Anda di halaman 1dari 85

Chapter 6 Web forming

Bo Norman

Chapter 6

Web forming
In a broad sense, the term "forming" describes the dilution in the short circulation of the thick
stock flow to a mix flow, the approach flow system, and the CD (cross machine direction)
distribution and jet generation of the mix by the headbox, as well as the creation of a wet web by
dewatering of the mix in the wire section.
The term "mix" will be used to denote the thick stock, after dilution with white water to a fiber
concentration low enough to avoid excessive flocculation. The term "consistency" will be avoided,
since it alludes to the rather imprecise evaluation of concentration traditionally performed by
different kinds of "consistency meters" characterized by the fact that they never actually intend
to directly measure concentration.
The mix is the last link in the chain:
- Pulp original fiber raw material, delivered from pulp production
- Stock treated pulp including eventual additives
- Thick stock homogenized stock, delivered from the machine chest
- Mix thick stock diluted in the short circulation and delivered to headbox and wire section.
Thick stock dilution in the short circulation should be considered as one part of the forming
process. This is logical since, in high concentration forming, no dilution of the thick stock is
required, thus neither is a short circulation loop. In this book, however, the thick stock dilution in
the short circulation is treated in a preceding chapter.
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 chapter, 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.
During the forming process, the basic structure of the final paper product is created. Besides
formation, the distribution of material components such as fines and fillers in the thickness
direction is important.
The degree of fiber orientation anisotropy and fiber orientation misalignment is increasingly
important, particularly to quantify these parameters at different positions in a sheet, from bottom
to topside. Regarding terminology, fiber orientation to many seems to specifically denote the
misalignment angle. It is recommended that anisotropy or misalignment respectively be added to
avoid misunderstandings.
Additional treatment within the areas of formation, fiber orientation, and material distribution in
the thickness direction, in combination with paper properties aspects, is included in Volume 16:
Paper Physics of this book series.
For overviews of forming see, e.g., Parker1 and Norman2.

6.1 Basis weight variability


Paper variability on all scales is detrimental to paper properties. The different scales can roughly
be described as follows:
- Micro-scale denotes the variations in a scale smaller than 0.1 mm. On this scale particle
Papermaking Part 1, Stock Preparation and Wet End - Page 1

Chapter 6 Web forming

size, microflow, and colloidal interactions are important.


- Small-scale covers the range of 0.1 to 20 mm, in which range fiber flocculation and
hydrodynamic conditions during forming are the main causes for irregularity. Formation describes
the local basis weight variations in this range and will be discussed further in a later section.
- Medium-scale covers the range of 20 mm to 10 m and is affected mainly by instabilities in
headbox flow 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 MD (machine direction). Such
variations will not be treated in this chapter.
6.1.1 The power spectrum
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 electrical signals. The power spectral density P(v) describes how the
variance 2 (square of standard deviation) of a signal is distributed on different frequency ranges
dv.
R1
2 = P (v)dv (1)
0

The frequency power spectrum has since been applied to characterize the structure of
turbulence for more than 60 years3. The frequency power spectrum has also been used for a
long time 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 fiber 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, since geometrical size rather than frequency often simplifies the tracing of the origins
for variations such as wire lengths, felt lengths, and cylinder circumferences. The frequency
spectrum is therefore transformed into a wavelength spectrum. The wavelength l is calculated
from the frequency n, and the scanning speed (or flow velocity) u from using the transformation.
l=

u
n

(2)

The wavelength corresponding to a floc of size L amounts to 2L, since in the analysis one
(sine)wave consists of one "positive" and one "negative" floc. To conserve the basic property of
the power spectrum, i.e., to describe the distribution of variance, the frequency spectral density
P(n) has to be transformed into the wavelength spectral density E(l) 4.

E(l) = k nu2 P nu (3)

k is a constant depending on the bandwidth of the incremental wavelength intervals used to


present the spectrum within the wavelength range of interest. Usually logarithmic bandwidths are
preferable, and the width of an individual wavelength interval is then proportional to the mean
wavelength of that interval.
From a two-dimensional map, such as a beta radiograph for formation evaluation as
described below, the wavelength spectrum can be calculated directly, without using the
frequency spectrum as an intermediate step.

Papermaking Part 1, Stock Preparation and Wet End - Page 2

Chapter 6 Web forming

"A Power spectrum quantifies the product of number and density of flocs in diffrent size
ranges, averaged over the complete sampling area. Wavelet mothods also inform about the
location of different variations, such as streaky structures. Wavelets can also rely on local
variation shapes different from the sinusoidal (power spectra), such as e.g. rectangular."
6.1.2 Fiber flocculation
A fiber under free rotation in a dilute fiber suspension covers a spherical volume with a maximum
diameter equal to the fiber length. The highest concentration of freely mobile fibers of length L is
then represented by the case of closely packed spheres of diameter L. The corresponding
volume concentration cv of fibers will then depend not only on the fiber length L but also on the
fiber diameter d, or rather the slenderness L/d of the fibers. The more slender the fibers are, the
lower will be the volume concentration allowable to still give the fibers full freedom of individual
movement.
cv =

1;5

( Ld )2

(4)

For fibers with a slenderness of 100, cv according to Eq. 4, will amount to 0.015% or 0.15 g/L.
This could be compared with the concentration used in laboratory sheet forming, where, e.g., 7 L
of water is required to form a 1 g paper sheet, i.e., a concentration of 0.14 g/L. The conclusion
then is that to form a sheet of paper with a negligible amount of disturbing fiber flocculation, the
fiber suspension initially has to be so dilute that the fibers are free to move independently. In
reality, however, the fibers in suspension will still contact each other to some extent, since they
will not be evenly distributed in space. A more random distribution will then generate areas with
local concentrations above and below average concentration respectively.
6.1.2.1 Sediment concentration
A small amount of cellulose fibers is added to a beaker of water, circa 0.5 g fibers per liter water.
If the beaker is left undisturbed, the fibers will gradually settle at the bottom of the beaker under
the influence of gravity (density of cellulose 1.5 g/cm3).
When the weight of the added fibers is known, it is possible to calculate the weight
concentration of fibers in the sediment from the height of the sediment, the so-called "sediment
concentration." The concentration is expressed either as a percentage or in g/L, where 1%
corresponds to 10 g/L.
The sediment consists of one mechanically connected fiber floc. The sediment concentration
is actually the lowest concentration at which a connected floc can be formed by the fibers. Any
amount of stirring in the beaker is sufficient to cause the fibers in the sediment to whirl up,
however, which indicates that the floc strength in the sediment is very low.
The level of the sediment concentration depends on the slenderness L/d of the fibers. For
normal cellulose fibers, the slenderness lies within an interval of 50300. The higher the
slenderness is, the lower the sediment concentration will be.
Table 1 shows the approximate sediment concentration for some different types of pulp.
Table 1. Sediment concentration, g/L.
Softwood fibers
Hardwood fibers
TMP
Groundwood pulp

23
34
46
59
Table 2 provides typical values of the mix concentration used when forming paper on
industrial paper machines. To avoid excessive fiber flocculation, forming concentration should not

Papermaking Part 1, Stock Preparation and Wet End - Page 3

Chapter 6 Web forming

greatly exceed the sediment concentration.


Table 2. Typical fiber concentrations in industrial forming, g/L.
Sack paper
Fine paper
Linerboard
Newsprint

1.5
4
5
10

6.1.2.2 Network generation


Carefully add dry, unflocculated fibers (e.g., cut polymer fibers) to a beaker filled with water until it
reaches a concentration clearly greater than the sediment concentration. Stir slowly with a spoon,
and place it finally in the center. It will immediately fall against the beaker edge.
Add turbulence energy using a small propeller agitator. When the stirring has been stopped,
the spoon can be placed in the center of the beaker, without falling against the beaker edge. The
fiber suspension thus no longer behaves as a liquid (a liquid cannot take up shear stresses
without continuing deformation) but instead as an elastic body, a fiber network, with considerable
network strength.
The physics of this network strength is that, during the turbulent stirring process, the fibers
are deformed from their natural shapes, which they strive to regain when the stirring ceases.
When they then strike adjacent fibers, the springing back is hindered, and the fibers become
interlocked in strained shapes in a network. As shown in Fig. 1, at least three contact points with
the surrounding fibers are required for a fiber to be kept in a bent shape.
Figure 1. Mobile fibers (top) and a fixed fiber (bottom).
Meyer and Wahren5 analyzed how the volume concentration cn (really cvn) at an average
number of contact points n with surrounding fibers depends on the slenderness L/d of the fibers.
They derived an approximate relationship according to the equation:
cn =

16L

3
1
( 2L
nd + n1 ) (n1 )d

(5)

Figure 2. Jacquelin-flocs.
Soszynski and Kerekes6 reported two experiments that confirm the validity of the above
model for network generation. They slightly tilted a beaker with fiber suspension and made it
rotate. After a period of time, uniform nearly spherical fiber flocs formed, so-called
"Jacquelin-flocs" (see Fig. 2).
When flocs had been formed, the fiber suspension was diluted with water, under continued
rotation of the beaker. If the original fiber concentration was low enough, the fiber flocs were
broken up by this dilution.
When the original concentration was progressively increased, conditions were eventually
reached where the flocs resisted the subsequent dilution. They named this critical original
concentration the "threshold concentration," and Fig. 3 shows the way in which it varies with the
slenderness of the fibers.
In the diagram, a curve has also been drawn according to Eq. 5 with an average of three
contact points between a fiber and the surrounding fibers. It is shown by the figure that there is
fair agreement between this curve and the experimentally determined threshold concentration.
Figure 3. Threshold concentration as a function of fiber slenderness. The solid line corresponds

Papermaking Part 1, Stock Preparation and Wet End - Page 4

Chapter 6 Web forming

to Eq. 5 with n = 3. Soszynski and Kerekes 6.


