprint at speeds up to 2000 m/min. Triple layered fabrics gave 3%
higher retention than double layered, but slightly lower dewatering rates.
However, both fabric designs could handle the increased amounts of water
brought into the former at higher speeds. Changes in fines distribution were
small.
It is pointed out that the interweave pattern between the layers in a triple-
layer fabric is critical, if internal wear is to be minimized [200,201].
Herzig and Johnson [202] designed a closed water flow loop in which the
desired amount of fibres can be injected and a web collected on an inserted
fabric, at dewatering rates similar to those during initial forming. Figure 49
shows the fibre deposition at four grammage levels. It is obvious that the
fabric structure has a big influence on the surface layer of the paper. Not until
a grammage of 10g/m2 does the fibre mat reasonably cover the knuckles.

502 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Figure 49 Fibre mat development for groundwood on a triple-layer fabric; 0.2, 4, 7

and 10 g/m2 respectively [202].

They calculated the mat pressure drop using the equation

ΔPmat = ΔPtotal − ΔPfabric (6)

where the fabric pressure drop is determined separately. According to the

fibre distributions described above, there is however a high correlation
between fabric openings and fibre locations, during the early stages of
dewatering,. This means that a low grammage web, separated from the fabric
would have a very low flow resistance, due to considerable openings in the
positions where the knuckles were detached. The combination of web and
fabric, therefore, give a much higher pressure drop than the sum of those of
the two individual components. This was earlier discussed by Meyer [203].
Dewatering resistance as a function of velocity v is generally modelled
using both the linear, viscous coefficient a and the second-order turbulent
induced dynamic coefficient b, see Equation 7.

12th Fundamental Research Symposium, Oxford, September 2001 503

B. Norman and D. Söderberg

ΔP
= av + bv2 (7)
L

Danby et al. made several studies of the influence of the forming fabric on
print quality. Initially, the different geometries of single, double and triple
layer fabrics on print evenness were demonstrated [204]. Periodic marking in
split sheet surfaces was evaluated. It correlated highly with respective fabric
surfaces and print marking [205,206]. With consideration of the fibre length
distribution, a correct choice of fabric mesh pattern (opening size and orien-
tation) had a large effect on retention level. A computer model was also
developed to predict print quality from the image of the surface layer of
paper samples after sheet splitting [207].
A similar flow loop as that of Herzig and Johnson was used by Jong et al.
[208]. The specific flow resistance of sheets of different grammage, formed by
LWC pulp, was measured. It could be noted among their results that in the
case “Closed loop Screen B”, a very high flow resistance was recorded at
20-30 g/m2, which dropped drastically at further fibre addition; see Figure 50.
This might indicate “sheet sealing”.
If no flow resistance were imposed by further fibre addition, the resistance
curve would follow the added broken curve. In reality, the extra fibres do
contribute to sheet resistance, but only to a degree corresponding to the
distance from the broken curve.

Figure 50 Effect of basis weight BW on mat flow resistance SFR for a LWC pulp
(Figure 7 in [208]). Added broken curve follows the equation SFR = k/BW.

504 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Fabric design may also be very important regarding rewetting. When the
actual web build-up is finalised, web dryness is increased over vacuum boxes
and couch roll. If water after this still remains in the fabric structure, a
substantial part of this water is likely to be sucked back into the paper web,
before and/or after its transfer to the pickup felt. Web rewetting by a single
layer forming fabric was studied by McDonald [209] on a pilot machine.
Samples for dryness evaluation were collected after the couch, in a grammage
range of 25–88 g/m2 using newsprint furnish. Rewetting was evaluated using
the Sweet method and gave a rewet value of 55 g/m2. This is a significant
amount of rewetting, especially for a low grammage product.
Simulation of initial fibre retention by the forming fabric has been made
using software developed to model the forming fabric in 3D, given the yarn
and weave parameters [210,211]. The geometric probability of initial reten-
tion for a fibre is calculated based on its given length and orientation.

6.5 Wet web resistance

Sayegh and Gonzalez [212] studied the compressibility of fibre mats during
drainage. A constant-rate drainage apparatus was used, and it was found that
the drainage resistance increased considerably with time at a constant water
flow rate. A Maxwell model was used to simulate the results. The time scale
used was from a few seconds and upwards, a too long time to be of value in
applications at papermaking conditions.
Vomhoff and Schmidt studied the effect of static compression on the dry-
ness of wet webs [213]. After ordinary filtration dewatering with low filtration
pressure, the wet web concentration will be in the order of 3–4%. As seen
from Figure 51, web concentration will increase to above 10% with the appli-
cation of a pressure of 10 kPa, a typical dewatering pressure in e.g., twin-wire
roll forming. Due to the dynamic effects, it is however still unknown how
much of this compression will take place in reality during a roll dewatering
phase ( ≈ 25 ms) or a blade pressure pulse ( ≈ 1ms) respectively.
Jong [214] studied the compression of thick wet mats with pressures up
to ca 10 kPa.
Sampson and Bridle [215] studied the flow pattern in a web using CFD. By
assuming a porosity gradient in the thickness direction, he demonstrated a
direction dependent flow resistance.
Sampson and Kropholler [216] modified a batch-drainage tester by intro-
ducing an electronic balance and a data acquisition system. The apparatus
yielded a full batch-drainage curve, representing dewatering on the forming
table to the dry line. Wet-pad concentration was evaluated to be in the range
2.5–3% for hardwood and 3–3.5% for softwood. These values were extremely

12th Fundamental Research Symposium, Oxford, September 2001 505

B. Norman and D. Söderberg

Figure 51 Web concentration as a function of static compression pressure [213].

insensitive to degree of beating and beater type. A model was presented for
the build-up of a fibre mat from a fibre suspension based on the response
determined in the above drainage device [217].
Mantar et al. [218] used a constant pressure drainage tester to study
dewatering of fibre suspensions of chemical and mechanical pulps using 17
kPa dewatering pressure. Dewatering rate was evaluated by measurement of
the drained volume. Influence of grammage (up to ca 400 g/m2), suspension
concentration (0.1–0.9%) and fines was included.
Ramarao and Kumar [219] modelled gravity drainage of pulp suspensions.
During gravity drainage, pressure, concentration and filtration resistance will
pass through a maximum with respect to time.
The model can be used to study drainage in conventional laboratory
devices such as hand sheet formers and freeness testers.
Wildfong et al. [220] studied filtration mechanisms of sheet forming. Con-
stant pressure filtration was generated by opening a valve to a vacuum cham-
ber with a pressure of −6 kPa. Dewatering rate was evaluated using optical
triangulation of the slurry surface. Realistic suspensions were taken as head-
box samples from newsprint, LWC and fine paper operations, but since the
retention values obtained were as high as 80–90%, fines content in the webs
formed were over-represented. Industrial like conditions (slice opening) were
used for initial slurry height. Although considerably more realistic than other
dewatering experiments reported, a slow initial dewatering (acceleration from
standstill) may be the most important deviation from industrial conditions
with a sudden jet to fabric impact. As related above, the initial dewatering
velocity may have an important influence on the build-up of the first fibre

506 Session 3: Forming

 Overview of Forming Literature, 1990–2000

layers on the fabric surface openings. The viscous resistance coefficient a

(Equation 7) was evaluated from the experiments. Before this, fabric pressure
drop was evaluated separately and subtracted from the total pressure drop.
An example of the evaluated resistance coefficient is shown in Figure 52.

Figure 52 Viscous resistance coefficient values a for different fibre suspensions [220].

In a separate investigation they studied the effect on dewatering resistance

of fine material and compression, which were found to be the two most
important parameters influencing specific filtration resistance at increasing
grammage [221].
Green et al. [222] suggested a design for a permeability cell, in which
dewatering resistance can be measured with controlled shear imposed upon a
dewatering event, to simulate forming conditions. A prototype not including
the shear device has been tested with three different pulps and different form-
ing fabrics [223]. The main principle is to use a fibre suspension of head-
box consistency and an initial channel height similar to the headbox slice
opening. A moving cylinder generates dewatering, and is accelerated and
decelerated during a test sequence.
The main difference from the real dewatering situation is the gradual
increase in dewatering velocity in the initial phase, compared to the high
velocity jet impact on a forming fabric.

12th Fundamental Research Symposium, Oxford, September 2001 507

B. Norman and D. Söderberg

7 DEWATERING PROCESSES
7.1 Jet impingement
The jet impingement on a forming fabric is a critical part of the paper forming
process, and has been studied experimentally and by modelling.
Shands [122] studied mix jump at jet impingement using linear wave theory,
and experimentally using laboratory flow models and pilot paper machine
trials. The experiments confirmed the theoretically suggested mechanism that
mix jump can result from wave amplification caused by the acceleration
changes that are produced as the mix jet lands on the wire. The effects of
headbox generated turbulence, free jet length, jet speed and curvature were
studied experimentally (see Figure 25).
A forming board arrangement with a radiused lead blade that can be used
to control mix jump at high machine speeds was also described.
Audenis [224] modelled the impingement of a plane inviscid liquid
jet on a forming fabric at different jet incidence angles. The deformability
and the porosity were considered separately. The present method extends
previous works in that part of the fluid is allowed to flow out of the
computational domain along the fabric and a more realistic boundary
condition has been applied at the porous fabric in terms of a permeability
law.
Jong [214] made a numerical and an experimental analysis of jet impinge-
ment on one wire and in a twin-wire nip. Measurements were made on the
Paprican pilot machine. There was a free wedge zone against one wire, also in
the twin wire case before contact with the forming roll took place. Reasonable
results were obtained, with roughly twice the amount of dewatering evaluated
numerically in comparison with experiments.
Dalpke, Green and Kerekes [225] modelled the impact of a jet onto a flat
wire using CFD software. Their main aim was to study the entrance region of
a twin-wire roll former, but the initial investigation treated the landing of a jet
on a flat wire. One viscous and one inertial term were used to characterise
wire flow resistance. In the main part of the calculations, fibre mat resistance
was neglected. Jet angle and velocity were important parameters. It was
found that the main part of dewatering took place within a distance equal to
the jet thickness, and that the pressure event further downstream was little
influenced by the different variables settings.

7.2 Fourdrinier dewatering

Loewen and Butler [226] applied infrared laser light reflection and correlation
techniques to evaluate mix surface flow velocity in a Fourdrinier section. At

508 Session 3: Forming

 Overview of Forming Literature, 1990–2000

450 m/min a jet excess speed of 7% was reduced to zero within 7 m (ca 1 s). It
was also possible to evaluate MD and CD velocity components at jet exit.
CD velocities agreed with fibre orientation misalignment angles of the
paper produced. Loewen [227] also discussed fibre orientation anisotropy and
misalignment in relation to jet-to-wire speed difference and headbox edge
flow adjustments.
Farnood et al. [228] used a laser velocity sensor to evaluate MD mix speed
and table activity and a gamma radiation gauge to evaluate drainage profile.
They studied effects of slip velocity, mix concentration and table activity on
paper formation. They suggested that such equipment could be applied for
on-line control of formation.
Kiviranta [229] made several studies of table activity in the Fourdrinier
section, and its effect on paper formation. Initially a photo clinometric
method to quantify mix surface irregularity was developed [230]. It was based
on a low angle sideways incident stroboscope flash, and a recording CCD
camera mounted over the wire section. Three main parameters were found to
be important in characterising activity: surface roughness, correlation length
and scale.
The factors affecting stock surface structure and paper formation for the
middle ply of folding boxboard was studied on a pilot machine [231]. The
same pilot machine was used to study the effects of mix concentration and
gravity foil parameters on fine paper formation [232] and on linerboard for-
mation [233]. These results were later also summarized [234].
The new method was tested on an industrial fine paper machine [235]. It was
found that table activity should be kept low in the initial dewatering stages. A
too severe activity decreased retention, which in turn automatically increased
retention aid addition, generating a higher degree of fibre flocculation.
Kiviranta and Dodson [53] used the pilot machine data by Kiviranta
described above, and included the Crowding factor N to replace the fibre
concentration. This improved the possibilities to predict formation levels. An
increase of N by 10 units corresponded to roughly 10% increase of local
grammage variations, which also corresponded to the range of influence on
formation by change of foils angles. The regression equations developed
could describe the effect of the parameters considered within a rather wide
range of operating conditions.
Foulger [236] investigated the submerged drainage concept of dewatering.
The new submerged drainage box replaces the conventional low vacuum box.
The main difference is that air is avoided by the use of a water filled box under
gentle vacuum. An increased dewatering capacity in the low mix concentra-
tion range, 1–2%, was reported, combined with a significant increase in the
retention level.

12th Fundamental Research Symposium, Oxford, September 2001 509

B. Norman and D. Söderberg

Zu and Chen [237] made a theoretical model of the vacuum generation by

gravity foils and step foils. Equations were developed based on lubrication
theory. For gravity foils, peak vacuum occurs at the tail of the foil, while step
foils show the opposite behaviour, with peak vacuum occurring near the front
tip of the foil, just after the step change. Comparisons of simulations and
experiments demonstrated the validity of the proposed equations.
A new type of foil element with an undulating upper surface has been
introduced [238,239]. Water is successively removed and returned, which
reduces the risk for “sheet sealing”, and with positive effects on retention
level. The design also makes it possible to generate constant mix activity,
independently of the actual dewatering rate. Sodergren and Neun [240]
applied classical fluid dynamic techniques to model the interaction between
the drainage surface and the slurry. Pilot trials were used to validate the fluid
model. While the forming fabric does contribute to the foil surface pressure
profile, the most significant input is from the shape of the foil surface.
An often-used method of modernisation of Fourdrinier machines was a
rebuild into a hybrid machine, by the installation of a top-wire unit at the end
of a wire section. This could be the Voith D-unit [1] or the Valmet MB-unit
[241,178], both including loadable blades below the wires. Also Beloit intro-
duced loadable blades beneath a top wire [170]. See further under twin-wire
dewatering.

7.3 Blade dewatering

Bando et al. [242] measured blade pressure pulses using the capillary probe
technique developed by Beloit [1]. They mainly looked at the maximum pres-
sure peaks using 5, 9 and 18 blades respectively in a typical “Bel Baie shoe”.
Parameters varied were wire bending stiffness (slight decrease of peak pres-
sure at increasing stiffness), pulp freeness (peak decrease at higher freeness)
and machine speed (slight peak increase at increased speed). It should be
remembered that since the pressure peaks are generated by local outer wire
deflections over the blades, changes in the dewatering rate at each blade will
influence local wire separation and thus pressure pulse events.
Bando et al. [244,243] also modelled the blade pressure pulses and studied
the effects on dewatering. They designed a filtration apparatus, in which
dewatering could be pulsated in a time of ca 15 s. In comparison with constant
dewatering rate, the pulses gave higher web concentration and higher filtra-
tion resistance. The modelled pressure pulses agreed satisfactorily with those
measured with a capillary probe; see Figure 53.
They also measured the outward and inward dewatering respectively and
compared these with the flows based on simulated pressure profile; see Figure 54.