Soszynski and Kerekes6 also demonstrated by another experiment that the mechanical
bending of the fibers is decisive for the strength of the flocs. They produced Jacquelin-flocs from
polymer fibers, and then removed some of the flocs from the rotating beaker. The removed flocs
were placed in water with a temperature above the softening temperature of the polymer. The
temperature treated flocs were then reintroduced into the rotating beaker, where they
immediately were broken up. This can be explained by stress relaxation in the fibers at the
temperature treatment, reducing the mechanically interconnecting forces within the flocs.
6.1.2.3 Crowding factor
Kerekes and Schell7 have introduced the concept of Crowding Factor N for the number of fibers
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 fiber of length L. The following relationship is then valid for the crowding factor
N.
2
(6)
N = 23 cv Ld
Experimentally, Kerekes has derived a relationship between the crowding factor and type of
fiber contacts at different concentrations (see Table 3).
Table 3. Fiber contacts at different crowding factor levels 7.
Crowding factor N
Concentration
Type of fiber contact
N<1
Dilute
Rare collisions
1 < N < 60
Semi-concentrated
Frequent collisions
N > 60
Concentrated
Continuous contact
It should be pointed out that the definition of the crowding factor assumes identical fibers.
Normally, a rather wide distribution of fiber lengths will be found in a practical situation. This
means that the crowding factor concept has to be treated with some caution in such cases, since
it is proportional to the square of fiber slenderness.
Since the density of cellulose is approximately 1.5, it would be expected that the volume
concentration of the fibers is lower than the weight concentration. However, because of the
swelling of the fiber wall and lumen volume, etc., concentration can be as high as twice the
weight concentration.
By eliminating the concentration between Eqs. 5 and 6, it is possible to express the number
of contact points n as a function of fiber slenderness L/d and the crowding factor N.
6.1.2.4 Network strength
The assumption that the strength of the fiber network is created by the mechanical interlocking of
the individual fibers predicts that the network strength increases with increasing fiber
concentration.
With measurements in a rheometer, Thaln and Wahren 8 demonstrated a relationship
between the shear failure strength B of the network and the fiber concentration c according to
the formula:
B = 0 (c cs )k (7)

where 0 and k are fiber dependent parameters and


cs is the sediment concentration.
The network strength is thus zero at the sediment concentration and then increases rapidly

Papermaking Part 1, Stock Preparation and Wet End - Page 5

Chapter 6 Web forming

with concentration. At high concentrations, the network strength is approximately proportional to


ck, where the exponent k lies close to two for most pulps.
At forming concentrations, colloidal phenomena can also influence the network strength. In
this way, the degree of flocculation in the paper product can be influenced by, e.g., the choice of
retention additives.
If a considerable amount of air is present in the system, this interacts with the fibers. Small air
bubbles get stuck in the fiber network and can influence the network forming, especially at low
concentrations.
6.1.2.5 Flocculation measurements
Qualitative studies of fiber flocculation in flowing fiber suspensions are possible using different
optical setups, but quantitative description of local concentration variations is extremely difficult,
due to secondary light-scattering effects. Most investigations have used transmitted light through
a transparent pipe or channel, but local reflection measurements have a potential of better
geometrical resolution.
Wgberg9 studied effects of chemical additives on flocculation in flowing fiber suspensions in
a pipe with 40 mm diameter. He used a laser light source, focused the illumination on one spot in
the pipe, and recorded the back-scattered light from that same spot. The optics were thus in
principle the same as those in confocal microscopy. From the signal, he evaluated the
flocculation spectrum up to 50 mm wavelength. He described the influence on flocculation of
chemical additives by taking the ratio between the recorded wavelength spectrum and a
reference spectrum without additives.
Beghello10 used a wide, flat channel with a reflecting background, provided light over the
whole area, and took two-dimensional pictures of the reflected light. It was not possible to
evaluate absolute concentration variations. Instead the 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. He studied the influence of, e.g., fiber concentration, fiber
surface properties, and chemical additives.
It should be pointed out that fiber flocculation in a mix is detrimental only to the extent that it
prevails also in the final paper product. Loose flocs, which break apart during the forming
process, do not have a negative effect. It is therefore necessary to complement the current
optical methods to evaluate fiber flocculation with methods to also evaluate the strength and
rheological properties of the flocs.
A general background within the area of fiber flocculation has been given by Kerekes 11,12.
6.1.3 Formation
The formation of a paper sheet, i.e., the local basis weight variations up to about 40 mm
wavelength, determined by the fiber distribution in the plane of the sheet, has a great influence
on many sheet properties; it is therefore desirable to be able to quantify this distribution. At least
two parameters, one overall intensity value which quantifies the total amount of local basis weight
variations, and one scale value, describing the distribution between small and large flocs, are
needed for a proper description of a specific paper sample. In analogy with the characterization
of macro scale basis weight variations recorded at the dry end, the coefficient of variation
(standard deviation divided by mean value) is a suitable measure of total variability.
In industry, formation evaluation is often done using light transmission methods, which are
sensitive to other optical parameters as well as local basis weight. Radiation sources other than
visible light are therefore required to quantify local basis weight variations.
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.
Papermaking Part 1, Stock Preparation and Wet End - Page 6

Chapter 6 Web forming

6.1.3.1 Optical methods


Traditionally, paper is studied against a backing light source and the variations in light
transmission are judged in order to assess the formation. This simplified method, however, can
give misleading results, especially in two cases:
- When the paper consists of components with different optical properties
- When the paper is calendered.
In the latter case, this is understandable if one considers that in calender blackening the more
transparent areas represent high local basis weight areas. They become transparent when
locally very high pressures are applied during calendering.
According to the Kubelka-Munk equations, which basically describe the optical properties of
paper sheets, light transmission is a function of local basis weight, local scattering coefficient,
and local absorption coefficient. It is therefore necessary that both scattering and absorption
coefficients are constant if light transmission is to be directly correlated to local basis weight only.
This is however not the case for the scattering coefficient in calendered sheets (see Komppa and
Ebeling13). All formation measurements based on light transmission should therefore take place
before calendering.
Further it is not possible to optically evaluate formation if a sheet consists of components with
different absorption and scattering coefficients such as fiber and filler.
In summary, optical formation meters have the drawback that they cannot give absolute
values of basis weight variations. However, they can still be valuable, for instance, to follow
changes on a specific machine. Comparisons between paper samples from different machines
are however more difficult.
Two of the more commonly used optical formation testers in North America are the NUI (Non
Uniformity Index) meter14 and the M/K Systems meter15. The NUI-meter analyzes light
transmitted through a rotating paper sample, and a total variation number is given as a formation
index. With the M/K Systems meter, on the other hand, the height of the amplitude distribution of
transmitted light is taken as a measure of formation, which means that a high value corresponds
to good formation. Recently the Kajaani formation tester has been introduced and, like the
M/K-systems meter, its output is calculated from the height of the transmitted light amplitude
distribution. Thus, also in this case, the formation index is higher for a more even paper sample.
6.1.3.2 Beta ray methods
To characterize formation in a correct way, a signal is required which represents the local basis
weight. Beta-radiation absorption is normally used to determine the paper basis weight on-line at
the dry end of the machine. For basis weight determination, beta-radiation is especially suitable
since it has the same absorption coefficient for different components in a paper and it also has a
negligible scattering coefficient. Unfortunately, it is not possible to measure formation on-line
since the very small measurement area would give inadequately low amounts of transmitted
radiation.
Different laboratory methods have however been developed utilizing beta-ray absorption to
measure formation.

Beta ray transmission


The Ambertec meter16 records beta ray transmission within local areas of 1 x 1 mm 2. The
measuring point is moved automatically using an x-y table, and 400 measuring points are
recorded on a paper sample size of 70 x 70 mm2. From the individual basis weight values, the
standard deviation is calculated with good accuracy, as well as normalized formation value
(see further below). The limited sample size often requires measurements on several samples to

Papermaking Part 1, Stock Preparation and Wet End - Page 7

Chapter 6 Web forming

give accurate formation values and also limits the possibilities to evaluate scale size.

Beta radiography
Beta radiography has been used since the early 1970s to evaluate formation 17. With this method,
an X-ray film in contact with the paper sample is exposed to the transmitted beta particles. In the
STFI method for beta radiography18, a C-14 beta radiation source with a size of 100 x 150 mm2
is used, which requires about 30 minutes of sample exposure, after which conventional film
development is performed. Figure 4 shows examples of beta radiographs.
Figure 4. Beta radiographs for two kraft paper samples, softwood kraft, 60 g/m2, twin-wire blade
forming. Left: Forming concentration 4 g/L. Right: Forming concentration 10 g/L.
Figure 5. Local basis weight variations along a newsprint sample using 0.1 mm measurement
resolution.
Simultaneously, calibration areas of constant basis weights within the range of interest are
exposed. After development, the film will show a negative, two-dimensional, gray-scale picture of
the basis weight distribution in the sample. The radiographs are analyzed in a standard desktop
scanner, using a resolution of 300 dpi (dots per inch), which corresponds to circa 0.1 mm
resolution. With the help of the calibration areas, the gray-scale values can be transformed to
basis weight valves. Thus a two-dimensional basis weight map can be generated, and formation
characteristics can be calculated (see below).
Figure 5 shows variations in the local basis weight along a sample of newsprint, recorded on
a beta radiograph. Local basis weight is seen to vary between about 25 and 60 g/m2.
6.1.3.3 Formation characteristics
The term "formation index" is reserved for optical formation meters, which give less well defined
measures of absolute formation level.
Formation number F denotes the coefficient of variation of local basis weight, that is, standard
deviation (w) divided by mean basis weight wm, and it is often expressed as a percentage.
F = (w)=wm (8)
If a paper sheet consists of a number of layers of equal structure, the formation number will
be lower, the higher the number of layers are in the sheet. It is therefore not relevant to compare
formation numbers between sheets of different basis weights. Comparisons are however possible
using normalized formation numbers Fnorm according to Eq. 9,
p
Fnorm = F wm =wn (9)
where wn is normalization basis weight and

wm sample basis weight.


Normalization according to Eq. 9 assumes that all layers in paper samples of different basis
weights have similar fiber distributions. 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 (see further discussion below).
In the STFI method, 60 g/m2 is normally used for normalization basis weight. It could be
p
pointed out that the normalized Ambertec meter values in g=m (which is a dimension hard to
interpret) can in fact be interpreted as a dimensionless normalized formation number according to

Papermaking Part 1, Stock Preparation and Wet End - Page 8

Chapter 6 Web forming

Eq. 9, using a normalization basis weight of 1 g/m2.


A more complete characterization of formation is given by the wavelength power spectrum.
The area beneath a power spectrum, as already mentioned above, is equal to the variance of the
parameter studied, and the area beneath a formation spectrum is thus equal to the square of the
formation number F2.
Because of the logarithmic wavelength scale, spectra presented as variance per unit
wavelength have the disadvantage that visually the importance of large flocs is underestimated.
To give a more correct picture of the contribution to the real area beneath the spectrum, in spite
of the logarithmic scale of the x-axis, a modified presentation is now used in the STFI method.
The variance is then presented for wavelength intervals proportional to the wavelength, instead of
intervals of constant wavelength.
This means in practice that the original spectral density is multiplied by the wavelength (and a
constant which is dependent on the bandwidth chosen) so that the spectral density at small
wavelengths decreases while, at large wavelengths, it increases.
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.
Figure 6. Modified STFI formation spectra for the two samples in Fig. 4. Solid line: Concentration
4 g/L, F=13.3%, F(0.33)=9.3%, F(330)=9.5%. Broken line: Concentration 10 g/L, F=19.1%,
F(0.33)=9.8%, F(330)=16.4%.
Figure 6 shows examples of modified STFI formation spectra. The formation is given for two
different forming concentrations. The figure shows clearly the dominating occurrence of large
flocs in the paper with the higher forming concentration. The spectra can be compared with the
radiographs in Fig. 4, which represent these two samples.
6.1.3.4 Formation of random sheets
It was suggested by Wrist during the 1960s19 that fiber distribution in a real paper could be more
even than that of a random sheet. 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 fiber suspension will be dewatered
at that position. Thus extra fibers will be deposited, and the overall basis weight evened out. This
was later studied by Norman et al.20 by calculation of the formation spectrum of a random sheet
and comparison using beta radiography with that of a well-formed laboratory sheet. 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 fiber flocculation
effects, by definition not present in the random sheet. This result has later been confirmed in an
investigation21 showing that the real sheet is more even than the random sheet at wavelengths
smaller than about 10 mm (see Fig. 7).
Figure 7. Formation spectra at 60 g/m2 for a laboratory sheet of softwood fibers, F=10.2%, and
the corresponding random sheet, F=14.1%21.
6.1.3.5 Dewatering effects on formation
As mentioned in the previous section, real paper sheets are more even in the small scale than
random sheets. This is due to the self-healing effects of the dewatering process. To avoid the
self-healing effect during paper forming, the dewatering rate must be extremely low, which leaves
time for considerable fiber flocculation by fiber-fiber interaction and sedimentation. In a

Papermaking Part 1, Stock Preparation and Wet End - Page 9

Chapter 6 Web forming

conventional laboratory sheet former, sugar was therefore added to the dilution water. The
amount of addition was such that the fluid density was equal to that of the fibers, and the
viscosity was highly increased. By extremely slow dewatering, thus avoiding the self-healing
effects, it was then possible to manufacture a real sheet with random fiber distribution 22. The
strength of such a sheet was lower than that of a standard laboratory sheet, which demonstrates
that the self-healing effect has some positive effect on paper strength.
Due to the increasing self-healing effect, the formation number compensated for the basis
weight level (see Eq. 9) should improve with increasing mean basis weight. This was
demonstrated for laboratory sheets formed at standard concentration 21. It can then be concluded
that a fourdrinier paper in principle has a better formation potential than a twin-wire formed paper,
and maybe also a better strength potential. The reason for this would be that the twin-wire paper
in reality consists of two plies, each with half the total basis weight. The self-healing effect is then
lower for each half than for the single sheet in fourdrinier forming.
Figure 8. Strength of single- and double-layered laboratory sheets respectively 21. Laboratory
sheets with 50/50 mixture of short/long fibers.
The strength of single- and double-layered laboratory sheets formed at standard
concentration was evaluated (see Fig. 8). From the diagram, it is clear that the strength potential
for a fourdrinier sheet is higher than that of a twin-wire sheet. In reality, however, many other
aspects have to be considered. The trend today is to use twin-wire forming in new paper
machines also for the production of strong papers. Main advantages are better control of the
dewatering process including medium-scale basis weight variability, especially at high machine
speeds, as well as lower equipment costs and space requirements.
6.1.4 Medium-scale basis weight variations
Maps of the two-dimensional basis weight distribution in the scale above the formation range,
that is, around 20 mm, up to about 10 m has been recorded by STORA using beta ray absorption
techniques for measurements in a laboratory scale. Figure 9 shows an example of such a basis
weight map.
Figure 9. Gray scale picture (original in color) of basis weight of full width paper web. Newsprint
quality, with each level curve indicating a deviation of 1 g/m2(23).
Such a map gives a good picture of the medium-scale variations in basis weight and is an
important tool for tracing the origin of different basis weight defects.
Recently, corresponding two-dimensional maps of light transmission have been recorded
on-line with the ABB Hyperscan system24. Even if only light transmission variations 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.

6.2 Headboxes
The main function of the headbox is to distribute the mix evenly across the width of the wire
section. This means, for example, that the flow from a pipe with a diameter of 800 mm shall be
transformed into a 10 mm thick and 10 000 mm wide jet, with absolutely the same flow rate and
flow direction at all points across the width, as indicated in Fig. 10.
Figure 10. Feed pipe for mix and cross-section of jet from headbox (not to scale).
Table 4. Typical jet thickness [mm] in industrial forming.
Papermaking Part 1, Stock Preparation and Wet End - Page 10

Chapter 6 Web forming

Newsprint
8
Fine paper
15
Linerboard
25
Sack paper
50
A simple equation that relates the required lip opening h [mm] to form a product of basis
weight w [g/m2] from a mix concentration of c [g/L] is given below:
w = Rhc (10)
where R is the retention factor. This equation presumes that jet speed equals wire speed.
The flow transformation by the headbox from the incoming pipe flow to the delivered plane jet
takes place in mainly three steps:
- The cross-direction distributor makes a first distribution of the mix across the machine
width.
- Pressure drop elements are introduced to even out the CD flow profile.
- A headbox nozzle generates the final jet.
6.2.1 Cross-direction distribution
A modern cross-direction distributor usually consists of a tapered header, a channel that runs
across the whole headbox width, from which discharge takes place successively through holes in
the channel wall25 (see Fig. 11).
Figure 11. CD distributor with main flow Q, discharges Q, and overflow q.
If a perfect headbox jet is to be delivered to the wire section, the mix flow Q through each
discharge hole must be equal. This means that the static pressure along the CD distribution
channel must be kept constant. The friction pressure drop along the channel must therefore be
compensated for by a corresponding pressure rise.
This pressure rise can be achieved by transforming part of the velocity energy in the flow
along the channel into static pressure by successively reducing the flow velocity. The channel
then functions in principle like a diffuser, even though the cross-sectional area A(x) decreases in
the flow direction, in contrast to the case in a conventional diffuser. This is possible since the
volumetric flow along the channel gradually decreases because of the discharge flows Q. The
following equation describes the flow velocity u(x) along the CD channel.
u(x) =

Qn(x)Q
A(x)

(11)

where n(x) is the number of flows Q discharged before position x. The cross-sectional area A(x)
along the channel changes so that, for given values of input flow Q and overflow q, a constant
static pressure is obtained along the whole channel. It will not be possible, however, to maintain
this constant pressure with other flow rates Q. If there is a pressure difference between the inlet
and outlet, the overflow rate q can be adjusted so that pressure agreement is attained. Some
pressure deviations inside the channel can nevertheless still remain.
The distance between the individual discharge holes should be so large that bridging of fibers
across two adjacent holes is avoided. Such fiber piling would lead to the build-up of detrimental
fiber flocs.
The discharge holes feed a tube bank or drilled plate. The higher the pressure drop is along
the tubes, the smaller the differences will be between the individual discharge flows Q caused
by variations in static pressure along the CD distribution channel. To achieve a high pressure
drop, the local flow velocity must be high, which means that the flow area in the tube bank should
Papermaking Part 1, Stock Preparation and Wet End - Page 11