510 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Figure 53 Comparison of pressure profile obtained by calculation and actual

measurement. 9 blades, radius of curvature 5 m, speed 700 m/min and basis weight 64
g/m2 [243].

Figure 54 Comparison of drainage rate between calculation and experiment; data as

in Figure 53 [243] outer wire deflections over the blades, changes in the dewatering rate
at each blade would influence wire separation and thus pressure pulse events.

12th Fundamental Research Symposium, Oxford, September 2001 511

B. Norman and D. Söderberg

Since the actual inward flow was much higher than simulated, they concluded
that the inner web resistance was low. This could be a result of washing out
effects by the blades, similar to that on a fourdrinier machine.
Bando et al. [39] also studied the effect of pressure pulses on orientation
anisotropy.
At increased jet to wire speed difference, they found that changes in
anisotropy only took place in the middle half of the sheet. They calculated
local flow velocity between the wires (see Figure 55), and evaluated cumulative
shear factor (m2/s) by integrating the absolute difference between wire and mix
speeds along the forming section. Fibre orientation anisotropy in the middle
layer increased with cumulative shear factor, but the effect of high amplitude
pulses was relatively much higher than that of lower pulses.

Figure 55 Velocity profile of pulp slurry between top and bottom wires as a function
of distance from jet impingement position for 5, 9 and 18 blades respectively [39].

The shape of blade pressure pulses was also studied by Zhao and Kerekes
[245]. A mathematical model was developed, assuming inviscid flow, a thin
blade, zero wire bending stiffness and constant permeability; see Figure 56
(left). They also made experiments on the Paprican pilot machine, and meas-
ured the pressure events using tappings along a flat blade, with wire deflection
over the back end; see Figure 56 (right).
Zhao and Kerekes [246] also tried to model the formation improvement
generated by blade pulses. They compared formation values (NUI) in roll
and roll-blade forming using a range of blade pulse amplitudes. They finally
correlated the formation improvement NUI with the calculated total mix
displacement due to the blade pulses and claimed a positive correlation.

512 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Figure 56 Left: Calculated pressure pulses at different wire speeds along a flat blade
with wire deflection over downstream end. Right: Comparison between calculated
and measured pressure pulses [245].

However, the spread in the data was much too high to motivate such conclu-
sions. It would be useful to characterise also formation scale, if conclusions
about the sheet forming mechanisms are to be drawn.
Nigam and Bark [247] considered the two-dimensional flow field between
two wires deflected over a curved blade. The case with impermeable wires and
irrotational flow was analytically solved; see an example in Figure 57.
Zahrai et al. [248] studied different aspects of twin-wire forming. Initially
a two-dimensional model was applied using a thin blade, and effects of
permeability and wire bending stiffness were considered [249,250]. Local suc-
tion pulses downstream of a blade were predicted. Analytical calculations
were compared with experimental results from the EuroFEX machine using
different blade shapes [251]; see Figure 58.
Green [252] reinvented the fact that the blade force (integrated blade
pressure) is proportional to the outer wire deflection angle over a blade.
Green et al. modelled different aspects of blade related pressure events and
studied:
• two-dimensional simulation of pressure pulses in blade formers [253], (like
Zahrai, see above) they found a downstream vacuum pulse;
• influence of wire stiffness, inner and outer wire tensions and blade width
[254];
• numeric analysis of blade pulses, in comparison with earlier analytical

12th Fundamental Research Symposium, Oxford, September 2001 513

B. Norman and D. Söderberg

Figure 57 Streamlines for the velocity field between two wires deflecting over a
curved blade (tip length 50 mm). No deflection of inner wire on ingoing and outgoing
side. Scales in m [247].

Figure 58 Fluid pressure over a 50 mm-wide triangular blade. Wire deflections at

front edge: 0 deg, top: 2,3 deg and back edge: 0.85 deg. Modeling (curve) and
measurements (crosses). Values on x-axis in mm and on y-axis in kPa [251].

514 Session 3: Forming

 Overview of Forming Literature, 1990–2000

solutions [255], the possibility to introduce decreasing permeability along

the dewatering zone was beneficial;
• the extension of the hydrodynamic pressure pulse generated at doctoring
of water layer [256];
• the effect of non-Darcy’s law web permeability on the hydrodynamics of
blade dewatering [257].
Shands and Wildfong [258] modelled the blade pressure pulses in a blade
plus counter blade arrangement. An example is given in Figure 59.

Figure 59 Modelled blade pressure peaks (left) and local mix velocities (right) in a
blade arrangement with 5 initial blades followed by 4 loadable blades and 5 stationary
blades. Wire tensions 6 kN/m and blade loads 0.5 kN/m [258].

Suction shoes
It was suggested by Odell, Pakarinen and Luontama [259] that when a twin
wire sandwich passes over a suction shoe, the wires will be bent into the shoe
openings by the suction forces; see Figure 60. The hypothetical result would
be an increased outer wire deflection around blade back and front edges,
which in turn would create increased pressure pulse amplitudes between the
wires.
Green [260,261] made an effort to model the mechanisms of suction shoe
dewatering, starting from a wire geometry as shown in Figure 61.
Roshanzamir, Green and Kerekes [262] attempted to model the suction
shoe event with a two-dimensional simulation, based on the 1-D method
earlier tried by Green. The argument that good agreement with the 1-D

12th Fundamental Research Symposium, Oxford, September 2001 515

B. Norman and D. Söderberg

Figure 60 Hypothetical wire deformation due to suction forces [259].

Figure 61 Hypothetical paths for two wires travelling along a suction shoe opening
[261].

method lends credence also to the 2-D method does not seem convincing.
Also in this case a considerable suction is predicted inside the outer web. The
modelled geometry of the wire runs in the different cases treated is, however,
not reported.
However, the physical model behind Figures 60 and 61 is incorrect since the
force bringing down the outer wire into the opening is negligible. The wire

516 Session 3: Forming

 Overview of Forming Literature, 1990–2000

will, therefore, travel along a mainly straight path between two adjacent
suction shoe blades.
When there is still free mix between the inner and outer webs, the web on
the inner wire will absorb the main part of the suction pressure, and the inner
wire would bend. The free mix would then support a vacuum high enough
to loosen the web from the outer wire. At a later stage, there will be one solid
web between the wires, which would loosen from the outer wire due to an
inside vacuum.
Suction application in the blade section of a twin-wire former will have a
major effect, previously not reported, on the generation of blade pressure
pulses.
In a blade arrangement with fixed blades on one side and loadable blades
on the opposite side, a pressure pulses is created in front of a loaded blade.
(Also about half the effect will occur at the two surrounding blades on the
opposite side, but this will not be further discussed now.) The basic reasons
for this mechanism is that when a loadable blade moves the contacting wire
closer to the opposing wire, the space for mix between the wires is reduced. To
still maintain the same rate of mix flow, a high-pressure zone will develop
upstream of the blade, and that pressure zone will locally move away the
opposite wire, so that the cross sectional area is mainly restored.
If now suction is applied beyond the opposite wire, the low pressure will
help to locally displace the opposite wire, and thus increase the cross sectional
area for mix flow. At some suction pressure level, the displacement of the
opposite wire will be sufficient to let the mix pass without the help of the
blade generated pressure pulse between the wires mentioned above. The blade
pressure pulse will thus disappear. The applied blade load will instead be
completely balanced by the local deflection of the contacting wire, which will
then increase in comparison with the deflection without suction (and then
increase the pressure pulses at the two surrounding blades on the opposite
side).

7.4 Roll dewatering

Martinez [263] studied the roll-dewatering phase in roll-blade forming on the
EuroFEX pilot machine. An example of dewatering pressure event is shown
in Figure 62.
A pressure sensor at the end of a thin flexible cord is initially released into
the jet (A), then into the twin-wire nip (B), and finally into the zone where the
wires leave the roll, generating a suction pulse (C). It is evident that the
pressure level does not reach the nominal level T/R until the final stages of
roll dewatering.

12th Fundamental Research Symposium, Oxford, September 2001 517

B. Norman and D. Söderberg

Figure 62 Pressure event in roll dewatering during roll-blade forming (trial N in

Figure 63). Roll diameter 1635 mm, wire wrapping 30 deg, wire tension 8 kN/m,
machine speed 800 m/min. Recorded on the EuroFEX machine in co-operation with
Valmet [263].

The pressure event during the initial part of roll forming has not yet been
modelled theoretically, and the local pressure minimum just before the final
maximum is also so far unexplained (see further below).
Martinez developed a physical model to predict the dewatering through the
two wires in twin-wire roll forming. A series of force and mass balances were
applied to the web, and expressions were derived for web thickness and solid-
ity based on fibre and process conditions. Using the measured pressure event,
dewatering rates were estimated with Darcy’s law. A series of experiments on
the EuroFEX machines were performed, and one test was used to calibrate
the parameters. The numerical solution predicted flow rates within 10%; see
Figure 63.
Boxer applied modelling techniques to a roll-blade former. He modified the
method by Martinez for roll dewatering, by replacing the numerical solution
method by an analytical one [264]. It is claimed that with the new method,
knowledge of the actual pressure event during dewatering is not required for
the solving of the dewatering flow rates [265].
Zahrai, Martinez and Dahlkild [266] presented a physical model to esti-
mate the thickness of the web in twin-wire formers with nearly constant
drainage pressure. Web thickness was found to be proportional to the square
root of forming distance. The constant of proportionality was related to
parameters such as compressibility and permeability of the web, headbox
consistency and machine speed.

518 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Figure 63 Comparison of the estimated to the predicted dewatering rates from both
the inner (A) and outer (B) fabric [263].

Zahrai, Bark and Martinez [267] modelled the dewatering from a curved
converging channel with moving walls, of which the outer one was permeable.
A numerical solution was obtained with increasing wall excess speeds relative
to the suspension for two fibre suspension models: one power law version and
one more “fluid-like”. It was shown that with increasing speed difference, the
fluid-like model resulted in an increased fibre concentration of the web at the
interface to the suspension, which was not the case using the power law
model.
Turnbull et al. [268] developed a 1-D model for a Crescent former (roll
former with one-sided dewatering) to study dewatering events, and specific-
ally the influence of longitudinal disturbances. Despite heavy damping,
the model showed significant amplification of disturbances near resonant
frequency. The model predicts that 2% disturbance in jet velocity then
may result in nearly 7% variation in basis weight. The wavelength for such
variations approximately equals the forming roll radius.
Chen et al. [269] modelled the CD mat thickness disturbances on a Cres-
cent former using a two-dimensional dynamic model with viscosity. The
study examined the CD variations due to the influence of major classes of
steady disturbances, and the coupled MD and CD variations due to dynamic
process disturbances in the fluid jet. Among many results, those from
unsteady disturbances show that the coupled MD and CD variations are
magnified near the fundamental natural frequency of a translating wire for
long wavelength disturbances.
Jong [214] made some roll forming modelling attempts and compared
the results with measurements on the Paprican pilot machine. One main

12th Fundamental Research Symposium, Oxford, September 2001 519

B. Norman and D. Söderberg

conclusion was that the model should be further developed before good pre-
dictions of dewatering could be expected.
In a recent overview of twin-wire forming, Malashenko [270] discussed a
typical roll pressure event in a roll-blade former and commented “Prior to
exit, there is an ever present and unexplained reversal in pressure . . .”. It
could be pointed out that such a local minimum is also present at the end of
the B-section in Figure 62.
Malashenko also stressed the importance of relative motion between wires
and connected web shear to different wire tensions and running radius over
rolls; see Figure 64.

Figure 64 Differential fabric displacement and web shear in twin-wire “S” path, [270].

Wildfong et al. [271] modelled drainage during roll forming, and validated
results using pilot paper machine data. Stock drainage data were evaluated
according to methods described above [220,221]. Roll dewatering pres-
sure was assumed constant, and the vacuum in the forming roll was neg-
lected. Prediction accuracy of drainage flow was within 15% of experimental
values.
A pilot plant study of roll dewatering of reslushed newsprint was made by
Gooding et al. [272]. They found that dewatering comprised two main parts,
one momentum driven and one tension-driven. Dewatering rates were
obtained by scooping with a special collection device, and further by flow
balances based on wire position estimations based on a probe recording
mechanical bending of the probe assembly. No separate measurement was
made of dewatering to the roll side, and the concluded values resulted in
some peculiarities. Normal fabric tensions were applied, but since the roll
diameter was only about half that of a modern industrial installation, the
dewatering pressures were twice the normal levels. Recorded retention levels
were high; for a fine fabric over 85% was obtained. Pressure events were

520 Session 3: Forming

 Overview of Forming Literature, 1990–2000

recorded using an air purged capillary tube introduced through the headbox,
and stopped at different positions along the forming zone, the same tech-
nique as that used by Beloit and Mitsubishi [242]. Pressure recordings reveal
some uncertainty about the probe tip location; sometimes it seems to have
been located inside the forming web.

7.5 Vacuum dewatering

Neun [273] studied vacuum box dewatering on an experimental paper
machine. Different vacuum levels and vacuum slot dimensions were used, and
web dryness was tested on blow samples. For a given vacuum level he found
that solids content C of the web as a function of time t could be described by
the equation:

C = b + m tanh(ct) (8)

where b, m and c are constants.

Figure 65 shows an example of experimental data and application of
Equation 8.
He also demonstrated how a segmented dewatering curve for a set of con-
secutive vacuum boxes could be designed:
Follow the bottom curve in Figure 65 for the desired time. Then move left
at constant solids content until the middle curve is reached. Follow the

Figure 65 Web solids as a function of dwell time for flat box at three vacuum levels.
The lines are drawn using Equation 8 [273].

12th Fundamental Research Symposium, Oxford, September 2001 521

B. Norman and D. Söderberg

middle curve for the relevant time period, and then move left to the third
curve, etc.
By application of this principle, he demonstrated considerable potential in
reductions of drag forces (fabric wear) and vacuum energy consumption
through improved strategies of vacuum application, replacing traditional
settings.
Neun and Fielding [274] studied vacuum box dewatering by field meas-
urements. They concluded that corresponding laboratory results were rea-
sonable absolute predictors. Increased grammage had an effect on slowing
down dewatering rather than on achievable dryness. Furnish had a large
effect, and higher temperature improved dryness.
Neun also evaluated the parameters m and c (Equation 8) for newsprint,
OCC and Kraft furnish at different vacuum levels [275,276,277].
Räisänen [278,279] studied flat box vacuum dewatering using the “Moving
Belt Drainage Tester”, MBDT (see description below). Using this equipment,
the same type of vacuum pulses are applied both during forming and vacuum
dewatering.
The process was considered to be basically a wet pressing event, and the
solid content C as a function of time t was described based on a first order
viscoelastic model

冢 冣
t
C = C0 + b 1 − e− r + dt (9)

The linear term was added to account for the effect on water removal of the
through-flow of air. The effect of different furnish parameters were studied
[280] and the laboratory evaluations were successfully compared with the
results for a pilot [281] and a commercial [282] paper machine respectively.
If, in fact, wet pressing is the dominating effect for water removal on
vacuum boxes, this part of the process should preferably be treated together
with the press section. Traditionally this has however not been the case, and
therefore this part is here still included in the forming section.
The MBDT was later also used by Mitchell and Johnston [283] to study
vacuum dewatering.
Shands and Hardwick [284] studied flat box vacuum dewatering on a pilot
machine for newsprint at commercial speeds. Dryness increase was found to
be mainly proportional to “vacuum impulse”, i.e., vacuum level multiplied
by dwell time. Also airflow requirements and drive loads were found to be
proportional to vacuum impulse.
Jones [285] developed a dynamic simulation model for flat box vacuum
dewatering using the experimental results by Neun (mentioned above). The
model is based on flow through compressible porous media. What appear to

522 Session 3: Forming

 Overview of Forming Literature, 1990–2000

explain the limitation of the flat box are the forces required to remove water
from the smaller pores of the sheet. Dewatering stops at the point at which
the surface tension forces in the pores exceed the applied vacuum. This can
explain why solids content levels off regardless of dwell time but increases
with increasing vacuum and temperature. Increased grammage slows drain-
age but does not reduce maximum solids content.

8 LABORATORY FORMING

Since fundamental research within the forming process performed on indus-

trial paper machines is generally too expensive and too restricted regarding
process design, means have been developed to perform such research in the
laboratory. Different kinds of laboratory sheet formers have therefore been
developed as well as new pilot machines for web forming.

8.1 Sheet forming

The main drawbacks with a standard hand sheet former are low forming
concentration, slow dynamics, no white water recirculation, isotropic fibre
orientation in the paper sheet and unrealistic fines distribution in the
Z-direction. The aim has therefore often been to increase forming concen-
tration, to make it possible to control fibre orientation anisotropy and
to rearrange dewatering to be closer to a real situation. White water recircula-
tion is however usually still missing.
Nazhad and Kerekes [286] modified a standard hand sheet former by
applying agitation before drainage of a mix with headbox concentration. This
means that mix height before drainage is ca 10 mm. Mixing was performed
using a plastic plate with rather closely spaced 10 mm diameter holes. The
plate was initially placed in contact with the upper surface of the mix and
then withdrawn vertically. Depending on the desired degree of uniformity of
the final sheet, 1–15 mixing actions could be performed. It is pointed out that
the retention of fine material is not simulated in this procedure, and that the
method therefore is only useful for determining the formation potential of
the long fibre component of the mix.
Using the equipment above, Nazhad et al. [287] demonstrated that both
paper formation and strength deteriorated when forming concentration
increased from 0.017% to 0.5% using TMP fibres of different size fractions.
Northern softwood was confirmed to give higher strength values than south-
ern pine. The strength for all samples decreased when grammage was reduced
to ca 30 g/m2.

12th Fundamental Research Symposium, Oxford, September 2001 523

B. Norman and D. Söderberg

The BetzDearborn Pulsed Drainage Device (BPDD) was developed in the

late 1980s [288]. The mix is deflocculated in a separate mixing chamber, and
then released onto the forming wire. A rotating foil below the wire will agitate
the mix during the drainage. Vacuum is applied below the wire, and the
dewatering rate is recorded (with a short delay) by measurement of the water
level at the bottom of the white water collection chamber; see Figure 66.
The BPDD equipment was later modified by Sutman [289] especially to
improve its predicting capabilities regarding dewatering times. By decreasing
the mix volume by a factor of four and aiming at a relevant grammage, a mix

Figure 66 The BetzDearborn Pulsed Drainage Device [289].

524 Session 3: Forming

 Overview of Forming Literature, 1990–2000

concentration close to that in an industrial application was utilised. A coarse

stainless steel screen replaced the wire, the foil speed below the screen was
increased and a moderate vacuum was applied at the beginning of a test. The
basic idea was to introduce more thickening dewatering, replacing the earlier
filtration dewatering.
The modified device gives dewatering times close to those in reality, while
the earlier design had given one order of magnitude too long times. The
vacuum during dewatering is recorded, and it is suggested that the peak to
equilibrium vacuum ratio would be an indicator of sheet formation, which
seems to be an oversimplification.
The Moving Belt Drainage Tester (MBDT) [290] is another device
developed to study dewatering under process-like conditions, especially the
final vacuum dewatering stage (see above); see Figure 67. The basic principle
of the MBDT is in principle similar to that of the BPDD. However, the

Figure 67 The Moving Belt Drainage Tester, MBDT [290].

12th Fundamental Research Symposium, Oxford, September 2001 525

B. Norman and D. Söderberg

agitation during dewatering is performed with a cogged, endless belt, which is

run below the wire. The geometry of the cogged belt and its running speed
will control the mode of agitation. It has been applied for the study of wire
section dewatering, including the use of retention aids [291,280].
Xu and Parker [292] developed the MBDF former, a variety of the above
former. Instead of dewatering studies, the main aim was to study forming
effects on sheet properties. At increased forming concentration formation
deteriorated, and the accompanying effect on paper mechanical properties
was low for short fibre pulps and significant for long fibre pulp.
Helmer et al. [293,294] have developed a system for the forming of aniso-
tropic hand sheets. The mix is delivered from a headbox, and flows across a
wire-covered vacuum box; see Figure 68. The sheets produced are of con-
siderable size, 260 × 320 mm2. Typical running data would be slice opening 10
mm, mix concentration 0.5%, and jet velocity 30 m/min. A typical tensile
stiffness index anisotropy obtained would be ca 1.8.

Figure 68 The University of Melbourne stationary wire laboratory former [293]

(Left) and operation limits at constant depth and vacuum [294] (Right).

For the channel flow on top of the wire, the Froude number Fr has signifi-
cance, with

v2
Fr = (10)
gd

with v flow velocity, g gravity factor and d mix height.

The application of the new former has so far concentrated on paper forma-
tion and anisotropy – other paper properties have not been reported.

526 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Sivén and Manner [295] designed a high-speed retention tester, the HSR-
Tester, based on a moving headbox and stationary twin-wire arrangement.
Inside the inner wire, drainage foils are mounted on a radius defining the
inner wire radius, and the set-up is rotated at ca 5 r/s. Outside of the inner
wire, the likewise stationary outer wire is tensioned. In between, the headbox
is placed, and when a sheet forming event is started it is retracted at ca 90 m/
min along a half-circle formed track, leaving the mix mainly stationary to be
drained through the two wires with the foil pulses agitation from inside. The
quality of the paper regarding formation and mechanical properties are not
reported so far. Its main aim is to study retention as a function of retention
aids addition [296].
Hammock and Garnier have developed a laboratory apparatus simulating
industrial twin-wire forming [297]; see Figure 69. The headbox width is 160
mm. The forming roll (diameter not specified) surface has 2 × 2 mm2 grooves
separated by 1 mm-wide ridges, to allow also inwards dewatering. After the
roll fixed blades on both sides and vacuum boxes bring final sheet dryness up
to 8–10%, which allows sheet sampling. The mix is pumped from a 400 L
reservoir, and bypassed back to the reservoir via a three-way valve before and
after the trial. Trial time is within the range 5–50 s. In a typical trial slice
opening was 4 mm, wire speed 320 m/min and mix consumption 140 L.

Figure 69 McGill roll-blade twin-wire laboratory sheet former [297].

12th Fundamental Research Symposium, Oxford, September 2001 527

B. Norman and D. Söderberg

It should be pointed out that in most laboratory devices described and

forming experiments reported, except for those in industrial systems and in
the EuroFEX system, a fibre suspension with the same composition as that of
the furnish is used. This means that it is not representative for a realistic
headbox mix, in which components with low retention are over-represented
due to the recirculation from the short and long circulations.

8.2 Pilot machines

An overview of paper pilot plants was presented in 1997 [298].
During the 1990s, the three dominating paper machine manufacturers all
started new pilot machines, mainly for printing paper development. A main
reason in all cases was to increase machine speed, since the speed potential of
the industrial machines started to surpass the pilot machines during the early
1990s. The machines were more complete than earlier versions by including
larger parts of the drying sections.
A new Voith machine was started in December 1990 in Heidenheim [154]
and after one year it was redesigned into a roll-adjustable blade twin-wire
version. Maximum speed was 2200 m/min and width 1 m.
Beloit started a new pilot machine Rockton in 1994 [299], with a max-
imum speed of 2400 m/min and 1 m width. During its last working years the
wire section was run in the roll-adjustable blade version.
Valmet started a pilot machine Jyväskylä in 1996 [300]. Design speed 2500
m/min and width 1 m. The wire section was of the roll-blade/adjustable blade
design.
Paprican built a new machine (with some reuse of existing components),
which was started in Pointe Claire in 1999 [301]. Headbox width 0.5 m, design
speed 1600 m/min and the machine is balanced for 2500 m/min. The wire
section is designed as a roll-adjustable blade unit. The white water system is
closed by the addition of a sand filter, in which material from the excess white
water is collected but not recycled. The stock system is designed for recircula-
tion of the furnish.
The EuroFEX machine at STFI, Stockholm, was started up already in
1982. It was balanced for 3000 m/min, design speed was 2400 m/min and
width 0.3 m. In 1991 the wire section was complemented to also include the
first ever roll-adjustable blade section [153]. The comparatively narrow
machine width was chosen to make a full day running without furnish
recirculation possible. This would not have been realistic with a wider
machine, considering the level of stock consumption. Continuous recovery of
fines and filler from the excess white water with a disc filter is enforced during
all runs.

528 Session 3: Forming

 Overview of Forming Literature, 1990–2000

The EuroFEX plant is today the only laboratory plant in which indus-
trial running conditions can be simulated if chemical additions and/or
multi-ply forming are involved, due to continuous fresh stock addition as
well as continuous recovery of material from excess white water. The dif-
ference from a pilot plant without these two design features was recently
demonstrated by Nalco Chemical Co [302]; see Figure 70. They made
similar test runs in the EuroFEX and the Beloit pilot plants. It is clear
from these tests that in the EuroFEX plant the white water chemistry
levels out between the different test points and that the levels recorded are
higher.

Figure 70 “Comparison of the SLM/FBRM mean chord length traces obtained for
XPM3 (Beloit) and EuroFEX pilot paper machines. The marks on the figure indicate
changes in retention chemistry or dose.” Figure 12 from [302].

9 FORMING, STRUCTURE AND PROPERTIES

Initially the generally spread misunderstanding that formation improvements

also mean paper strength improvements should be commented upon. A cor-
rect statement is instead that under some circumstances, both strength and
formation will be changed in the same direction.

12th Fundamental Research Symposium, Oxford, September 2001 529

B. Norman and D. Söderberg

• The most obvious way of improving both formation and paper strength
is to reduce forming concentration, assuming that current drainage
limitations are not superseded.
• An often-used method to worsen formation in hand sheet forming is to
impose a delay time before drainage. Formation as well as strength will
then deteriorate.
On the contrary, there are several occasions where the opposite situation
will be true, i.e., formation improvements will not be accompanied by
strength improvements:
• Excessive addition of retention aid can improve formation, but paper
strength will be unchanged.
• Manipulation of the activity on a fourdrinier wire may improve formation,
but this will normally be associated with a reduction in paper strength.
• An industrial method to improve formation in twin-wire forming is to
introduce blade pressure pulses. However, in this case paper strength will
deteriorate [117].
A “general rule” was formulated by Fredlund [303] as a summary to a
range of EuroFEX pilot plant experiments:
Strength potential in forming lies in the quality of the headbox jet. Manipula-
tions at later stages can improve formation but not paper strength.

9.1 Single-layer forming

In 1992, Holik made an overview of the state-of-the-art technology within
forming at Sulzer Escher Wyss [304,305].
Ohlsson [306] studied twin-wire forming of mechanical pulp containing
printing papers on the EuroFEX machine. It was found that formation
improved with decreasing slice opening, and not until around 5 mm, forma-
tion started to deteriorate due to the high forming concentration. It should be
remembered that the small slice opening resulted in a high nozzle contrac-
tion, with its positive effects on paper formation. This was however not
appreciated at that time.
Albinsson, Swerin and Öberg [307] studied the effect of blade dewatering
in the EuroFEX machine on formation and filler retention. It was found that
with a suitable retention aid strategy (addition components, dosage time
strategy), increased mechanical blade action due to increased outer wire ten-
sion could be used to improve formation without reducing the retention level.
Swerin and Mähler [308] studied the effect of fibre suspension properties
on final sheet formation in twin wire roll-blade forming on the EuroFEX

530 Session 3: Forming

 Overview of Forming Literature, 1990–2000

machine. When flocculants (cationic polyacrylamide and bentonite clay) were

added, higher shear (mix-to-wire speed difference) during dewatering was
needed to produce a less flocculated paper. At minimum speed difference,
formation number, filler retention and drainage (given as the amount drained
over the forming roll) all showed maximum values.
Nilsson [309] studied different aspects of twin-wire roll forming. He found
that a positive velocity difference between mix and wires would align fibres in
the negative MD (when viewed in a sample cross sectioned in MD), and vice
versa for a negative velocity difference. A tape splitting test on a paper sample
easily confirms this. A tape is attached in the MD on a paper surface. When it
is peeled off in the negative MD, few fibres will be pulled out, when the fibre
orientation is in the negative MD. With the opposite fibre orientation in the
sample, an increasing amount of fibres will be pulled out; see also [310].
Nilsson further studied the effect on paper formation of the mix to wire
velocity difference. He found that the traditional formation optimum at a
certain velocity difference between mix and wires moved towards increased
velocity differences for increased floc strength in the mix (Crowding factor,
chemical additives etc.).
Nordström and Norman [117] studied paper properties using a TMP fur-
nish in roll and roll-blade forming respectively, using different degrees of
headbox nozzle contraction. The effects on fibre orientation anisotropy were
discussed above. The effects on small-scale and large-scale formation are
shown in Figures 71 and 72.
At low and medium headbox nozzle contraction ratio (traditional for
hydraulic headboxes), small-scale as well as large-scale formation behaved in
the traditional way.
This means that optimum formation was obtained at a certain positive and
negative mix-to-wire speed difference respectively [1].
Large-scale formation optimum using a high nozzle contraction ratio was
obtained in a wide range around zero mix-to-wire speed difference (if the low
value at −60 m/min speed difference is disregarded), and the formation values
were the same as the optimum at low and medium contraction; see Figure 71.
The addition of blade pressure pulses (in combination with reduced roll
dewatering) improved large-scale formation in a wide speed difference range.
Small-scale formation optimum using a high nozzle contraction was
obtained at zero mix-to-wire speed difference, and the formation value was
significantly lower than that at low and medium contraction ratio; see Figure
72. This indicates that the fibre flocculation structure in the jet is improved at
high contraction, which is a result of floc stretching and breakage in the
elongational flow in the headbox nozzle (as also discussed in a previous sec-
tion). However, even a small increase of mix-to-wire speed difference during

12th Fundamental Research Symposium, Oxford, September 2001 531

B. Norman and D. Söderberg

Figure 71 Large-scale formation as a function of mix-to-wire speed difference in roll

and roll-blade twin-wire forming on the FEX machine. “Square” – high nozzle
contraction, “filled circle” – medium contraction, “unfilled circle” – high
contraction, “+ marks” – high contraction and blade pulses after the roll [117].