Chapter 6 Web forming

be relatively small. This means that the (relative) open area at the inlet of the tube bank, i.e., the
ratio of the total area of the discharge holes to the channel wall area is small (in the order of 10
percent). To feed a downstream chamber from only 10 percent open area would however
generate unacceptable flow instabilities. To improve the stability of the downstream flow, the
relative open area at the outlet from the tube bank therefore may be increased considerably.
This can be done in two alternative ways:
- The tubes or holes are expanded with gradual or sudden increases in diameter, so that the
flow cross section is successively increased.
- The tube centers are brought together, so that the solid area between the individual jets
delivered is decreased.
After the tube bank, there may be an equalization chamber, followed by a secondary tube
bank to generate a pressure drop with a function to further improve the velocity CD profile.
Depending on the design of such a chamber, a headbox is called either an "air-cushion headbox"
or a "hydraulic headbox." The latter term is also used if the tube bank directly feeds the headbox
nozzle.
6.2.2 Air-cushion headboxes
The air-cushion headboxes are a development of the original, completely open headboxes,
where gravity was the only driving force for the outflow through the headbox nozzle.
At high machine speeds, however, too high a mix height would be required in an open
headbox to achieve the required jet speed. A jet speed of 250 m/min requires a mix height of 0.9
m. At higher speeds, it was therefore preferable to close the headbox and place a pressurized air
cushion over the mix in the equalization chamber, to create the driving forces required for the jet
velocity generation. The compressible air cushion has a damping effect on pressure fluctuations
in the mix flow entering the headbox.
A traditional way of evening out velocity profiles, which is applied in, e.g., wind tunnels, is to
let the flow pass through a number of nets which cause pressure drops and thus generate lateral
flow from areas with locally high flow velocities. Unfortunately, nets cannot be applied in a paper
machine headbox since fibers would rapidly clog them. Instead, rolls perforated with circular
holes are placed in the equalization chamber. Their function is to generate a pressure drop and
thereby even out the flow velocity profile across the headbox width. They are kept in slow rotation
to reduce the tendency for large fiber flocs to form at the hole edges. Early air-cushion
headboxes could be equipped with up to five perforated rolls, while two are used in modern
versions (see Fig. 12).
Figure 12. Air-cushion headbox with two perforated rolls.
The degree of downstream velocity disturbances originating from a perforated roll is
determined mainly by the size and spacing of the holes. The relative open area of a perforated
roll should not be below 50 percent. This still generates a rather turbulent flow at the entrance to
the headbox nozzle.
Air-cushion headboxes are now used mainly for moderate machine speeds, for the
manufacture of different specialty papers, and for some kraft paper machines, which require very
large jet thickness.
6.2.3 Hydraulic headboxes
Hydraulic headboxes were designed specifically for twin-wire forming. A main requirement was
small nozzle dimensions to allow a short free jet from the headbox into the gap between the two
wires. Hydraulic headboxes lack the traditional air cushion and are available either with or without

Papermaking Part 1, Stock Preparation and Wet End - Page 12

Chapter 6 Web forming

an equalization chamber. If disturbing pressure pulsations occur in the approach mix flow, it is
necessary to install a separate air-cushion pulsation damper before the headbox.
In a hydraulic headbox with an equalization chamber, the perforated rolls in the air-cushion
headbox are replaced by a secondary tube bank between the chamber and the headbox nozzle,
to improve the velocity CD profile. Although it is impossible in air-cushion headboxes, a stationary
arrangement is applicable in hydraulic headboxes, since the turbulence level is higher (due to
higher flow velocities) so that fiber accumulation at edges on the upstream side of the pipes is
avoided. The secondary tube bank is often called "turbulence generator," although its main
objective is to reduce local flow defects.
As mentioned earlier, the relative open area at the downstream end of the pipe package
should be as large as possible to stabilize the nozzle flow. In the first hydraulic headboxes, the
headbox nozzle was fed by rather widely spaced tubes, which gave relatively high turbulence
levels. To counteract the generation of large-scale fluctuations, Beloit divided the nozzle into
several narrow channels with the help of thin, flexible separation vanes (see Fig. 13), a technique
which was introduced at the end of the 1960s26.
Figure 13. Beloit Converflo headbox nozzle with separation vanes.
Voith developed a design for improved flow stability during the 1970s, by modification of the
outlet section of the pipe package feeding the headbox nozzle 27. At the inlet side, the pipe
package consisted of widely spaced cylindrical pipes (low open area). Toward the outlet of the
pipes, the shape was gradually modified to a hexagonal form. These hexagons could then be
packed closely together, so that a high open area was attained.
Figure 14. Flow pattern in Escher Wyss step diffuser headbox.
An alternative design by Escher Wyss, also from the 1970s, is the so-called "step-diffuser
headbox"28, in which there is no equalization chamber (see Fig. 14).
In the tube bank from the CD distributor, area increases take place through sudden
expansions of the pipe diameters. The final stage has a square cross section in order to permit
close packing. Along a pipe wall, a boundary layer of water develops, and vertical walls can then
generate streaks in the paper if several pipes are located on top of each other. However, a
sudden increase in flow cross section area means that the boundary layer is broken up. This
reduces the problems of basis weight streaks when several units are stacked vertically.
A similar principle but with a rectangular discharge shape is now used by Valmet 29. Valmet
has chosen to use some sideways displacement between the different rows to avoid basis weight
streaks (see previous paragraph), although this in turn causes some problems at the headbox
edges. To avoid jet flow misalignment, corrective mix additions sometimes are made at the
headbox edges.
6.2.4 Headbox nozzle
In the headbox nozzle, consisting of top and bottom lips, the mix is accelerated and leaves the
nozzle outlet in the form of a plane jet. The lip opening determines the initial thickness of the jet.
The ideal jet is perfectly flat, which requires that it contains no turbulence. When such a jet lands
in the wire section, the wire contact will take place along a straight line, perpendicular to the MD.
In reality, however, the jet surfaces are never planar, but deteriorate gradually with the distance
from the nozzle end. The irregular jet contact with a wire will then cause flow disturbances and
can influence web formation and local fiber orientation.
To generate a jet with the smallest possible velocity variations, feeding the nozzle from a
large relative open area is important, as discussed above. The acceleration of the suspension in
Papermaking Part 1, Stock Preparation and Wet End - Page 13

Chapter 6 Web forming

the nozzle generates a pressure drop, and this will reduce local velocity defects and reduce the
degree of relative turbulence. Thus a high degree of acceleration using a large nozzle contraction
ratio (inlet divided by outlet cross sectional areas) will be positive.
Figure 15 shows results from studies of one channel in a Beloit headbox (see Fig. 13), fed
with relatively widely spaced pipes. At the largest slice opening, a velocity streak from each feed
pipe is evident. When the slice opening is reduced by a factor of three, i.e., the nozzle contraction
ratio is increased by a factor of three, the velocity peaks nearly disappear.
Figure 15. Velocity profiles at nozzle outlet for different slice openings d030.
During the 1990s, Beloit replaced the widely spaced feed holes by step diffusers with final
gentle transition from circular to rectangular sections. These are tightly packed sideways to give a
high relative open area at the nozzle inlet.
The mix height in the main chamber of an air-cushion headbox is normally 5001 000 mm,
and the mix is fed directly from this chamber to the headbox nozzle. This means that the
contraction ratio of the headbox nozzle is large. This large contraction is needed to counteract
the high turbulence level downstream of a perforated roll.
A large stock height at the entrance to the headbox nozzle permits the use of a large jet
thickness with acceptable jet quality. On, e.g., sack paper machines, a jet thickness of 50 mm
may be applied.
Hydraulic headboxes, on the contrary, were originally designed for much smaller jet
thickness. It then seldom exceeded 1520 mm for printing paper forming in twin-wire machines.
Recently, however, hydraulic headboxes have been developed also for sack paper production,
and the number of pipe layers in the package feeding the headbox nozzle has then been
increased to give an acceptable jet quality.
The lower mix heights in hydraulic headboxes than in air-cushion headboxes often results in
a lower degree of large-scale turbulence in the jet, and this in turn reduces large-scale basis
weight variations in the paper product.
Turbulence of all scales is harmful to the planeness of the free jet and then also to paper
formation. Turbulence in the mix during dewatering in the wire section is also likely to be
detrimental to paper formation. Turbulence should therefore be avoided at these stages of
forming, while at earlier stages it might be positive for deflocculation purposes and for decreasing
fiber orientation anisotropy.
The level of relative turbulence generated by the mixing between the individual jets entering
the headbox nozzle will be reduced by the acceleration of the nozzle flow. During the
acceleration phase, however, some increase in the absolute turbulence level will take place in all
three directions. This is partly a result of "vortex stretching." The thickness of the boundary layers
inside the headbox will decrease along the nozzle, due to the strong negative pressure gradient
in the flow direction. Recent wind tunnel experiments indicate that, depending on the magnitude
of the nozzle contraction ratio, a boundary layer transition from turbulent to laminar state can take
place before the nozzle outlet31.
Figure 16. Polar fiber orientation distribution. Left: Isotropic laboratory sheet. Right: Anisotropic,
machine made sheet.
Besides producing a stable jet, a high degree of acceleration leads to increased fiber
orientation anisotropy32,33, which is undesirable for many paper grades. Fiber orientation
anisotropy is also discussed by Niskanen in Volume 16: Paper Physics, and will be further
discussed in a later section.