Figure 72 Small-scale formation as a function of mix-to-wire speed difference

and headbox nozzle contraction in roll and roll-blade twin-wire forming, data as in
Figure 71 [117].

532 Session 3: Forming

 Overview of Forming Literature, 1990–2000

dewatering generates a deteriorated small-scale formation. This is surprising,

since all previous investigations have shown a formation improvement with
some degree of shear during dewatering. The reason could be that the degree
of small-scale flocculation in the jet after strong elongation in the head-
box nozzle is so low that all shear during dewatering only can increase gram-
mage variations. The addition of blade pressure pulses after roll dewatering
slightly worsens small-scale formation in the whole speed difference range
investigated. Retention level and paper mechanical properties were either
maintained or impaired (especially Z-toughness) for roll-blade forming in
comparison with pure roll forming.
Nordström and Norman [311] further found that the separation distance
between the blades in roll-blade forming of TMP could be reduced from 116
mm to 46 mm without any negative effects on paper properties.
Nordström and Norman [312] studied the effect of jet quality on paper
properties in roll-blade forming of TMP. The high quality jet was produced
with high nozzle contraction, resulting in a paper tensile stiffness MD/CD
ratio of 4 at minimum mix-to-wire shear (thus generated by the orientation
anisotropy in the jet), and a more flocculated jet from a low contraction
nozzle, with a MD/CD ratio of ca 1.5. The detrimental effect on formation of
jet flocculation was reduced by introducing low level blade pulses after initial
roll forming, and also some increase in final paper tensile stiffness MD/CD
ratio was obtained. The lower Z-toughness for roll-blade dewatering com-
pared to pure roll dewatering was only to a minor degree influenced by jet
quality and blade force level.
Nordström and Norman [313] studied the influence of the proportion of
roll dewatering in roll-blade dewatering by changing the roll-covering angle
of the initial roll forming. It was found that the improvement in the large-
scale formation was best exploited with a sufficiently large proportion of roll
dewatering. The deterioration with lower degree of roll dewatering might be
caused by negative effects on the thin wet webs at the separation of the wires
from the roll surface as discussed above.
Odell [259,314] discussed Valmet roll-blade forming technology in relation
to different sheet structure properties, such as formation/retention, layered
orientation and material distribution in the Z-direction. In Figure 73 an
example is given of the influence of suction shoe vacuum in a roll-blade
former on filler distribution in the Z-direction.
Verkasalo, Odell and Korhonen [241] described the development of LWC
base paper from fourdrinier over hybrid formers to roll-blade twin-wire
formers.
Erkkilä, Pakarinen and Odell [43] quantified fibre orientation profiles
in the Z-direction using the tape-stripping method described above, with

12th Fundamental Research Symposium, Oxford, September 2001 533

B. Norman and D. Söderberg

Figure 73 Effect of suction shoe vacuum on controlling filler distribution [259].

different forming principles and jet-to-wire speed ratios; see Figure 74. They
motivated the different profile shapes by different dewatering rates and differ-
ent speed differences between web and mix:
• the low degree of anisotropy towards the wire surfaces is due to the fast
drainage in those areas, which leaves little time for fibres to orient due to
shear;
• the lower degree of anisotropy at the centre with two-sided dewatering is
due to reduced shear velocities due to viscous damping;

Figure 74 Layered orientation anisotropy during “rush” (left) and drag (right)
forming conditions for three forming principles [43].

534 Session 3: Forming

 Overview of Forming Literature, 1990–2000

• the asymmetry of fourdrinier sheets is due to the unsymmetrical shear

effects along the dewatering section.
The local anisotropy minimum in the centre of a gap dewatered sheet could
alternatively be an effect of space limitations for fibre rotation at this stage, as
well as a natural, geometric accumulation in the centre of fibre flocs.
Jansson [44] studied the effect of headbox nozzle contraction ratio on fibre
orientation anisotropy for roll forming at minimum shear conditions during
dewatering; see Figure 75. According to this figure, anisotropy is low at sheet
edges even when orientation around sheet centre is high. One reason for this
has later been shown by Asplund [120] to be the lower degree of fibre orienta-
tion anisotropy in the headbox nozzle boundary layers (see Figure 22).
Another reason could be the effects of local micro flow disturbances at
jet-to-wire contact.
Erkkilä, Pakarinen and Odell [315] made further studies of the layered
structure of twin-wire formed paper. It is initially proposed that the role of
turbulence is extremely important when the suspension enters the forming
section. Elimination of turbulence might produce frozen flow with the
absence of relative motion between fibres, leading to flocculation and poor
formation. In contrast, small-scale turbulence in the flow is supposed to
create constant deflocculating action in the stock.
These views on fibre suspension turbulence are debatable (see also in
section Headboxes). It is true that the traditional model for turbulence decay
in a fluid will tell that the large eddies are broken down into smaller units, and
this process gradually continues until the final viscous dissipation into heat in

Figure 75 Fibre orientation anisotropy with different headbox nozzle contraction

ratios for roll formed paper at zero mix-to-wire speed differences; squares – low
contraction, triangles – high contraction [44].

12th Fundamental Research Symposium, Oxford, September 2001 535

B. Norman and D. Söderberg

the micro-scale. However, the situation in a fibre suspension is drastically

different. Bennington [60] claims that a large part of the dissipation takes
place directly between the fibres. Small-scale turbulence is also efficiently
damped by the presence of the fibres. A smaller part of the fluid energy will
then be transformed into the micro-scale range. Thus mainly more large-scale
turbulence will survive, and this large-scale turbulence, if present during the
dewatering phase, is highly detrimental to paper formation.
There is no known experimental evidence that turbulence present during
dewatering has a positive effect on paper formation. The basic idea should
instead be to affect the fibre suspension inside the head box, so that the fibre
flocs are as weak as possible and can be easily broken apart by the shear
allowable during dewatering.
Turbulence should be avoided at all stages of forming, since it is the main
source for reflocculation.
Using the tape stripping method, Erkkilä et al. analysed the different layers
for four different parameters; see Figures 76 and 77.

1. Anisotropy;
the definition of “anisotropy” in this work, Equation 11, differs from that
in Equation 2; see Figure 8.

“Anisotropy” = 1 − a/b ≈ 1 − CD/MD (11)

2. Standard deviation of anisotropy; calculated on 3.3 × 3.3 mm2

measurement areas.
3. Floc index.
4. Orientation angle.

It is clear from Figure 76 that “anisotropy” is considerably lower at

minimum mix-to-wire speed difference than with a negative speed difference.
The aligning effects from some shear considerably suppress local anisotropy
variations during dewatering
Fibre orientation misalignment traditionally changes from positive to
negative angle when going from negative to positive speed differential. The
misalignments shown in Figure 77 therefore would indicate that minimum
shear occurs slightly below jet/wire ratio1.04, rather than at 1.02.
From Figure 77 it is evident that flocculation at minimum mix-to-wire
shear during dewatering is substantially higher, especially at web centre. One
reason for the ever-present better formation close to the wires could be that
dewatering will initially move the free fibres to these positions. The more
flocculated state of mix would then be left to dewater in the later stages, at

536 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Figure 76 Layered orientation structure for roll-blade forming at 1200 m/min, LWC
50 g/m2. “Anisotropy” (Left) and Standard deviation of “anisotropy” (Right). Jet/wire
ratio on the low side; 1.02 corresponds to minimum shear during dewatering [315].

Figure 77 Layered floc index (left) and misalignment angle (right), three different
thickness ranges from sheet bottom to top; different jet/wire ratios, data as in
Figure 76 [315].

web centre. This could explain why shear during dewatering mainly has a
positive effect in the late dewatering phases – this is simply when the flocs are
to be dewatered.
The results reported by Erkkilä et al. also include rush conditions, but
space does not allow its presentation at this time.
The effect of differences in vacuum levels in forming roll surface and form-
ing shoes was also studied, but had only limited influence on sheet structure.
The mounting of vanes inside a headbox nozzle is well known to have a
decreasing effect on fibre orientation anisotropy in the final paper product.
As shown above in Figure 22, vane insertion reduces fibre orientation
anisotropy in the headbox jet.

12th Fundamental Research Symposium, Oxford, September 2001 537

B. Norman and D. Söderberg

Erkkilä et al. [315] studied the effects on the paper of headbox vanes and
roll-bade forming; see Figure 78.
It is obvious that there is no influence of vane length itself (with the actual
dimensions), only of end shape. However, the turbulent eddies generated by
the blunt vane end probably lasted all the way to the dewatering stage. The
reducing effect of turbulence on anisotropy and the negative effect on floc
index are clearly demonstrated.

Figure 78 The effect of different vanes on layered “anisotropy” (Left) and floc index
(Right). The 500 and 600 mm vanes end 220 and 120 mm respectively, before slice
opening [315].

Figure 79 shows a fibre orientation vector map, where anisotropy magni-

tude is represented by individual line length and orientation by line angle. The
high degree of local variations is clearly demonstrated.
Lloyd and Chalmers [45] applied sheet splitting using the tape-method and
optical fibre orientation analysis to characterise the structure of anisotropic

Figure 79 Fibre orientation map for layer next to sheet top using blunt tip vanes;
data as in Figure 78 [315].

538 Session 3: Forming

 Overview of Forming Literature, 1990–2000

laboratory sheets. It was also demonstrated that variations in the Z-direction

of fibre orientation misalignment and its variability correlated with the
occurrence of cockle on machine made paper.
Räisänen and Paananen discussed the influence of furnish properties on
linerboard quality using hybrid and roll-blade forming principles [178].
Mohlin [316] performed pulp evaluation using the EuroFEX system.
Comparison with standard pulp evaluation techniques demonstrated that
different ranking orders could be obtained in the two cases.
Mohlin [317] also studied the influence of fibre dimensions on formation
and strength properties using the pilot machine EuroFEX. 100% chemical
pulp and also LWC-type furnishes were tested. Small-scale formation was
improved by decreasing fibre coarseness and increasing fines content and
large-scale formation was also influenced by fibre flocculation. The ratio
between the two formation numbers was suggested to be a measure of floc
formation, and decreased with decreasing crowding factor. There was no
general correlation between formation and strength. About the same strength
was obtained on the fourdrinier and in the twin-wire roll-adjustable blade
former, in spite of their very different formation properties. Fibre flocculation
had a larger effect on burst strength than on tensile strength.
Odell and Pakarinen [318] made a recent overview, aiming at fibre orienta-
tion related defects on different scales, and the effects on curl, and local
dimensional instabilities.
An interesting application of layered orientation information as that
described above is to construct a 3-D orientation map of the paper cross-
section; see Figure 80. Local orientation and orientation variability are
interpreted as effective shear and turbulence present during dewatering.
However, it should be pointed out that some of the structure character-
istics in the final paper might originate from upstream conditions in the
headbox.

9.2 Multi-layer forming

Traditionally, multi-layer forming of tissue products has been applied
industrially already since the 1970s. Some recent headbox flow investigations
are included under Headboxes. However, probably due to the secrecy within
the tissue area, no results from recent tissue applications can be found in
literature.
Already during the 1970s, a three-layer headbox was installed on a new 10
m wide Bel Baie twin-wire linerboard machine; Page and Hergert [319]. 25%
mixed waste was placed in the centre layer and it was claimed that the reduc-
tion in burst value was only 6% in comparison to a 20% loss using single layer

12th Fundamental Research Symposium, Oxford, September 2001 539

B. Norman and D. Söderberg

Figure 80 CD cross-section of a paper sample. Local MD orientation anisotropy

(Top) and variability in CD anisotropy (Bottom), [318].

forming on a fourdrinier machine. Page and Hergert also discuss the advan-
tages for paper mechanical properties using different layering in alternative
cases.
During the 1990s several installations of 2-layer headboxes on linerboard
machines have been made. This makes it possible to manufacture white-top
liner with only one forming unit.
During the last decade much work has been performed to improve the
quality of 3-layered printing papers (see section Headboxes). One main prob-
lem in comparison with 2-layer linerboard forming is the low grammage of
the surface layers; brighter surface layers cannot optically cover a darker
centre layer. The centre layer brightness therefore should not deviate too
much from that of the surfaces.
Häggblom-Ahnger [320] made several pilot-plant trials to study three-layer
forming of printing papers. Häggblom-Ahnger and Eklund [321] studied the
effects of a CTMP middle layer in a 3-layer copy paper. Positive effects on
mechanical properties were found, but the surface layer cover was not even
enough.
In a study of the location of softwood fibres (30%) in a three-layer office
paper bending stiffness was found to increase with a central placement;

540 Session 3: Forming

 Overview of Forming Literature, 1990–2000

Häggblom-Ahnger [322]. At the same time surface smoothness and forma-

tion improved. For runnability purposes, the amount of softwood fibres
could be reduced with central placement.
Filler distribution in the Z-direction can be controlled using selective add-
ition in three-layer forming [323,314]. There is also a large control potential
through selective retention aid addition in the different layers.
An overview of layering possibilities was made by Lloyd [324].

9.3 Multi-ply forming

Current methods for single-ply forming and combinations into multi-ply
products are summarised in Wire section designs. There is no relevant litera-
ture available from the last decade, on the multi-ply aspects of board
properties.

ACKNOWLEDGEMENTS

Dr Hannes Vomhoff is thanked for valuable views on the original manuscript

and Tekn-Lic student Greger Asplund for skilful help with many of the
illustrations.

REFERENCES

1. Norman, B., “Overview of the physics of forming”, in 9th Fundamental Research

Symposium, Cambridge, pp. 3: 73–149 (1989).
2. Lindström, T., “Overview of the chemistry of forming”, in 9th Fundamental
Research Symposium, Cambridge, pp. 311–412 (1989).
3. Norman, B., “Web forming”, Papermaking Part 1, Stock Preparation and Wet
End in Papermaking Science and Technology, Fapet OY, Helsinki, 8(6): pp.
191–250 (2000).
4. Beghello, L., et al., “A device for measuring fiber floc sizes in highly turbulent
fibre suspensions”, Nord. Pulp Pap. Res. J., 11(4): pp. 254–259 (1996).
5. Keller, S., Luner, P., and Pawlak, J., “Characterization of the spatial variability of
paper formation using a continuous wavelet transform”, in International Paper
Physics Conference, San Diego, pp. 357–368 (1999).
6. Keller, S., Lewalle, J. and Luner, P., “Analysis of paper variability using the
continuous wavelet transform”, Paperi ja Puu, 81(6): pp. 440–446 (1999).
7. Söderberg, D., “A comparison between the flow from a paper machine headbox
and a low Reynolds number water jet”, in Tappi Engineering Conference,
Anaheim, pp. 1155–1172 (1999).

12th Fundamental Research Symposium, Oxford, September 2001 541

B. Norman and D. Söderberg

8. Johansson, P. Å. and Norman, B., “Methods for evaluating formation, print

unevenness and gloss variations developed at STFI”, in Tappi Process and Prod-
uct Quality Conference, Cincinnati, p. 140 (1996).
9. Cresson, T., Tomimasu, H. and Luner, P., “Characterisation of paper formation.
Part 1: Sensing of paper formation”, Tappi J., 73(7): pp. 153–159 (1990).
10. Cresson, T. and Luner, P., “The characterization of paper formation – Part 2: The
texture analysis of paper formation”, Tappi J., 73(12): pp. 175–185 (1990).
11. Cresson, T. and Luner, P., “The characterization of paper formation – Part 3: The
use of texture maps to describe formation”, Tappi J., 74(2): pp. 167–175 (1991).
12. Cresson, T. and Luner, P., “Description of the spatial grey level dependence
method algorithm”, Tappi J., 73(12): pp. 220–222 (1990).
13. Dodson, C., “A rational approach to formation analysis and spreadsheet schema
for data interpretation”, Tappi J., 76(5): pp. 153–156 (1993).
14. Ambertec, “Beta formation tester”, Ambertec OY, Espoo, Finland. (1990).
15. Waterhouse, J., “Effect of papermaking variables on formation”, Tappi J., 76(1):
pp. 129–134 (1993).
16. Tomimasu, H., et al., “Comparison of four paper imaging techniques:
b-radiography, electrography, light transmission and soft X-radiography”, Tappi
J., 74(7): pp. 165–176 (1991).
17. Keller, S., “Paper formation measurement by electron beam transmission imaging
and analysis using wavelet transforms”, SUNY, Syracuse, PhD Thesis (1996).
18. Dodson, C., “A universal law of formation”, (Research Note). J. Pulp Pap. Sci.,
16(4): pp. 136–137 (1990).
19. Lloyd, M., “Turbulent layer mixing during stratified forming of woodfree paper”,
Pulp and Paper Chemistry and Technology, KTH, Stockholm, Licentiate thesis
(1997).
20. Deng, M. and Dodson, C., “Paper: An engineered stochastic structure”, Tappi
Press, Atlanta (1994).
21. Trepanier, R., Jordan, B. and Nguyen, N., “Specific perimmeter: a statistic for
assessing formation and print quality by image analysis”, Tappi J., 81(10):
pp. 191–196 (1998).
22. I’Anson, S., “Identification of periodic marks in paper and board by image
analysis using two-dimensional fast Fourier transforms. Part 1: The basics”, Tappi
J., 78(3): pp. 113–119 (1995).
23. I’Anson, S., “Identification of periodic marks in paper and board by image
analysis using two-dimensional fast Fourier transforms. Part 2: Forming and press
section marks”, Tappi J., 78(7): pp. 97–106 (1995).
24. Wahjudi, U., Duffy, G. and Kibblewhite, P., “Effect of grammage and concentra-
tion on paper sheet formation of pinus radiata kraft pulps”, in Appita Annual
Conference, pp. 671–678 (2000).
25. Bernié, J.-P. and Douglas, M., “Local grammage distribution and formation of
paper by light transmission image analysis”, Tappi J., 79(1): pp. 193–202 (1996).
26. Bouydain, M., Colom, J. and P. J, “Using wavelets to determine paper formation
by light transmission image analysis”, Tappi J., 82(7): pp. 153–158 (1999).

542 Session 3: Forming

 Overview of Forming Literature, 1990–2000

27. Daunais, R. and Garner, R., “The NUI formation tester – an evaluation, modifi-
cation and comparison with other techniques”, in Inernational Paper Physics
Conference, Mont Gabriel, p. 43 (1987).
28. Kallmes, O. and Ayer, J., “Light scanning system provides qualitative formation
measurement”, in International Paper Physics Conference, Mont Gabriel, p. 209
(1987).
29. Popil, R. and Jordan, B., “The wavelength dependence of transmitted light forma-
tion measurement”, in International Paper Physics Conference, pp. 61–69 (1995).
30. Lloyd, M. and Norman, B., “Effect of headbox slice geometry during the stratified
forming of woodfree paper”, Nord. Pulp Paper Res. J., 12(4): pp. 252–259 (1997).
31. Moineau, D., Bloch, J-F. and Eymin, G., “On-line 2D formation sensor: A new
tool to evaluate the quality of paper formation”, in International Paper Physics
Seminar, Grenoble, pp. 29–38 (2000).
32. Hiertner, M., “Basis weight varies faster than on-line systems can measure”, in
Control Systems ’98, Porvoo, p. 307 (1998).
33. Ferguson, H., “Full sheet image system becomes control reality”, Pulp & Paper,
71(10): p. 75 (1997).
34. Mathias, T., “Paper industry’s first full sheet imaging system”, in 52nd Appita
Annual Conference, pp. 429–433 (1998).
35. Hagart-Alexander, C., “Wet end measurement system for MD and CD quality
monitoring and control”, in 86th Annual Meeting, PAPTAC, Monteral, pp.
231–233 (2000).
36. Niskanan, K., Kajanto, I. and Pakarinen, P., “Fibre orientation”, Paper Physics,
Papermaking Science and Technology, Fapet OY, Helsinki, 16(1.4): pp. 37–50
(1998).
37. Titus, M., “Ultrasonic technology – measurements of paper’s orientation and
elastic properties”, Tappi J., 77(1): pp. 127–130 (1994).
38. Nomura, T., “Introduction of new SST-3000 and development of sonic sheet
tester uses”. Japan Tappi J., 48(1): pp. 215–219 (1994).
39. Bando, T., et al., “Drainage mechanism on a twin-wire former. Part 3: Depend-
ence of the structural anisotropy of paper sheet on the pulsating drainage”, Japan
Tappi, 49(49): pp. 841–848 (1995).
40. Masuda, K., Hasuike, M. and Bando T., “Characterization on 3-D geometric
arrangement of fibres in a paper sheet”, in ISF, Yokohama, p. 243 (1994).
41. Hasuike, M., Kawasaki, T. and Murakami, K., “Evaluation method of 3-D
geometric structure of paper sheet”, J. Pulp Pap. Sci., 18(3): pp. J114–120 (1992).
42. Hasuike, M., Masuda, K and Bando, T., “Characterization of three-dimensional
sheet structure”, Mitsubishi Heavy Industries, Technical Review, 32(3), 127–131
(1995).
43. Erkkilä, A.-L., Pakarinen, P. and Odell, M., “Sheet forming studies using layered
orientation analysis”, Pulp Paper Canada, 99(1): pp. 81–85 (1998).
44. Jansson, M., “Fibre orientation anisotropy – variations in the z-direction”, (in
Swedish). Pulp and Paper Chemistry and Technology, KTH, Stockholm, MSc
Thesis (1998).

12th Fundamental Research Symposium, Oxford, September 2001 543

B. Norman and D. Söderberg

45. Lloyd, M. and Chalmers, I., “Use of image orientation analysis technique to
investigate sheet structural problems during forming”, in APPITA, Melbourne,
pp. 495–502 (2000).
46. Scharcanski, J. and Dodson, C., “Local spatial anisotropy and its variability”, J.
Pulp Pap. Sci., 25(11): pp. 393–397 (1999).
47. Thorpe, J., “Exploring fiber orientation within copy paper”, in International
Paper Physics Conference, San Diego, pp. 447–457 (1999).
48. Xu, L., Parker, I. and Filonenko, Y., “A new technique for determining fibre
orientation distribution through paper”, in International Paper Physics Confer-
ence, San Diego, pp. 421–427 (1999).
49. Kao, S. and Mason, S., “Dispersion of particles by shear”, Nature, 253(February
20): pp. 619–621 (1975).
50. Norman, B. et al., “Hydrodynamics of papermaking fibres in water suspension”,
in Fibre-Water Interactions in Paper-Making, Oxford, pp. 195–250 (1977).
51. Steen, M., “Modeling fiber flocculation in turbulent flow: a numerical study”,
Tappi J., 74(9): pp. 175–181 (1991).
52. Kerekes, R. and Schell, C., “Characterization of Fibre Flocculation Regimes by a
Crowding Factor”, J. Pulp Pap. Sci., 18(1): pp. J32–38 (1992).
53. Kiviranta, A. and Dodson, C., “Evaluating fourdrinier formation performance”,
J. Pulp Pap. Sci., 21(11): pp. J379–384 (1995).
54. Kerekes, R. “Perspective on fibre flocculation in papermaking”, in International
Paper Physics Conference, Niagara, pp. 23–31 (1995).
55. Andersson, S., “Netwok disruption and turbulence in fibre suspensions”, Dep of
Chemical Engineering and Design, CTH, Gothenburg, PhD Thesis (1998).
56. Wikström, T., “Flow of pulp suspensions at low shear rates”, Dep of Chemical
Engineering Design, CTH, Gothenburg, Licentiate Thesis (1998).
57. Wikström, T. and R. A, “Yield stress of pulp suspensions”, Nord. Pulp Paper Res.
J., 13(3): pp. 243–250 (1998).
58. Andersson, S. and Rasmuson, A., “The network strength of non-flocculated fibre
suspensions”, Nord. Pulp Paper Res. J., 14(1): pp. 61–70 (1999).
59. Andersson, S., Nordstrand, T. and Rasmuson, A., “The Influence of Some Fibre
and Suspension Properties on Pulp Fibre Friction”, J. Pulp Pap. Sci., 26(2): pp.
67–71 (2000).
60. Bennington, C. and Mmbaga, J., “Liquid phase turbulence in pulp fibre suspen-
sions”, in Fundamental Research Symposium, Oxford (2001).
61. Schmid, C. and Kingenberg, D., “Mechanical flocculation in flowing fiber
suspensions”, Phys. Rev. Letters, 84(2): pp. 290–293 (2000).
62. Huhtanen, J.-P., “Non-Newtonian Flows in Paper making”, Energy and Process
Engineering, Tampere University of Technology, Tampere, Techn.lic. Thesis
(1998).
63. Ross, R. and Klingenberg, D., “Simulation of flowing wood fibre suspensions”,
J. Pulp Pap. Sci., 24(12): p. 388 (1998).
64. Schmid, C., Switzer, L. and Klingenberg, D., “Simulation of fiber flocculation:
Effects of fiber properties and interfiber friction”, J. Rheol., 44, pp. 781–809 (2000).

544 Session 3: Forming

 Overview of Forming Literature, 1990–2000

65. Schmid, C. and Klingenberg, D., “Properties of fiber flocs with frictional and
attractive interfiber forces”. J. Coll. Int. Sci., 226, pp. 136–144 (2000).
66. Shah, P., et al., “The role of turbulent elongational stresses on deflocculation in
paper sheet formation”, Tappi J., 83(4): pp. 8 pages (2000).
67. Kaji, H., Monma, K. and Katsura, T., “Fractal analysis of flocculation in pulp
suspension”, in International Paper Physics Conference, Hawaii, pp. 291–297
(1991).
68. Kallmes, O., Kallmes, P. and Bishop, B., “Monitoring flocculation in the paper
machine”, Tappi J., 77(7): pp. 194–198 (1994).
69. Zhao, R. and Kerekes, R., “The effect of suspending liquid viscosity on fiber
flocculation”, Tappi J., 76(2): pp. 183–188 (1993).
70. Paul, T., Duffy, G. and Chen, D., “New insights into the flow of pulp suspen-
sions”, in 54th Appita Annual Conference, Melbourne, pp. 645–654 (2000).
71. Giri, M., Simonsen, J. and Rochefort, W., “Dispersion of pulp slurries using
carboxymethylcellulose”, Tappi J., 83(10): pp. 14 pages (2000).
72. Kerekes, R. and Schell, C., “Effects of fibre length and coarseness on pulp floccu-
lation”, Tappi J., 78(2): pp. 133–139 (1995).
73. Beghello, L., “The tendency of fibers to build flocs”, Laboratory of Paper
Chemistry, Åbo Akademi University, Åbo, PhD Thesis (1998).
74. Raghem-Moayed, A. and Kuhn, D., “Turbulent flocculation measurement”,
J. Pulp Pap. Sci., 26(4): pp. 163–165 (2000).
75. Raghem-Moayed, A., “Characterisation of fibre suspension flows at papermaking
consistencies”, Dept of Chemical Engineering and Applied Chemistry, University
of Toronto, Toronto, PhD Thesis (1999).
76. Wågberg, L. and Nordqvist, T., “Detection of polymer induced flocculation of
cellulosic fibres by image analysis”, Nord. Pulp Pap. Res. J., 14(3): pp. 247–255
(1999).
77. Pierre, C., “A new sensor for the measurement of fibre flocculation in the stock. A
tool for the sheet formation control”, in COST E14, Lisbon (2000).
78. Swerin, A., “Flocculation and fibre network strength in papermaking suspansions
flocculated by retention aid systems”, Division of Paper Technology, KTH,
Stockholm, PhD Thesis (1995).
79. Tuomisaari, M., Gullichsen J. and Hietaniemi, J., “Floc disruption in medium-
consistency fiber suspensions”, in International Paper Physics Conference, Hawai,
pp. 609–612 (1991).
80. Hietaniemi, J. and Gullichsen, J., “Floc disruption in medium consistency fibre
suspensions: Empirical results and their interpretation”, in International Paper
Physics Conference, Niagara-on-the-Lake, pp. P29–38 (1995).
81. Cichoracki, T., Gullichsen, J. and Paulapuro, H., “High consistency forming: A
new concept”, Tappi J., 84(3): pp. 8 pages (2001).
82. Björkman, U., “Flow of flocculated fibres”. Stockholm: Paper Technology, KTH.
283. (1999).
83. Helmer, R., Covey, G. and Lai, L., “Approach flow stock mixing”, in 52nd Appita
Annual Conference, pp. 223–228 (1998).