Papermaking Part 1, Stock Preparation and Wet End - Page 14

Chapter 6 Web forming

Fiber orientation distribution can be described in a polar diagram and often shows an elliptic
shape (see Fig. 16).
The orientation anisotropy ratio can be expressed as a/b, and is the misalignment angle.
A method to analyze fiber orienta-tion at different levels in the z-direction of a paper sample
has been developed34. The paper samples are delaminated using adhesive tape, and fiber
orientation anisotropy in each layer is evaluated using image analysis techniques.
Figure 17. Fiber orientation anisotropy at different levels in a paper sample, using low (8.5:1,
squares) and high (30:1, triangles) headbox nozzle contraction ratio respectively. FEX twin-wire
roll forming experiments, 60 g/m2, hardwood/softwood-mixture, 700 m/min. Minimum velocity
difference between mix and wires 35.
Figure 17 demonstrates the anisotropy effects using a low and a high nozzle contraction ratio
respectively.
Since a velocity difference between mix and wires increases orientation anisotropy (see
further below), this difference was kept to a minimum in Fig. 17. The results then indicate to what
degree fiber orientation anisotropy existed already in the headbox jet. There can be two reasons
for the lower orientation anisotropy toward the paper surfaces using the high nozzle contraction
ratio:
- The boundary layers along the top and bottom lips in the headbox nozzle might have
prevented the MD orientation of the fibers.
- The disturbances at the interaction between jet surfaces and wires might have rearranged
an original MD fiber orientation in the jet.
The accelerating flow in the headbox nozzle also has a deflocculating effect. The flocs are
stretched in the flow direction and can be broken down into smaller units, especially for
mechanical pulps. The influence of headbox nozzle contraction on fiber orientation anisotropy
and on web formation is further demonstrated in a later section.
It should be remembered that there is also some flow contraction in the initial part of the free
jet. The total contraction ratio is the product of nozzle and jet contraction ratios.
Sderberg showed theoretically and experimentally that deviations from a flat velocity profile
will generate surface instabilities in a plane jet36. This occurs when low velocity boundary layers
are accelerated in the emerging free jet. The jet surfaces will become more uneven further away
from the lip opening, and it is therefore vital to keep the free jet length to a minimum.
Headbox instabilities generated by the side walls and the tube bank have been analyzed by
Aidun37.
6.2.5 Jet angle and velocity
As mentioned previously, the mix jet flow requirements are that the speed and flow direction are
constant across the whole width of the machine and that also the jet has a constant thickness in
the CD. Figure 18 shows the discharge from a headbox lip opening and the free jet. The top lip
often ends with a slice. This has two functions: firstly, to facilitate a local change in the lip opening
(see later in this section) and, secondly, to reduce the thickness of the boundary layers formed
along the lips through local mix acceleration.
Figure 18. Outflow from a headbox with a sloping top lip and a vertical slice.
Because of the jet contraction, the thickness h at "vena contracta" is always less than the
geometrical slice opening ho. The ratio is called the contraction coefficient of the jet, where

Papermaking Part 1, Stock Preparation and Wet End - Page 15

Chapter 6 Web forming

h
h0

(12)

The contraction coefficient and jet angle depend on the angle of inclination of the top lip,
the relative downward protrusion Y/h o of the slice and the relative horizontal protrusion of the
bottom lip, X/ho. A vertical movement of the slice ho can adjust the lip opening.
The angle and thereby the location of jet landing in the wire section can be controlled by
changing the protrusion X of the lower lip. Equations for calculating jet contraction and jet angle
have been given by Kerekes38.
The jet speed u at vena contracta can in the general case be calculated with the aid of the
energy equation (Bernoulli equation):
p + 2 u2 + gHL + U = const (13)
where p is pressure,
density,
g acceleration by gravity,
HL height level and
U internal energy.
The friction pressure drop from the pressure measurement position, marked at the lower left
in the headbox in Fig. 19, to the discharge is pf. Part of it is transformed into heat through
viscous dissipation and thereby increases the internal energy U. Usually this part can be
neglected and is therefore not considered in the discussion below.
Figure 19. Outflow from a hydraulic headbox.
The pressure drop pf and the pressure p1 in the headbox can be expressed as "pressure
heights" by the transformations
pf = gHf (14)
p1 = gH1 (15)
The equation for conservation of the volumetric flow Q further gives
Q = h1 u1 = h0 u (16)

u=

If Eqs. 1316 are combined, the following relationship is obtained:


s
2gH1

Hf
HL
H1 H1
2
h
1 h 0
1

(17)

This equation is thus valid for a hydraulic headbox, in which the friction pressure drop in the
headbox, the dynamic pressure at the pressure measurement points, and the height differences
cannot be neglected.
In air-cushion headboxes, the mix speed inside the headbox is often so low that both the
friction pressure drop and the dynamic pressure can be neglected. An air-cushion headbox also
normally works horizontally. Equation 17 then reduces to the well-known equation:
p
u = 2gH1 (18)
Here H1 is the sum of the mix height in the headbox chamber and the pressure height in the
air chamber relative to the surrounding atmosphere.
6.2.6 CD basis weight profile
Papermaking Part 1, Stock Preparation and Wet End - Page 16

Chapter 6 Web forming

An uneven feed from the CD distributor will generate some transverse flow components in the
headbox nozzle. Such sideways flow components cause quality problems, since the fiber
orientation distribution in the final paper product will no longer be symmetrical around the
machine direction, and fiber orientation misalignment will result. If the headbox is equipped with
an intermediate chamber, this allows some equalization through lateral flow; consequently, there
will be less transverse flow components in nozzle and jet.
The lip opening profile, through local deformations of the slice can control the final CD
distribution of the mix in the jet. The local position of the slice is controlled by "slice screws,"
which are placed across the whole width of the headbox at intervals of 100150 mm. On modern
machines, the slice is equipped with position indicators so that the slice profile can be easily
evaluated. Remote control of the slice screws makes automatic control of the lip opening profile
possible, based on the basis weight profile recorded at the dry end of the machine.
A local reduction in lip opening reduces the mix outflow at this position but, at the same time,
transverse velocity components are generated on each side (Fig. 20).
Figure 20. Transverse flow components at local lip opening reduction.
The magnitude of the lateral flow components depends to a high degree on the design of the
headbox nozzle. With a low headbox nozzle contraction angle, typical for a modern hydraulic
headbox, over four times as much sideways flow can be generated as with a traditional air
cushion headbox with vertical front wall39.
Further, in a conventional drying section, CD shrinkage will show a minimum at the center
and increase toward the front and back sides of the machine. This means that, with an even CD
basis weight profile after the wire section, the final profile will follow the trend of the shrinkage
profile. Thus the CD profile from the headbox should compensate for this, which means that the
lip opening should decrease toward the front and back sides. And, as discussed above, such
local changes in lip opening generate fiber orientation misalignment. Consequently, misalignment
is then impossible to avoid if basis weight profiles are controlled using the conventional method of
lip opening adjustment.
6.2.6.1 Dilution control
A new way of controlling the basis weight profile is through local dilution, where white water is
added selectively across the headbox to compensate for streaks with local basis weight
deviations. The principle of CD concentration control has been introduced to mainly avoid fiber
misalignment problems. In principle, it allows the use of a constant lip opening all across the
machine; thus, the generation of sideways flow in the headbox nozzle could be avoided. Schultz
made an early design with a radial, central CD distributor feeding the headbox via a number of
hoses. In each hose, dilution water could be added 40, see Fig. 21.
Figure 21. Radial flow design for CD distribution.
One principal advantage with the radial distributor is that, unlike an ordinary manifold
channel, it will deliver equal flow rates in each hose, independently of the total level of mix flow
rate.
Figure 22. Voith principle for addition of dilution water.
The dilution principle has been shown to be effective in industrial applications and is now a
standard for new headboxes. According to the Voith design 41, dilution water is added to each
pipe in a single row package exiting from the CD distribution channel. To keep the total flow

Papermaking Part 1, Stock Preparation and Wet End - Page 17

Chapter 6 Web forming

through each pipe unchanged, the additions are made at a certain angle to the pipe axis. There is
also a downstream cross section restriction where the main pressure drop is generated.
In the Beloit design42, water addition already takes place in the CD distribution channel.
Upstream of each pipe feeding the headbox, dilution water addition is made. In a Valmet
headbox43, dilution water is added in each pipe exiting from the CD distribution channel, and the
addition takes place at a step increase in pipe diameter.
A further advantage of the dilution control method compared with lip opening deformation is
that more narrow streaks can be affected, using a CD distance between dilution points even
smaller than 50 mm.
Separate valves are used to control the individual additions of dilution water. Small additions
would require small valve openings and, when unfiltered white water is used, this could lead to
valve blocking with fiber material. Therefore specially designed valves are used, and a minimum
of 5%10% of the mix flow is usually added for dilution.
6.2.7 Multilayer headboxes
In principle, a multilayer headbox includes a separate CD distribution channel for each layer and
flow separation between the different layers throughout the headbox. In the headbox nozzle,
separation vanes are applied, for which bending stiffness as well as thickness along the vanes is
important. Especially the geometrical design of the downstream ends and the vane length in
relation to nozzle length are critical parameters.
Figure 23. Flow instability with a layered headbox nozzle 44.
The turbulence level in the wake behind a vane should be minimized. This requires a tapered
downstream vane shape with a very thin tip. The fine-scale turbulence generated will then
dissipate comparatively quickly in the mix. A blunt downstream vane edge will generate periodic
fluctuations in the flow, with a larger mixing action between the different layers.
There are two conflicting requirements regarding the length of the separation vanes 44. With
extended vanes, the different layers can be kept completely separated even some distance
outside of the headbox nozzle. However, a considerable amount of turbulence will be generated
along the extended vanes; therefore, small-scale downstream mixing between the layers will be
considerable.
If the vanes end already inside of the converging part of the headbox nozzle, the boundary
layers will be thin due to the negative pressure gradient during the flow acceleration in the nozzle.
Furthermore, the downstream wake will also be stabilized by the negative pressure gradient
before the nozzle outlet. However, fiber flocs of considerable size can still be present in the
different channels and, for geometrical reasons, these can later protrude into adjacent layers.
Further developmental work is needed before a high-quality, layered jet can be generated.
Application of the new technology for CD profiling using dilution control is necessary to control
the basis weight profile of the individual layers in multilayer headboxes.
6.2.8 High-concentration headboxes
Conventional forming generally takes place at concentrations of less than one percent. Forming
at higher concentrations would mean smaller volumetric flows thus smaller transport pipe
diameters, pump sizes, etc. High-concentration forming therefore has a potential for both reduced
investment costs and lower energy consumption.
Figure 24. Headbox for high-concentration forming (STFI).
A high-concentration headbox working at a concentration of about 3% was developed at STFI