12th Fundamental Research Symposium, Oxford, September 2001 545

B. Norman and D. Söderberg

84. Ringnér, J. and Rasmuson, A., “Characterization of fibre suspensions using X-ray
computed tomography and image analysis”, Nord. Pulp Pap. Res. J., 15(4): pp.
319–325 (2000).
85. Kellomäki, M. et al., “Fiber flocculation measurement in pipe flow by digital
image analysis”, in International Paper Physics Conference, San Diego, pp.
461–463 (1999).
86. Karema, H. et al., “Transient fluidisation of fibre suspension in straight channel
flow”, in International Paper Physics Conference, San Diego, pp. 369–379 (1999).
87. Li, T.-Q., et al., “Velocity measurements of fiber suspensions in pipe flow by the
magnetic resonance imaging method”, Tappi J., 77(3): pp. 145 (1994).
88. Li, T.-Q. and Ödberg, L., “Flow of pulp suspension through an abrupt contrac-
tion studied by flow encoded nuclear magnetic resistance imaging”. Nord. Pulp
Paper Res. J., 10(2): pp. 133 (1995).
89. Li, T.-Q. et al., “Pipe flow behaviour of hardwood pulp suspensions studied by
NMRI”, J. Pulp Pap. Sci., 21(12): pp. J408–414 (1995).
90. Li, T.-Q. and Ödberg, L., “Studies of flocculation in cellulose fibre suspansions by
NMR imaging”, J Pulp Pap. Sci., 23(8): pp. J401 (1997).
91. Houvila, J. et al., “The Valmet Headbox Family-A New Level of Customer
Orientation”, in XII Paper Technology Days, Turku, pp. 128–139 (2000).
92. Pantaleo, S., “Headboxes – Past, Present and Future”, in CPPA Annual Meeting,
Montreal, pp. A131–139 (2000).
93. Schultz, H.-J. “The use of standard units for headbox design”, in 4th
International Conference: New available techniques and current trends, Bologna,
pp. 6–45 (1992).
94. Schultz, H.-J., “Der BTF-Universalverteiler. ein verbindenes Trennelement
zwichen kontasntem Teil und Stoffauflauf”, Wochenbl f. Papierfabr. (11/12): pp.
461–465 (1992).
95. Knoller, H., Foulger, M. and Parisian, J., “A new dilution control system for
existing headboxes”, in 86th Annual Meeting, PAPTAC, Montreal, pp. 17–21
(2000).
96. Begemann, U., “Module Jet, das neue Stoffauflaufkonzept mit Stoffdickte-
Querprofilregelung”, Das Papier, 47(10A): pp. V149 (1993).
97. Begemann, U., “The new headbox concept with CD consistency control”, in
Annual Meeting, Technical Section CPPA, Montreal, pp. A305–310 (1995).
98. Heinzmann, H., “Faserorientierung – Querprofil”, Wochenbl f. Papierfabr., (4):
pp. 121–126 (1995).
99. Pantaleo, S., “A new headbox design featuring consistency profiling decoupled
from fiber orientation response”, Tappi J., 78(11): pp. 89 (1995).
100. Pantaleo, S., “Basis weight profiling experience of the concept IV-MH headbox”,
in Annual Meeting, Technical Section CPPA, Monteal, pp. B193–199 (1996).
101. Nyberg, P. and Malashenko, A., “Dilution control headbox – choises, threats
and solutions”, in 83rd CPPA Annual Meeting, Montreal, pp. 17 (1997).
102. Wurster, H. and Weitkämper, K., “Stoffdichteabhängige Querprofilregelung –
Betriebserfahrungen”, Das Papier, 49(10A): pp. V99–102 (1995).

546 Session 3: Forming

 Overview of Forming Literature, 1990–2000

103. Wyse, R. et al., “Consistency profiling – A new technique for CD basis weight
control”, in Annual Meeting, Technical Section CPPA, Montreal, pp. A267–273
(1995).
104. Heaven, M. et al., “Recent Experiences with Consistency Profile Control Instal-
lations”, in 84th Annual Meeting, Technical Section CPPA, Montreal, pp. 261
(1998).
105. Smits, M. “Results from a consistency profiling headbox and a method of
analysing the CD profile performance”, in 52nd Appita Annual Conference, pp.
397–401 (1998).
106. Pantaleo, S. “A headbox design minimizing MD fiber orientation streaks while
maintaining high resolution basis weight control”, in 85th Annual Meeting,
PAPTAC, Montreal, pp. 113–119 (1999).
107. Collins, D., R. Leclerc and J. Perrault. “Operation of a moduljet dilution
headbox on Dolbeau PM5”, in 85th Annual Meeting, PAPTAC, Montreal, pp.
101–106 (1999).
108. Lee, J. and Pantaleo, S., “Headbox Flow Analysis”, J. Pulp Pap. Sci., 25(12):
pp. 437–441 (1999).
109. Hämäläinen, J. and P. Tarvainen. “CFD-based shape and control optimization
applied to a paper machine headbox”, in 86th Annual Meeting, PAPTAC, Mon-
treal, pp. 99–102 (2000).
110. Fortier, L. “A new family of high turbulence headboxes”, in 76th Annual Meet-
ing, PAPTAC, Montreal, pp. 197–201 (1990).
111. Haptmann, E., Vyse, R. and Mardon, J., “The wake effect as applied to modern
hydraulic headboxes, Parts 1 & 2”, Pulp Paper Can., 91(9: 10): pp. T357–364 and
T369–377 (1990).
112. Shariati, M. et al., “Numerical and experimental models of flow in the con-
verging section of a headbox”, in Tappi Papermakers Conference, Vancouver,
pp. 685–693 (2000).
113. Aidun, C., “Opportunities to improve CD strength and to reduce fibre com-
sumption in boardmaking”, in PIRA International Conference: Scientific and
Technical Advances in Boardmaking, Heathrow (1999).
114. Aidun, C., “Methods and apparatus to enhance paper and board forming
qualities”, IPST patent, US 6 153 057 (2000).
115. Nordström, B., “Effects of headbox design and dewatering conditions on twin-
wire forming of TMP”, Pulp and Paper Chemistry and Technology, KTH,
Stockholm, PhD Thesis (1995).
116. Nordström, B. and Norman, B., “Influence on sheet anisotropy, formation,
Z-toughness and tensile stiffness of reduced feed area to a headbox nozzle”,
Nord. Pulp Pap. Res. J., 9(1): p. 53 (1994).
117. Nordström, B. and Norman, B., “Effects of roll-blade and roll forming on
formation, retention, mecanical properties and anisotropy – A preliminary
comparison”, J. Pulp Pap. Sci., 21(7): pp. J223–229 (1995).
118. Ullmar, M., “On fiber alignment mechanisms in a headbox nozzle”, Pulp and
Paper Chemistry and Technology, KTH, Stockholm, Tekn Lic Thesis (1998).

12th Fundamental Research Symposium, Oxford, September 2001 547

B. Norman and D. Söderberg

119. Ullmar, M. and Norman, B., “Observation of fiber orientation in a headbox

nozzle at low consistency”, in Tappi Engineering Conference, Nashville, pp. 2:
865 (1997).
120. Asplund, G., “Fibre orientation anisotropy at different levels in a headbox jet”,
(in Swedish) Division of Paper Technology, KTH, Stockholm, Masters Thesis
(1999).
121. Chuang, S., “Microturbulence generation in a papermacine headbox”, in Tappi
Engineering Conference, pp. 205–213 (1982).
122. Shands, J., “Mechanics of stock jump at jet impingement”, J. Pulp Pap. Sci.,
17(3): pp. J92–J99 (1991).
123. Farrington, T., “A numerical investigation of three tissue machine headboxes”,
in AIChE Proceedings, Forest Products Symposium, pp. 177–188 (1991).
124. Hua, L. et al., “Flow distribution in a hydraulic headbox”, in Tappi Engineering
Conference, Annaheim, pp. 1123–1133 (1999).
125. Bandhakavi, V. and Aidun, C., “Analysis of turbulent flow in the converging
zone of a headbox”, in Tappi Engineering Conference, Anaheim, pp. 1135–1154
(1999).
126. Parsheh, M. and Dahlkild, A., “Modelling the flow around elastic guiding vanes
in the converging zone of a headbox”, in Tappi Engineering/Process and Product
Quality Conference, Anaheim, pp. 1433–1452 (1999).
127. Parsheh, M., “Flow in contractions with applications to headboxes”, Dep of
Machanics, Kungl. Tekniska Högskolan, Stockholm (2001).
128. Baker, J., Beuther, P. and Farrington, T., “Impact of three dimensional headbox
flows on tissue layer purity”, in AIChE Symposium Series, Advances in Pulp and
Papermaking, pp. 132–142 (1995).
129. Lloyd, M. and Norman, B., “Layer mixing during three-layer stratified forming:
The role of vane length and mix-wire speed difference”, Tappi J., 81(11): pp.
194–208 (1998).
130. Lloyd, M. and Norman, B., “Use of stepped vanes to alter headbox turbulence
levels during stratified forming of woodfree paper”, Appita, 52(2): pp. 98–104
(1999).
131. Parsheh, M. and Dahlkild, A., “Numerical modelling of mixing in stratified
headbox jet”, in Tappi Engineering Conference, Nashville, pp. 1159–1173 (1997).
132. Li, A., Neill, E. and Rogers, T., “Fundamental studies of mixing in a stratified
headbox jet”, in Annual Meeting, PAPTAC, Montreal, pp. A53–56 (2000).
133. Li, A., Rogers, T. and Shands, J., “Mixing intensity analysis of a 2-layer stratified
headbox jet”, Tappi J., 83(12): pp. 16 pages (2000).
134. Norman, B., “Hydrodynamic developments in twin-wire forming – an overview.
Paperi ja Puu, 78(6–7): p. 376 (1996).
135. Matsson, J. and Alfredsson, H., “Curvature- and rotation-induced instabilities
in channel flow”, J. Fluid Mech., 210, pp. 537–563 (1990).
136. Alfredsson, H. and Matsubara, M., “Streaky structures in transition”, in Tran-
sitional Boundary Layers in Aeronautics, (ed.) H. v. Ingen, Elsevier, pp. 374–386
(1996).

548 Session 3: Forming

 Overview of Forming Literature, 1990–2000

137. Söderberg, D. and Alfredsson, H., “Experimental and theoretical investigations

of plane liquid jets”, Eur. J. Mech. B/Fluids, 17(5): pp. 689–737 (1998).
138. Söderberg, D. and Alfredsson, H., “Experiments concerning the creation of
streaky structures inside a planar water jet”, in Tappi Engineering Conference,
Nashville, pp. 1205–1222 (1997).
139. Söremark, C., Johansson, G. and Kiviranta, A., “Characterization and elimin-
ation of fibre orientation streaks”, in Tappi Engineering Conference, pp. 97–104
(1994).
140. Aidun, C. and Kovacs, A., “Hydrodynamics of the forming section: the origin of
non uniform fibre orientation”, Tappi J., 78(11): pp. 97–106 (1995).
141. Schlupp, K., Karrila, S. and Paulapuro, H., “An experimental study of a high
speed slice jet”, in Annual Meeting, Technical Section CPPA, Monteal, pp.
B199–205 (1995).
142. Schlupp, K., “Characteristics of a headbox slice jet at elevated speed”, Depart-
ment of Forest Products Technology, HUT, Helsinki, Licentiate Thesis (2000).
143. Lindqvist, A., “Structures in the flow from paper machine headboxes”, Division
of Fluid Mechanics, Luleå University of Technology, Luleå, Licentiate Thesis
(1996).
144. Li, A., Neill, E. and Rogers, T., “Jet trajectory from a converging nozzle”, in
Tappi Engineering Conference, Nashville, pp. 1197–1204 (1997).
145. Aidun, C., “Hydrodynamis of streaks on the forming table”, Tappi J., 80(8): pp.
155–162 (1997).
146. Aidun, C., “Quantitative evaluation of the forming jet delivered from four
different hydraulic headboxes using high speed digital imaging”, Tappi J., 81(4):
pp. 172–179 (1998).
147. Ono, K. et al., “Forming jet surface velocity profile measurements with high
speed digital imaging”, Tappi J., 84(3): pp. 13 pages (2001).
148. Söderberg, D. and Alfredsson, H., “Experiments concerning the origin of
streaky structures inside a plane water jet”, J. Pulp Pap. Sci., 26(11): pp. 395–400
(2000).
149. Baker, D. and Reimer, K., “Twin-wire gap forming: Leading edge technology for
newsprint in the 90s”, Pulp Paper Canada, 93(6): pp. T170–175 (1992).
150. Banning, J., “Vorrichtung zum Führen der Siebe einer Doppelsiebpartie einer
Papier-oder Kartongmaskin”, Dörries patent, DE 35 03 242 (1986).
151. Norman, B., “Pilot plant machines and the future”, in Pira Conference: New
technologies in multiply and multilayer structures, Bristol, pp. Forum panel
presentations (1991).
152. Norman, B. et al., “The STFI-former for roll/blade dewatering”, in 4th Inter-
national Conference New Available Techniques and Current Trends, Bologna,
pp. 114–121 (1992).
153. Nordström, B. and Norman, B., “Development of the STFI former”. Nord. Pulp
Pap. Res. J., 9(3): p. 176 (1994).
154. Voith, “PM 2000-neue Versuchspapiermaschine bei Voith”, Wochenbl f.Papier-
fabr., 119(3): p. 98 (1991).

12th Fundamental Research Symposium, Oxford, September 2001 549

B. Norman and D. Söderberg

155. Zippel, F., “Formation und retention bei Doppelsieb-Formern”, Wochenbl f.

Papierfabr., (18): pp. 802–808 (1990).
156. Bück, R. et al., “Twin-wire paper-web former”, Voith patent, WO 91/02842
(1991).
157. Schmidt-Rohr, V. and Kohl, B., “Der Duoformer CFD – eine Neuentwicklung
auf dem Gebiet Blattbildungssysteme”, Wochenbl. f. Papierfab. (11/12): pp.
455–460 (1992).
158. Schmidt-Rohr, V., “Betriebserfahrungen mit dem Roll-Gap-Former (Duoformer
CFD) bei grafischen Papieren”, Wochenbl. f. Papierfabr. (11): pp. 441–446
(1994).
159. Egelhof, D. and Bubik, A., “Der Einfluss des Maschinenbaus auf die
Blattbildung”, Wochenbl f. Papierfabr., (4): pp. 111–117 (1994).
160. Moser, H., “Forming technology today and tomorrow”, in 6th International
Conference on New Available Techniques, Stockholm, pp. 450–453 (1999).
161. Palm, C. and Cedra, I., “Inbetriebnahme und erste Erfahrungen mit einer
Online-SC-Papiermachine”, ipw Science and Technology ?(11): pp. T191–T194
(2000).
162. Moser, H., “DuoFormer TQv”, in 86th Annual Meeting, PAPTAC, Monteal,
pp. 255–256 (2000).
163. Moser, H. et al., “DouFormer TQv – a new twin wire former”, Together (10):
pp. 17–21 (2000).
164. Moser, H., “State of the art forming technology”, in 87th Annual Meeting
PAPTAC, Monteal, pp. C121–126 (2001).
165. Mohrhardt, G. et al., “Wet end, method for the production of a fibrous material
web”, Voith patent, WO 9828489, prior. Dec. (1996, 1998).
166. Partanen, H., “Web forming method and apparatus”, Valmet patent, EP
0397430, Prior. May (1989, 1994).
167. Ahonen, P. et al., “Former development for an environmetally concious paper
mill”, in 1st EcoPaperTech Conference, Helsinki, pp. 341–351 (1995).
168. Verkasalo, L. and Kinnunen, J., “Experiences of new forming technology”, in
6th International Conference on New Available Techniques, Stockholm, pp.
471–477 (1999).
169. Clarke, R., “The evolution of Beloit forming technology”, Pulp Paper Can.,
96(8), pp. T281–285 (1995).
170. Shands, J., Irwin, J. and Wildfong, V., “An examination of two former rebuild
options for fourdrinier machines producing light weight coated grades”, in Tappi
Engineering Conference, pp. 801–806 (1995).
171. Walcher, G., “Improved newsprint operation with a Bel Baie IV”, in Annual
Meeting, Technical Section CPPA, Montreal, pp. A109–113 (1997).
172. Lockie, M., “Filling in the gaps with new forming technology”, Pulp & Paper
Int., (February): pp. 45–47 (1998).
173. Crouse, J., “Personal communication, Beloit” (1998).
174. Iwata, H., Goto, D. and Matumoto, M., “Application of novel dewatering elem-
ent for high speed former”, in 2000 Pan Pacific Conference, pp. 75–82 (2000).