Papermaking Part 1, Stock Preparation and Wet End - Page 18

Chapter 6 Web forming

during the early 1970s45. The basic idea is to fluidize the fiber suspension, and then let the
turbulence decay during flow in a 23 mm narrow channel, thus geometrically hindering the
forming of larger fiber flocs. The decay event takes place in a winding channel, in which
alternating shear is introduced, to some extent bringing the individual fiber flocs together.
The headbox "nozzle" in this case consists of a converging channel, widening to the lip
opening required for the targeted basis weight. As an example, according to Eq. 10, a 150-g/m2
product formed at a mix concentration of 30 g/L (3%) would require a lip opening of 5 mm (at 100
percent retention level).
The forming process at high concentration is rather an extrusion procedure. The jet leaving
the headbox consists of a mechanically connected fiber network. Much of the final sheet structure
has then already been formed inside the headbox. The process therefore differs considerably
from a conventional forming process where the sheet is built up during drainage in the wire
section (see the next section). Further development of the process has been carried out at
Valmet 46.

6.3 Wire sections


The water is drained from the mix in the wire section, and a wet web is formed. In principle, the
dewatering process can be of two different kinds, filtration or thickening (see Fig. 25).
The filtration type of dewatering dominates conventional forming (in contrast to high
concentration forming), i.e., fibers are successively deposited flat on a wet web as the
suspending water is removed, and a mix with the same concentration as that in the jet from the
headbox is still present above the wet web. The strongly layered web structure characteristic of
conventional paper is formed in this way.
Figure 25. Dewatering through filtration (left) and thickening (right)1.
When all free mix has been dewatered, the remaining dewatering takes place as a thickening
process, which means a progressive compression of a fiber network. The dewatering of fiber
flocs remaining in the mix also takes place according to the thickening principle.
In fourdrinier forming, downward dewatering of the mix takes place through a horizontal wire,
and this was the principle of the original paper machine. During the 1970s, twin-wire forming was
increasingly used for printing paper production. The mix is enclosed between two wires and
two-sided dewatering is then possible, which can generate products with less asymmetry, i.e.,
less two-sidedness. Twin-wire dewatering is now also increasingly used for the production of
packaging paper and board.
In high concentration forming (as mentioned in previous section), the fiber suspension has so
high a concentration, ca. 3%, that a fiber network with considerable strength is already formed
inside the headbox. The following dewatering then is a pure thickening process and the resulting
web structure is considerably more "three-dimensional" than that of the layered, conventional
webs.
6.3.1 Fiber deposition
It is of vital importance that the fiber distribution in the web is as uniform as possible so that the
best formation is attained. Traditionally, there are three main ways of improving the fiber
distribution, viz. dewatering, oriented shear, and turbulence 1 (see Fig. 26), and to this we have
now also added elongational flow33.
(a) Turbulence can be positive at an early stage of forming, i.e. before the jet exit from the
headbox nozzle. Then it can contribute to the homogenization of the mix, by breaking down extra
hard and extensive flocs. However, even though flocs can be disrupted during turbulent

Papermaking Part 1, Stock Preparation and Wet End - Page 19

Chapter 6 Web forming

conditions, flocs will eventually reform during turbulence decay.


Any turbulence in the mix jet leaving the headbox nozzle will contribute to jet surface
roughness, which will interact with the forming screen surface and cause disturbances in paper
formation. Similarly, any turbulence present during dewatering has an adverse effect on
formation. Such a negative effect can easily be demonstrated by continued stirring during
handsheet forming.
(b) It has been shown that fiber flocs can be broken apart through a stretching effect in an
elongational flow47. Elongational flow is developed mainly by the acceleration in the headbox
nozzle, and an improvement in paper formation can be obtained using a high nozzle contraction
ratio33. This is further discussed in a later section. Possibly, formation improving extensional flow
can also be generated by the pressure pulses in twin-wire blade forming 33.
(c) Dewatering has a self-healing effect on local variations in basis weight, as described in a
previous section.
(d) At normal forming concentrations, fiber flocs occur to some extent. The weakest flocs can
be broken apart during the dewatering process through oriented shear. The shear field can also
move flocs to new, less disturbing positions.
Oriented shear also influences fiber orientation anisotropy. When one end of a fiber during
dewatering is deposited on the wet web surface, the shear field will tend to align the fiber in the
direction of shear48,32. The origin of shear is a velocity difference between mix and wire. A
positive velocity difference will align fibers in the forward MD, while a negative difference will align
it in the negative MD. In both cases, an increase in fiber orientation anisotropy will result.
Figure 26. Principles for improving the evenness of fiber distribution: (a) turbulence, (b)
elongational flow, (c) dewatering and (d) oriented shear.
6.3.2 Forming fabrics
A detailed description of the properties and application of forming wires (alternative fabrics) is
given in Chapter 7. In this section, only a basic description is included to facilitate the reading of
the remainder of this chapter.
The forming fabric through which the mix is dewatered and on which the wet web is deposited
consists of a woven fabric connected into an endless band. The fabric is characterized by the
threads (materials, diameters) and the weave pattern.
The traditional bronze wire was of single-layer design, with a simple, symmetric, two-shaft
weave pattern where the structure is repeated every two threads both in MD and CD. The
smooth surface of a bronze wire was well suited to give a good paper surface structure.
During the 1970s, the bronze wire was replaced by the polymer fabric since the increasing
machine speeds gave the bronze wire too short a life span, due to the poor wear resistance of
the bronze threads. The life span was increased radically by a transition to wear-resistant
polymer threads.
The design of polymer fabrics was developed from single-layer to double-layer and
multi-layer, with much more sophisticated weave patterns than that of the bronze wires. To give a
fine, high draining paper side surface and at the same time a coarser, wear-resistant bottom
surface, multilayer designs seem to offer the best possibilities. There is still much development
taking place within the area.
From a paper quality point of view, the three-dimensional fabric surface structure has a large
impact both on paper surface structure and on dewatering resistance. The distribution of open
areas for dewatering and knuckles on the fabric surface has a large impact on the initial web
Papermaking Part 1, Stock Preparation and Wet End - Page 20

Chapter 6 Web forming

build-up and thus on paper surface structure (see Fig. 27).


Figure 27. 0.2, 4, 7, and 10 g/m2 fiber mats on a triple layer fabric49.
To generate a smooth paper surface, a flat fabric surface would be preferable. However, a
certain degree of surface three-dimensionality facilitates dewatering, since it will then be possible
for individual fibers to bridge the dewatering openings, without causing complete blockage.
6.3.3 Dewatering resistance
In pulp characterization, the Schopper Riegler number (SR) or the Canadian Standard Freeness
level (CSF), as evaluated in standard laboratory equipment, is often used to predict the
dewatering properties. However, two main reasons prevent pulp characterization as such from
being completely successful to predict drainage potential during papermaking:
- The structure of the web will depend on the details of the forming process.
- The web is highly compressible; therefore, the drainage resistance of a specific web will
vary with the dewatering pressure event.
Early studies of dewatering resistance were carried out by Ingmanson and co-workers 50,
according to the theories of filtration. They based their analysis on the Kozeny-Carman equation:
dQ
dt

3
1 (1C) 1
p
K S2C2

(19)

where dQ/dt is rate of drainage per unit area of the web,


p pressure gradient across the web,
C volume fraction of the web occupied by solids,
S specific surface area of the solids per unit volume,
viscosity of the fluid and
K Kozeny constant.
The Kozeny-Carman equation assumes laminar flow through parallel capillaries in an
incompressible medium. Neither of these assumptions is true in the case of wet paper webs. The
Kozeny-Carman equation in its present form thus cannot be used for predictions, and substantial
modifications must be incorporated before it can successfully be applied to paper forming.
Radvan51 summarized fundamental work on fourdrinier dewatering. Among other things, he
concluded that application of Eq. 19 results in values of the surface area S one order of
magnitude different from the surface area estimated using other methods.
As already mentioned, the compressibility of the wet web has a strong influence on the
dewatering rate since, already at the pressure levels used in a modern wire section, web
concentration will increase considerably (see Fig. 28).
Figure 28. Web concentration as a function of compression pressure 52.
After ordinary filtration dewatering with low dewatering pressure, the wet web concentration
will be in the order of 3%4%. As seen from the figure, web concentration will increase to above
10% with the application of a pressure of 10 kPa, a typical dewatering pressure level in, e.g.,
twin-wire roll forming. Filtration resistance is of course considerably increased at the higher
concentration level.
During the dewatering process in the wire section, a fiber suspension is forced against a
compressible, porous web, backed by the surface of a woven forming wire. The solid particles in
the suspension are to a certain degree trapped (definition of retention level), initially by the wire
surface and later by the accumulating wet web. The resistance to flow through the wet web

Papermaking Part 1, Stock Preparation and Wet End - Page 21

Chapter 6 Web forming

formed is then influenced by several parameters such as:


- Mix concentration and material composition
- Fiber properties (dimensions, degree of swelling, elastic properties)
- Chemical conditions, especially the agglomeration of fines and fillers to fiber material, using
retention aids
- Characteristics of the headbox jet
- Characteristics of the forming wire
- Time event of the pressure driving the dewatering.
The complexity of the overall situation has partly been considered during roll forming by
Martinez53, but further work is necessary to make accurate predictions possible.
For practical use, empirical equations are often used to evaluate dewatering capacity. One
example is the equation by Wahlstrm and OBlenes54:
t = kc w (p)

(20)

where tis dewatering time,


c mix concentration,
w deposited basis weight,
p dewatering pressure and
k, and n empirical constants.
6.3.4 Fourdrinier forming
The traditional paper machine is the fourdrinier machine, which was introduced around 1820 (see
Chapter 2). A horizontal wire was used, and all dewatering was one-sided. Classical fourdrinier
dewatering has been described in some detail 1,19,51. Figure 29 shows an example of a modern
fourdrinier machine.
Figure 29. Basic characteristics of a fourdrinier paper machine.
Originally, all dewatering relied on gravity effects and supporting rolls were introduced only to
keep the wire horizontal, while causing a minimum of friction drag. Today, dewatering devices are
applied below the fourdrinier wire with the object of creating dewatering effects and also of
providing means for controlling the degree of fiber flocculation in the sheet formed.
The setting of the jet onto the wire and the initial dewatering are extremely important for both
the dewatering capacity and the product quality. The jet should ideally hit the wire immediately in
front of a supporting forming board. If the jet hits the wire much ahead of this board, too rapid
dewatering will result in a deteriorated formation, whereas air will be trapped and disturb forming
if the jet hits the wire on top of the forming board.
Excessive pressure pulsations in the mix feed flow to the headbox generate jet velocity
pulsations when hydraulic headboxes are used. In fourdrinier forming applications, such velocity
pulsations can result in serious basis weight variations. Through amplification effects due to MD
wave generation on the wire, basis weight variations up to ten times as large as the jet velocity
variations have been observed 55.
6.3.4.1 Table rolls
As the machine speed was gradually increased, a speed range was entered where the rotating
wire support rolls the table rolls developed dewatering ability. Work performed by, e.g.,
Wrist19 described the basic principle for table roll dewatering, which was understood during the
1950s.
Papermaking Part 1, Stock Preparation and Wet End - Page 22

Chapter 6 Web forming

At a table roll, a suction pressure peak is generated in the downstream expansion zone
between wire and roll (see Fig. 30). This causes a local downward deflection of the wire
immediately after the roll, and a corresponding upward movement of the wire then has to occur
before the next dewatering element is reached. The vertical motions of the wire generate
instabilities in the mix on the wire and, to some degree, may improve final web formation.
Figure 30. Comparison between the suction pulses of a table roll (above) and of a foil element
(below) at 660 m/min19.
Before a table roll, the wire normally holds a water layer remaining from the preceding roll
dewatering on the bottom side, and this water is initially pressed up through the wire before the
above-mentioned suction pulse is generated. The washing effect which is thus introduced results
in a lower fines (and filler) content in the web on the wire side, i.e., the side of the paper facing
the wire during dewatering. At the same time, the drainage resistance in the lower part of the web
can decrease.
The amplitude of the table roll suction pressure pulse is proportional to the square of the wire
speed. There is therefore an upper speed limit for the application of table rolls of approximately
500 m/min, above which the vertical shaking is so intense that the high degree of disturbances
generated destroys final web formation. At higher machine speeds, the table roll can therefore no
longer be used as a dewatering element.
6.3.4.2 Foil elements
Foils were introduced during the 1960s19 and took over as the standard dewatering element at
machine speeds where table rolls could no longer be used. A foil element consists of a stationary
blade fitted below the wire, the upper side of which forms a small expansion angle with the wire,
generating an under-pressure. However, by choosing a small enough angle, sometimes below
one degree, a suitable suction pressure can be generated even at high wire speed. Figure 30
shows the difference in under-pressure pulse between a table roll and a foil element.
Figure 31. Reaction force F per unit width on wire with speed u, generated by the deflection of a
water layer of thickness h at an angle .
Much of the water sucked down by a foil element will cling to the wire underside and be
doctored off by the next foil element. To minimize the pressure pulse generated by the reaction
force F per unit width, from the deflection of the water layer thickness h, the front angle of the
downstream foil element should be as small as possible (see Fig. 31 and Eq. 21)56.
F =

2(1cos )
sin

h u2 (21)

If the front edge of the foil element is too sharp, wire life will deteriorate. The pressure pulse
generated by the front edge, and thus the amount of water pressed back through the wire, will be
much smaller than that in the table roll case since, in the latter case, all water under the wire will
be pressed through the wire.
To increase the dewatering effect, several foil elements may be mounted in low-vacuum
boxes. Dewatering then also takes place between the foil elements.
The overall optimization of the individual foil angles and the separation distances between the
elements is a complex procedure. To change the foil angle using the conventional design, the foil
element had to be replaced. However, modern designs allow on-line adjustment of foil angle.
Schmidt57 performed an experimental dewatering study and evaluated the instability level
introduced by watching the surface of the mix on the wire using a strobe light. The amplitude of
surface disturbances was graded as "activity" in numbers from 1 (smooth surface) to 10 (severe
Papermaking Part 1, Stock Preparation and Wet End - Page 23

Chapter 6 Web forming

mix jump). Among other things, he concluded that an increase in foil angle along the wire section
gave an optimum combination of dewatering capacity and web formation.
The mix layer surface interface with air on a fourdrinier machine and its effects on web
formation have also been studied by Kiviranta 58, using optical, geo-technical techniques.

Step foils
A step foil consists of a flat, short supporting surface from which a small (a few millimeters)
vertical step leads to a lower surface, parallel with the first surface. A step foil unit consists of a
number of foil elements fitted onto a vacuum box.
To change the degree of activity on a wire without having to change foil elements, as was
required with traditional foil elements, the vacuum level in the box can be changed. The higher
the vacuum is, the greater is the downward suction and thereby the greater the degree of activity
on the wire.

Iso-flo foils
Iso-flo foils are horizontal support blades mounted on a vacuum box. Every second blade is
slightly lowered so that, already at a moderate vacuum, the wire is sucked down toward the lower
supporting blades. The mix thereby receives a vertical pulsation movement. The vertical distance
between the blades then controls the degree of activity on the wire. The vacuum level and
thereby the dewatering capacity can be increased without influencing the vertical agitation, which
is not possible using traditional foils or step foils.
6.3.4.3 Dry suction boxes and couch roll
After the dry line, high-vacuum boxes with flat covers containing circular holes or slots increase
web concentration. Final suction dewatering is made on the couch roll. The vacuum levels range
from 1540 kPa in the vacuum boxes to 4080 kPa in the couch roll.
Dewatering in this range is a combination of web compression and friction effects by
transmitted air59,60.
6.3.4.4 Structure modification
To improve the formation on fourdrinier machines, the wire shake was moving the upstream part
of the wire section sideways, generating oriented shear in the mix on the wire. This was originally
introduced at low machine speeds, but it has later also been applied at speeds as high as
600700 m/min. As described above, "shaking" on modern machines is instead performed in the
vertical direction.
A dandy roll is a wire-covered open roll of cylindrical design. It is applied on the wire at a
position where the average concentration is around 2%3%, and it improves web formation by
introducing shear in the mix. The dandy roll is not a dewatering element, and the suspension
concentration is therefore approximately the same on the upstream and the downstream sides of
the roll. An even better effect would be achieved if the dandy roll were applied at an earlier stage
on the wire, but then severe downstream spouting results.
Dandy rolls are difficult to apply on high-speed machines for two reasons:
- If the roll is placed too early on the fourdrinier, the drops thrown towards the dry end will
create marks in the wet web.
- If the roll were moved to a higher dryness range, it is more difficult to keep clean, and
increasingly so at higher speeds.
As far as the effect of web formation improvement is concerned, the twin-wire former can be
considered as a development of the dandy roll, since it can be run with low in-going fiber
concentration and also at high speeds. Many fourdrinier machines, especially those