550 Session 3: Forming

 Overview of Forming Literature, 1990–2000

175. Attwood, B. and Moore, G., “An introduction to the theory and practice of
multiply forming”, Leatherhead: PIRA. 179. (1994).
176. Attwood, B. and Loy, K., “Pressure forming developments in multi-ply paper-
board forming”, Tappi J., 76(10): pp. 160–164 (1993).
177. Halmschlager, G., “Duoformer Top – a new former for the production of top
plies of packaging grades”, in Tappi Engineering Conference, Nashville, pp.
1239–1243 (1997).
178. Räisänen, K. and Paananen, J., “Forming process requirements set by a furnish
type in linerboard production”, in 6th International Conference on New
Available Techniques, Stockholm, pp. 454–457 (1999).
179. Halmschlager, G., “Douformer Base – compact concept for increased board and
packaging paper production”, in 6th International Conference on New Available
Techniques, Stockholm, pp. 460–464 (1999).
180. Norman, B., “Fibre coarseness (weight per unit length) should be replaced by
fibre grammage (weight per unit area)”, Nord. Pulp Paper Res. J., 13(2): p. 166
(1998).
181. Dodson, C., “The effect of fibre length distribution on formation (Research
note)”, J. Pulp Pap. Sci., 18(2): pp. J74–J76 (1992).
182. Dodson, C. and Fekih, K., “The effect of fibre orientation on paper formation”,
J. Pulp Pap. Sci., 17(6): pp. J203–J206 (1991).
183. Dodson, C., “Flocculation and orientation effects on paper formation statistics”,
Tappi J., 75(1): pp. 167–171 (1992).
184. Kuhn, D. and Dodson, C., “Measurement of fibre orientation effects on
formation”, in 78th Annual Meeting, Technical Section CPPA, Montreal,
pp. A283–A286 (1992).
185. Dodson, C. and Serafino, L., “Flocculation, dispersion and dynamic scenarios
for formation”, Nord. Pulp Paper Res. J., 8(2): pp. 264–272 (1993).
186. Deng, M. and Dodson, C., “Random star patterns and paper formation”, Tappi
J., 77(3): pp. 195–199 (1994).
187. Farnood, R., Dodson, C., and Loewen, S., “Modeling flocculation. Part I:
Random disc model”, J. Pulp Pap. Sci., 21(10): pp. J348–355 (1995).
188. Farnood, R. and Dodson, C., “The similarity law of formation”, in Inter-
national Paper Physics Conference, Niagara-on-the-Lake, pp. 5–12 (1995).
189. Farnood, R. et al., “Modelling flocculation: A gallery of simulated flocculated
papers”, Nord. Pulp Paper Res. J., 12(2): pp. 86–89 (1997).
190. Norman, B., “Letter to the editor”, Nord. Pulp Paper Res J., 12(3): p. 210
(1997).
191. Sampson, W. et al., “Hydrodynamic Smoothing in the Sheet Forming Process”,
J. Pulp Pap. Sci., 21(12): pp. J422–426 (1995).
192. Norman, B. et al., “The effect of localised dewatering on paper formation”, in
International Paper Physics Conference, Niagara-on-the-Lake, pp. 55–59 (1995).
193. Lucisano, M. and Norman, B., “The forming and properties of quasi-random
laboratory paper sheets”, in International Paper Physics Conference, San Diego,
pp. 331–340 (1999).

12th Fundamental Research Symposium, Oxford, September 2001 551

B. Norman and D. Söderberg

194. Kilpeläinen, R. et al., “Stock Preparation and Wet End. Part 1: forming fabrics”,
in Papermaking Science and Technology, Fapet OY, Helsinki 8(7): pp. 251–283
(2000).
195. Johnson, D., “Effects of jet impingement on Bal Baie machines”, Pulp Paper
Canada, 95(5): pp. 12–15 (1992).
196. Fejér, M., “Der Einfluss des Blattbildungsmediums Sieb auf die Formation”,
Wochenbl. f. Papierfab., (20): pp. 839–846 (1993).
197. Adanur, S., “Effects of forming fabric structural parameters on sheet properties,
Tappi J., 77(10): pp. 187–195 (1994).
198. Hampson, P., “Fibre length and fabric aperture”, in 52nd Appita Annual Confer-
ence, pp. 377–382 (1998).
199. Heinola, M. et al., “The influence of forming fabrics on the retention and drain-
age on gap formers”, in Annual Meeting, Technical Section CPPA, Montreal,
pp. A59–64 (1996).
200. Cole, S., “Huytexx – a new breed of forming fabric”, in 87th Annual Meeting,
PAPTAC, Montreal, pp. A5–7 (2001).
201. Keen, P. and Cole, S., “Unique forming fabric structures deliver added value”, in
87th Annual Meeting PAPTAC, Montreal, pp. A57–60 (2001).
202. Herzig, R. and Johnson, D., “Investigation of thin fiber mats formed at high
velocity”, Tappi J., 82(1): pp. 226–230 (1999).
203. Meyer, H., “The effect of wire screens on forming fiber mats”, Tappi J., 52(9):
pp. 1716–1723 (1969).
204. Danby, R., “The impact of forming fabric structures on print quality”, in CPPA
Annual Meeting, Montreal, pp. 33–38 (1993).
205. Danby, R., “Forming fabric design definition and selection – The means to
machine and fiber efficiency”, in 85th Annual Meeting, PAPTAC, Montreal
(1999).
206. Danby, R. and Plouffe, P., “Print quality improvements through forming fabric
design changes”, Pulp Paper Can., 101(9): pp. 66–69 (2000).
207. Dudley, W. et al., “Engineered approach and unique forming fabric design pro-
vide mill with dramatic paper & print improvements”, in Tappi Engineering
Conference, Atlanta, pp. 9 pages (2000).
208. Jong, J. H., Baines, W. D. and Currie, I. G., “Experimental Characteristics of
Forming Fabrics and Fibre Mats”, J. Pulp Pap. Sci., 25(3): pp. 95–99 (1999).
209. McDonald, D., “Web rewetting by forming fabrics”, in Tappi Engineering
Conference, Anaheim, pp. 597–599 (1999).
210. Barratte, C. et al., “Simulation of initial fibre retention by forming fabric”, in
54th Appita Annual Conference, Melbourne, pp. 415–420 (2000).
211. Keen, P., “Bringing virtual reality to forming fabric design”, Paper Technology
(December): pp. 47–50 (1999).
212. Sayegh, N. and Gonzales, T., “Compressibility of Fibre Mats During Drainage”,
J. Pulp. Pap. Sci., 21(7): pp. J255–261 (1995).
213. Vomhoff, H. and Schmidt, A., “The steady-state compressibility of saturated
finre webs at low pressure”, Nord. Pulp Pap. Res. J., 12(4): pp. 267–269 (1997).

552 Session 3: Forming

 Overview of Forming Literature, 1990–2000

214. Jong, J., “Characteristics of jet impingement, drainage and compression in a

forming roll of a twin-wire machine”, Department of Mechanical and Industrial
Engineering, University of Toronto, Toronto, PhD Thesis (1998).
215. Sampson, W. and Bridle, I., “Simulation of the directional dependence od flow
through fibre mats”, Nord. Pulp Paper Res. J., 10(2): pp. 145–149 (1995).
216. Sampson, W. and Kropholler, H., “Batch-drainage curves for pulp characteriza-
tion. Part 1: Experimental”, Tappi J., 78(12): pp. 145–151 (1995).
217. Sampson, W. and Kropholler, H., “Batch-drainage curves for pulp characteriza-
tion. Part 2: Modeling”, Tappi J., 79(1): pp. 151–160 (1996).
218. Mantar, E., Co, A. and Genco, J.M., “Drainage Characteristics of Pulp
Slurries”, J. Pulp Pap. Sci., 21(2): pp. J44–J50 (1995).
219. Ramarao, B. and Kumar, P., “A Model for the Gravity Drainage of Paper-
making Suspensions”, Nord. Pulp Paper Res. J., 11(2): p. 86 (1996).
220. Wildfong, V. et al., “Filtration Mechanics of Sheet Forming. Part I: Apparatus
for Determination of Constant-Pressure Filtration Resistance”, J. Pulp Pap. Sci.,
26(7): pp. 250–254 (2000).
221. Wildfong, V. et al., “Filtration Mechanics of Sheet Forming. Part II: Influence
of Fine Material and Compression”, J. Pulp Pap. Sci., 26(8): pp. 280–283 (2000).
222. Hung, L., Leung, W. and Green, S., “Pulp fibre mat permeability measurements
under simulated forming conditions”, in Tappi Engineering Conference,
Atlanta, pp. 5 pages (2000).
223. Duplantis, W. and Green, S., “Measurements of fibre mat permeability with a
new device”, in 87th Annual Meeting PAPTAC, Montreal, pp. B7–12 (2001).
224. Audenis, G., “On the impingement of a plane liquid jet on the wires of a paper
machine”, Dep. of Mechanics, KTH, Stockholm, Licentiate Thesis (1999).
225. Dalpke, B., Green, S. and Kerekes, R., “Modelling of jet impingement in twin-
wire paper-machines: Impingement on one fabric”, Tappi J., 84(2): pp. 19 pages
(2000).
226. Loewen, S. and Buttler, A., “Laser speed measurement of forming section flow
properties”, Pulp Paper Can., 93(12): pp. 161–164 (1992).
227. Loewen, S., “Fibre orientation optimization”, Pulp Paper Can., 98(10): pp.
T391–396 (1997).
228. Farnood, R., Loewen, S. and Dodson, C., “Machine direction profiles of stock
speed and table activity”, in Annual Meeting, Technical Section CPPA,
Montreal, pp. B241–245 (1997).
229. Kiviranta, A., “Table activity on the fourdrinier: Its characterization and its
effect on formation”, Laboratory of Paper Technology, HUT, Helsinki, PhD
Thesis (1993).
230. Kiviranta, A. et al., “Measurement of table activity on a Fourdrinier paper
machine”, J. Pulp Pap. Sci., 19(6): p. J234 (1993).
231. Kiviranta, A. “Characterization and optimization of Fourdrinier table activity
on the middle ply of a folding boxboard machine”, Paperi ja Puu, 74(8): pp.
640–649 (1992).
232. Kiviranta, A. and Paulapuro, H., “The role of table activity in the manufacture

12th Fundamental Research Symposium, Oxford, September 2001 553

B. Norman and D. Söderberg

of fine paper”, in 4th New Available Techniques and Current Trends Conference,
Bologna, pp. 46–72 (1992).
233. Kiviranta, A. and Paulapuro, H., “Characterization and optimization of
Fourdrinier table activity in linerboard manufacture”, Paperi ja Puu, 74(9): pp.
728–737 (1992).
234. Kiviranta, A. and Paulapuro, H., “The role of fourdrinier table activity in the
manufacture of various paper and board grades”, Tappi J., 75(4): pp. 172–186
(1992).
235. Ahonen, P., Kiviranta, A. and Kaipila, J., “Effect of Fourdrinier table layout on
the wet-end performance of a fine paper machine”, Paperi ja Puu, 74(10): pp.
802–806 (1992).
236. Foulger, M., “Submerged drainage: A new concept of dewatering”, in Annual
Meeting, Technical Section CPPA, Montreal, pp. A65–A68 (1995).
237. Zu, J. and Chen, J., “A theoretical study of vacuum force on gravity foils and
step foils”, Tappi J., 82(11): pp. 93–98 (1999).
238. Evans, D. and Fielding, S., “VID formation system”, in 54th Appita Annual
Conference, Melbourne, pp. 437–442 (2000).
239. Palm, D. and Sodergren, O., “VID dormation system”, in 87th Annual Meeting
PAPTAC, Montreal, pp. A185–190 (2001).
240. Sodergren, O. and Neun, J., “Developments in activity generation on
Fourdriniers”, Tappi J., 83(10): pp. 10 pages (2000).
241. Verkasalo, L., Odell, M. and Korhonen, H., “Forming of LWC base papers”, in
Tappi Engineering Conference, Nashville, pp. 1231–1238 (1997).
242. Bando, T., Iwata, H. and Nagano, A., “Drainage mechanism on a twin-wire
former. Part 1: Factors affecting on the drainage phenomena”, Japan Tappi,
48(7): pp. 948–954 (1994).
243. Bando, T., Adachi, T. and Iwata, H., “Drainage mechanism on a twin-wire
former. Part 2: Computer simulation on pulsating drainage”, Japan Tappi,
48(11): pp. 1493–1498 (1994).
244. Iwata, H. et al., “Development of Mitsubishi new former for papermaking
machine”. Mitsubishi Heavy Industries, Technical Review, Technical Review,
Vol. 30, No. 2, 123–128 (1993).
245. Zhao, R. and Kerekes, R., “Pressure distribution between forming fabrics in
blade gap formers: Thin blades”, J. Pulp Pap. Sci., 21(3): pp. 97–103 (1995).
246. Zhao, R. and Kerekes, R., “The effect of consistency on pressure pulses in blade
gap formers”, Paperi Ja Puu, 78(1/2): pp. 36–38 (1996).
247. Nigam, M. and Bark, F., “An analytical method to calculate the flow past a
blade in twin-wire formers”, Dep of Mechanics, KTH, Stockholm, 1997: 7
(1997).
248. Zahrai, S., “On the fluid mechanics of twin-wire formers”, Dep of Mechanics,
KTH, Stockholm, PhD Thesis (1997).
249. Zahrai, S. and Bark, F., “On the fluid mechanics of twin wire blade forming in
paper machines”, Nord. Pulp Pap. Res. J., 10(4): pp. 245–252 (1995).
250. Zahrai, S. and Bark, F., “Analytical and numerical approaches to solve the

554 Session 3: Forming

 Overview of Forming Literature, 1990–2000

model equations for the flow of pulp in blade forming process”, in AIChE
Symposium Series, Pulping, Papermaking and Chemical Preparation, pp. 68–73
(1996).
251. Zahrai, S., Bark, F. and Norman, B., “An analysis of blade dewatering in a
twin-wire paper machine”, J. Pulp Pap. Sci., 23(9): pp. J452–459 (1997).
252. Green, S., “Analytical and computational modeling of twin-wire blade forming”,
J. Pulp Pap. Sci., 23(7): pp. J353–358 (1997).
253. Roshanzamir, A., Green, S. and Kerekes, R., “Two-Dimensional Simulation of
Pressure Pulses in Blade Gap Formers”, J. Pulp Pap. Sci., 24(11): pp. 364–368
(1998).
254. Green, S., Zhao, R. and Kerekes, R., “Pressure Distribution Between Forming
Fabrics in Blade Gap Formers: Blades of Finite Width and Fabrics of Finite
Stiffness”, J. Pulp Pap. Sci., 24(2): pp. 60–66 (1998).
255. Green, S. and Kerekes, R., “Numerical analysis of pressure pulses induced by
blades in gap formers”, Tappi J., 81(4): pp. 180–187 (1998).
256. Roshanzamir, A. et al., “Hydrodynamic Pressure Generated by Doctoring in
Bade Gap Formers”, in Tappi Engineering Conference, pp. 1181–1189 (1999).
257. Roshanzamir, A., Green, S. and Kerekes, R., “The Effect of Non-Darcy’s Law
Drainage on the Hydrodynamics of Blade Gap Formers”, in Tappi Papermakers
Conference, Vancouver, pp. 701–708 (2000).
258. Shands, J. and Wildfong, V., “A twin wire former rebuild option for improved
formation and drainage”, in Tappi Engineering Conference, pp. 53–69 (1998).
259. Odell, M., Pakarinen, P. and Luontama, M., “Customizing roll and blade form-
ing to control paper structure”, in Valmet Paper Machine Days, Jyväskylä, pp.
21–42 (1996).
260. Green, S., “Modeling suction shoes in twin-wire blade forming: Theory”, Tappi
J., 82(9): pp. 136–142 (1999).
261. Green, S., “Modelling Suction Shoes in Twin-Wire Forming: Results”, J. Pulp
Pap. Sci., 26(2): pp. 53–58 (2000).
262. Roshanzamir, A., Green, S. and Kerekes, R., “Two-Dimensional Simulation of
Suction Shoes in Gap Formers”, J. Pulp Pap. Sci., 26(4): pp. 158–162 (2000).
263. Martinez, M., “Characterizing the dewatering rate in roll gap formers”, J. Pulp
Pap. Sci., 24(1): pp. 7–13 (1998).
264. Boxer, T., “Quantitative modelling of an industrial roll-blade gap former”,
Department of Paper Science, UMIST, Manchester, PhD Thesis (1999).
265. Boxer, T., Dodson, C. and Sampson, W., “Analytic solution to the Martinez
dewatering equations for roll gap formers”, J Pulp Pap. Sci., 26(11): pp. 391–394
(2000).
266. Zahrai, S., Martinez, M. and Dahlkild, A., “Estimating the thickness of the web
during twin-wire forming”, J. Pulp Pap. Sci., 24(2): pp. 67–72 (1998).
267. Zahrai, S., Bark, F. and Martinez, M., “A numerical study of cake formation in
2-D cross-flow filtration”, J. Pulp Pap. Sci., 24(9): pp. 281–285 (1998).
268. Turnbull, P. et al., “One-dimensional dynamic model of a paper forming
process”, Tappi J., 80(1): pp. 245–252 (1997).

12th Fundamental Research Symposium, Oxford, September 2001 555

B. Norman and D. Söderberg

269. Chen, E., Schultz, W. and Perkins, N., “A two-dimensional viscous model
of a wet paper forming process”, in Tappi Engineering Conference, pp. 21
(1998).
270. Malashenko, A. and Karlsson, M., “Twin Wire Forming – An Overview”, in
Annual Meeting PAPTAC, Montreal, pp. A189–201 (2000).
271. Wildfong, V. et al., “Drainage during roll forming – Model validation using pilot
papermachine data”, in Tappi Engineering Conference, Atlanta, pp. 11 pages
(2000).
272. Gooding, R., McDonald, D. and Rompré A., “Measurement of drainage
around a forming roll”, in 87th Annual Meeting, PAPTAC, Montreal, pp.
A125–138 (2001).
273. Neun, J., “Performance of high vacuum dewatering elements in the forming
section”, Tappi J., 77(9): pp. 133–138 (1994).
274. Neun, J. and Fielding, S. “High vacuum dewatering optimization”, in Tappi
Papermakers Conference, pp. 307–312 (1994).
275. Neun, J., “High vacuum dewatering of brown paper grades”, in Tappi Paper-
makers Conference, pp. 259–265 (1995).
276. Neun, J., “High-vacuum dewatering of newsprint”, Tappi J., 79(9): pp. 153–157,
(1996).
277. Gagnon, J. and Neun, J., “High vacuum dewatering on fourdriniers and for-
mers”, in Annual Meeting, Technical Section CPPA, Montreal (1996).
278. Räisenen, K., “Water removal by flat boxes and a couch roll on a paper machine
wire section”, Laboratory of Paper Technology, HUT, Helsinki, PhD Thesis
(1998).
279. Räisänen, K., “High-vacuum dewatering on a paper machine wire section – a
literature review”, Paperi ja Puu (Papper och Trä), 78(3): pp. 113–120 (1996).
280. Räisänen, K., Paulapuro, H. and Karrila, S., “The effects of retention aids,
drainage conditions and pretreatment of slurry on high-vacuum dewatering; a
laboratory study”, Tappi J., 78(4): pp. 140–147 (1995).
281. Räisänen, K., Karrila, S. and Maijala, A., “Effect of vacuum level and suction
time on vacuum assisted drainage of a paper machine wire section”, Appita J.,
48(4), pp. 269–274 (1995).
282. Räisänen, K., Karilla, S. and Maijala, A., “Vacuum dewatering optimization
with different furnishes”, Paperi ja Puu (Papper och Trä), 78(8): pp. 461–467
(1996).
283. Mitchell, C. and Johnston, R., “Pulsating suction during vacuum dewatering
and its effect on the rate and extent of water removal”, in 54th Appita Annual
Conference, Melbourne, pp. 443–447 (2000).
284. Shands, J. and Hardwick, C., “Dewatering on a high vacuum flat box”, in 86th
Annual Meeting, PAPTAC, Montreal, pp. A113–117 (2000).
285. Jones, G., “Dynamic simulation of dewatering in high vacuum flat boxes”, in
Tappi Engineering Conference, Anaheim, pp. 197–209 (1999).
286. Nazhad, M. and Kerekes, R., “Producing handsheets at headbox consistency”,
in Annual Meeting, PAPTAC, Montreal, pp. A203–208 (1999).

556 Session 3: Forming

 Overview of Forming Literature, 1990–2000

287. Nazhad, M. et al., “The influence of formation on tensile strength of paper

made from mechanical pulps”, Tappi J., 83(12): pp. 9 pages (2000).
288. Lin, J. and Schuster, M., “Predicting drainage retention and formation develop-
ment on high alkaline paper machines”, in Tappi Papermakers Conference,
pp. 253–266 (1992).
289. Sutman, F., “Development and validation of a new drainage testing method-
ology”, in Tappi Papermakers Conference (1999).
290. Karrila, S., Räisänen, K. and Paulapuro, H., “The moving belt drainage tester
(MBDT)”, in Tappi Papermakers Conference, Nashville, pp. 275–299 (1992).
291. Räisenen, K., Karrila, S. and Paulapuro, H., “Wire section siulation with the
moving belt drainage tester (MBDT)”, in Tappi Papermakers Conference,
Atlanta, pp. 103–113 (1993).
292. Xu, L. and Parker, I., “Simulating the forming process with the moving belt
drainage former”, in 53rd Appita Annual Conference, Rotorua, pp. 169–173
(1999).
293. Helmer, R. et al., “Preliminary development of a laboratory former for oriented
sheets”, in Appita J, pp. 465–470. (2000).
294. Helmer, R. et al., “Flow phenomena and paper forming”, Appita J., 54(1): pp.
54–60 (2001).
295. Sivén, S. and Manner, H., “High speed retention tester”, in 86th Annual
Meeting, PAPTAC, Montreal, pp. A37–41 (2000).
296. Sivén, S. and Manner, H., “Multicomponent retention systems in twin wire
forming”, in Apita Annual Meeting, pp. 215–220 (2000).
297. Hammock, C. and Garnier, G., “The laboratory twin-former. Part 1: Design and
performance”, Tappi J., 83(10): pp. 13 pages (2000).
298. Lockie, M., “Getting competitive through pilot plant trials”, Pulp Paper Int.,
(August): pp. 28–33 (1997).
299. Hynnek, M. and Shands, J., “Beloit Corporation’s new pilot paper machine”, in
Annual Meeting, Technical Section CPPA, Montreal, pp. A51–53 (1996).
300. Valmet, “Valmet’s new tool in printing paper R& D”, Paperi ja Puu, 78(4): pp.
139, (1996).
301. Crotogino, R. et al., “Paprican’s new pilot paper machine”, in CPPA Annual
Meeting, pp. A121–A127 (1999).
302. Gerli, A., Keised, B. and Strand, M., “The use of a flocculation sensor as a
predictive tool for paper machine retention program performance”, Tappi J.,
83(10): pp. 16 pages (2000).
303. Fredlund, M., “Unpublished summary of FEX investigations” (1996).
304. Holik, H., “Blattbildung – Grundlage für Papierqualität, Teil 1”, Wochenbl f.
Papierfabr., (10): pp. 375–380 (1992).
305. Holik, H., “Blattbildung – Grundlage für Papierqualität, Teil 2”, Wochenblatt f.
Papierfabr. (11/12): pp. 443–454 (1992).
306. Ohlsson, K., “Influence of raw material and forming process on quality of
woodcontaining printing papers”, (in Swedish). Dep Pulp and Paper Chemistry
and Technology, KTH, Stockholm, Licentiate Thesis (1993).

12th Fundamental Research Symposium, Oxford, September 2001 557

B. Norman and D. Söderberg

307. Albinsson, C.-J., Swerin, A. and Ödberg, L., “Formation and retention during
twin-wire blade forming of a fine paper stock”, Tappi J., 78(4): pp. 121–128
(1995).
308. Swerin, A. and Mähler, A., “Formation, retention and drainage of a fine paper
stock during twin-wire roll-blade forming. Implications of fibre network
strength”, Nord. Pulp Paper Res. J., 11(1): pp. 36–42 (1996).
309. Nilsson, B., “Fibre orientation effects in twin wire forming”, (in Swedish). Dep.
Paper Technology, KTH, Stockholm, Licentiate Thesis (1993).
310. Eguchi, A. et al., “Flow characteristics in a headbox and fiber orientation of
paper. Part 2: Verification by the experimental paper machine”, Japan Tappi,
47(7): pp. 887–889 (1993).
311. Nordström, B. and Norman, B., “Effects on formation, retention and mechan-
ical properties of forming blade distance in a twin-wire roll-blade former”,
Appita J., 48(1): pp. 19–24 (1995).
312. Nordström, B. and Norman, B., “Effects of headbox jet quality and blade pulse
force on formation, retention and mechanical properties during roll-blade form-
ing with curved blades”, Nord. Pulp Pap. Res. J., 10(1): pp. 33–40 (1995).
313. Nordström, B. and Norman, B., “Effect on paper properties and retention of the
proportion of roll dewatering during twin-wire roll-blade forming of TMP”,
J. Pulp Pap. Sci., 22(8): p. J283 (1996).
314. Odell, M., “Paper structure engineering”, in 53rd Appita Annual Conference,
Rotorua, pp. 155–160 (1999).
315. Erkkilä, A.-L., Pakarinen, P. and Odell, M., “The effect of forming mechanisms
on layered fiber structure in roll and blade gap forming”, in Tappi Papermakers
Conference, pp. 389–400 (1999).
316. Mohlin, U.-B., “Pulp evaluation for modern papermaking”, in 6th International
Conference on New Available Techniques, Stockholm, pp. 343–350 (1999).
317. Mohlin, U.-B., “Fibre dimensions-formation and strength”, in International
Paper Physics Seminar, Grenoble, pp. 5–14 (2000).
318. Odell, M. and Pakarinen, P., “The complete fibre orientation control and effects
on diverse paper properties”, in Tappi Papermakers Conference, Cincinatti, pp.
27 pages (2001).
319. Page, R. and Hergert, R., “Enhancement of Paper Properties and Fibre
Economics Through Web Stratification”, Appita J., 42(1): pp. 33–41 (1989).
320. Häggblom-Ahnger, U., “Three-ply office paper”, Laboratory of Paper Chemistry,
Åbo Akedemi University, Åbo, PhD Thesis (1998).
321. Häggblom-Ahnger, U. and Eklund, D., “The role of the middle ply in multilay-
ered copy paper”, Appita Journal, 51(3): pp. 199–204 (1998).
322. Häggblom-Ahnger, U., “Optimum location of softwood sulphate pulp in
three-ply office paper”, Tappi J., 82(6): pp. 181–187 (1999).
323. Häggblom-Ahnger, U., “Layering of office paper”, Paper and Tiber, 80(7): pp.
508–513 (1998).
324. Lloyd, M., “Stratified (multilayer) forming – a technology for the new millenium?
in 53rd Appita Annual Conference, Rotorua, pp. 161–168 (1999).

558 Session 3: Forming

 Transcription of Discussion

OVERVIEW OF FORMING LITERATURE,

1990–2000

Bo Norman and Daniel Söderberg

Royal Institute of Technology

Bill Sampson Department of Paper Science, UMIST

In your review you mention that there are some sort of grey areas, different
conditions can give different influence on formation and strength, and I won-
dered with your high contraction nozzle where you observe better formation
at zero jet to wire speed difference, whether there was any influence on
strength in that region?

Bo Norman
Well our basic finding is that if you do these improvements in the head box
such that you have an improved jet coming out then you will get better
strength. If you introduce some of these shear influences in dewatering you
will lose strength. So when stock is coming onto the wire section it is too late
to make any strength improvements; these have to be upstream in the head
box. We think that is one conclusion.

Jean-Claude Roux EFPG

You have presented what you believe to be the best forming twin wire
machines with adjustable blades and with fixed blades. What are your feel-
ings; can you extend this philosophy to hybrid forming machines?

Bo Norman
I think that the hybrid former is a retro-fit for older machines. I do not think
that you would fit this technology to new machines. For an optimum machine
you should not combine Fourdrinier dewatering forming and twin wire
dewatering, that will not be the best option.

12th Fundamental Research Symposium, Oxford, September 2001

Discussion

Jean-Claude Roux
Why do we do it like that?

Bo Norman
We want to dewater paper under minimum shear. If you start dewatering on a
Fourdrinier wire and you add a top wire the suspension will enter into a
pressure zone that will decelerate the suspension which will inevitably cause
shear. This should be avoided in the final, perfect paper machine but we have
a long way to go so with today’s machines, of course for rebuilds on a
Fourdrinier machine you will often have a positive formation effect when
introducing these top wires.

Jean-Claude Roux
You mean we lose orientation profile at the beginning of the hybrid former,
and it is not possible to achieve it later in the process?

Bo Norman
Yes, the best alternative is however to add stationary and adjustable blades in
the top wire unit, to be able to control and minimize shear effects.

Session 3