Papermaking Part 1, Stock Preparation and Wet End - Page 24

Chapter 6 Web forming

manufacturing printing grades, have been modernized to hybrid formers (see below) by the
introduction of "extended, dewatering dandy rolls" in the form of top wires.
The fiber orientation anisotropy in a web leaving the wire section depends partly on the
orientation in the mix jet from the headbox and partly on the change in orientation during the
dewatering process. During dewatering, longitudinal shear at the interface between wire/web and
mix, caused by a certain speed difference between jet and wire, introduces the combing effect
mentioned above. The fiber orientation anisotropy cannot be reduced during the dewatering
process; it can only be increased by oriented shear.
Danielsen and Steenberg performed the first investigation of the distribution in the z-direction
of a sheet of fiber orientation anisotropy by using polar diagrams to describe the distribution of
fiber orientation61. The orientation was found to be stronger on the wire side of a sheet since this
side is formed first and the relative motion between wire/web and mix gradually decreases as the
dewatering process proceeds.
Figure 32. Sheet anisotropy (breaking length) and formation as a function of jet to wire speed
difference48.
It is well documented that in fourdrinier forming, a difference between jet and wire speeds
affects formation as well as fiber orientation. When the speed difference deviates from zero by
small amounts, formation improves. This could be caused by the breaking apart of fiber flocs
through shear, or it could be a result of the supply of extra fibers for local low basis areas through
relative movement of mix in the plane of the web. At larger speed differences, disturbing effects
will cause formation to deteriorate.
Svensson and sterberg studied sheet anisotropy and formation on an experimental paper
machine (see Fig. 32).
They demonstrated that anisotropy and formation are dependent on the jet-to-wire speed
difference. Traditionally the influence of jet-to-wire speed ratio had been studied.
It should be noted that in addition to the effects of fiber orientation anisotropy, the breaking
length anisotropy of a finished paper is also affected by the web tension conditions during
drying.
Lately, the concern about fiber orientation misalignment has increased, partly because of
increased product demands and partly because of improved instrumentation to detect such
conditions. Fiber orientation misalignment can be caused by a headbox jet velocity vector not
exactly in line with the MD or by sideways flow generated on the fourdrinier wire section.
Holik and Weisshuhn62 demonstrated that even with an angular deviation between jet and
machine directions as small as a half degree, a considerable cross flow velocity arises on a
modern high-speed paper machine. The sum of this cross flow velocity component and the
difference between machine speed and jet MD-speed forms a velocity vector, representing the
relative suspension movement on the wire (see Fig. 33).
Figure 33. The relative mix velocity vector urel as a function of the difference between jet and
wire speeds, uj- u w and cross flow ucross. Left: uj- uw> 0, Center: uj- uw= 0, Right: uj- uw < 0.
They pointed out that local variations in the angle of this relative velocity vector are largest
when the jet to wire velocity difference is low and that the misalignment angle will change sign
with a change from excess jet velocity to excess wire velocity.
6.3.5 Twin-wire forming
The higher the machine speed is, the longer the wire section is required to dewater a given
product. On a fourdrinier at high machine speeds, problems with an unstable liquid surface
Papermaking Part 1, Stock Preparation and Wet End - Page 25

Chapter 6 Web forming

increase, depending on instabilities in the mix and on a considerable friction between mix and
air.
Attention was already being given to these problems during the 1950s, and experiments then
began in which the mix was trapped between two wires, so-called "twin-wire forming." This new
principle had the advantages of avoiding the free surface between mix and air, of increasing the
dewatering capacity by allowing dewatering through two wires (see Fig. 34), and of the possibility
of symmetrical dewatering and thus also symmetrical sheet structure. Since, in each direction,
half the dewatering will take place through half the basis weight, dewatering capacity will in
principle quadruple compared to one-sided dewatering.
Figure 34. Basic principle of two-sided dewatering.
In twin-wire formers, two webs are formed, each half of the total basis weight. This means
that a lower retention level can be expected in comparison with fourdrinier forming since retention
increases with web basis weight. Also the potential of sheet strength will decrease, to some
extent, since the self-healing effect on fiber distribution will be lower at lower basis weight levels
(see previous section).
Principles for twin-wire forming were patented as early as during the last century, but not until
the 1950s were machines developed that could be used for practical paper and board production.
Figure 35. Roll former according to Websters U.S. patent63.
David Webster invented and demonstrated the twin-wire roll former principle for printing
paper production during the 1950s. Figure 35 shows the principle design demonstrated in the
form of a patent drawing.
During the 1970s, this principle was commercialized as the Papriformer (Dominion
Engineering, Paprican) and the Periformer (KMW).
Brian Attwoods Inverform design (Fig. 36) was developed at St. Annes Board Mill64. This
was a twin-wire arrangement in which the separate board layers were successively deposited on
a long, bottom fourdrinier wire. Apart from the first layer, all layers were dewatered upward
through top wires, which created twin-wire nips together with the web on the bottom wire.
Figure 36. Inverform unit for dewatering of a board ply.
During the 1960s, twin-wire blade formers with stationary dewatering elements were
developed by Beloit (Bel Baie) and Black Clawson (Vertiforma).
In pure twin-wire formers, the mix jet is delivered directly into the gap between the two wires,
hence the term "gap former."
In a hybrid former, a fourdrinier section precedes the twin-wire nip. Most hybrid machines
were originally fourdrinier machines that were later complemented with an upper wire for
dewatering upward at the end of the wire section.
During the 1970s, twin-wire forming became the dominating principle for newly installed
printing paper machines. The following advantages could be attained with twin-wire dewatering:
- Increased dewatering capacity
- More symmetric paper product
- Lower basis weight variability
- Better formation
- Lower linting (fiber release during printing).

Papermaking Part 1, Stock Preparation and Wet End - Page 26

Chapter 6 Web forming

During the 1990s, twin-wire machines for the production of packaging paper have grown in
importance, with the main reason being their ability to produce paper with less variability in basis
weight and their high dewatering capacity.
6.3.5.1 Roll forming
As mentioned above, David Webster developed the original roll former concept during the early
1950s65. This design avoided the simultaneous mechanical guiding of both wires at the same
location, which was a task impossible to perform with the required accuracy and the main reason
why the older designs never had been successful. The secret to success was to never try to
simultaneously guide both wires; at all positions at least one of the wires should be able to move
freely in the transversal direction.
Twin-wire roll formers dominated the installation of new printing paper machines during the
1970s. In a twin-wire roll former, a mix jet is injected into the nip between two wires wrapping a
rotating roll. Dewatering can take place through both the outer and inner wires. The latter water is
kept in the open roll surface and released after the inner wire has been removed from the roll
(see Fig. 37). The drained water is thereafter led to the wire pit. The water passing through the
outer wire is also led to the wire pit.
Figure 37. Twin-wire roll former with two sided dewatering.
The outer wire is given a tension T [kN/m] and, using classical mechanics, the fluid pressure
p [kN/m2] generated in the mix to support the wire is then given by the equation
p=

T
R

(22)

where R [m] is the local radius of curvature of the wire.


The outer wire approaches the forming zone along a straight path, R = . There is a gradual
decrease of outer wire radius from infinity to roll radius R and a corresponding gradual build-up of
a dewatering pressure (see Fig. 54). The roll radius is an acceptable approximation for wire
curvature R after the initial phase since the separation distance between the outer wire and the
roll surface is small relative to the roll radius.
The roll former can be characterized as an easily run (few variables) and reliable process,
which can accept some degree of imperfections in the mix jet without too negative effects on
paper formation. With the headboxes used during the 1970s, due to the gentle type of
dewatering, roll forming did not give top levels of formation. However, later experiments indicate
that headboxes designed for a high fiber orientation anisotropy level in the mix jet can produce
roll formed paper with improved formation (see further discussion below).
The undisturbed dewatering process gives a relatively high retention level, which reduces the
need for retention additives.

Centrifugal effects
If a fluid element of radial extension h and density moves with velocity u along a path with a
radius of curvature R, a centrifugal pressure p c develops according to the equation:
2

pc = h uR (23)
It is often claimed that the dewatering pressure in a roll former is generated by the centrifugal
forces. This is fundamentally wrong, since the dewatering pressure is only created by the applied
tension in the curved outer wire. Because of the influence of the centrifugal force, however, the
pressure drop across the inner wire decreases by the amount pc65. To compensate for this and
thus maintain the same pressure drop across the outer and the inner wires, a corresponding

Papermaking Part 1, Stock Preparation and Wet End - Page 27

Chapter 6 Web forming

vacuum pressure p c is created with the aid of the suction zones inside the roll (see Fig. 37).
The ratio between centrifugal pressure pc (Eq. 23) and wire-generated pressure pw (Eq. 22)
can be described by the following equation:
pc
pw

hu2
T

(24)

If the centrifugal pressure were to exceed the wire pressure, i.e., if the ratio in Eq. 24 were to
exceed unity, unstable conditions would result. This means that there is a practical upper limit for
the ratio pc/pw of unity. From Eq. 24, it is clear that these conditions are reached independently of
the local radius of wire curvature. Figure 38 shows corresponding values of maximum fiber
suspension radial thickness h at wire speed u.
Figure 38. Maximum wire speed u for stable dewatering conditions as a function of suspension
radial thickness h and wire tension T66.

Jet to wire speed difference


On a fourdrinier machine, minimum fiber orientation anisotropy is achieved at a jet/wire speed
ratio of unity, which corresponds to equal velocities of jet and wire.
When the mix jet from the headbox enters the wire gap of a roll-former unit, it is met by a
dewatering pressure, gradually rising to p = T/R, which means that the jet is decelerated. If the
energy equation (Bernoulli's equation) is applied, we will get the following relationship between
the mix velocity uj in the free jet and the decelerated mix velocity um between the wires:

u2
j
2

+0=

u2
m
2

T
R

(25)

where is mix density.


If the lowest possible fiber orientation is desired, the mix after the deceleration at the inlet to
the forming zone should assume the same speed as that of the wires. The combing effect of
oriented shear is then minimized during the rest of the dewatering process. According to Eq. 25
the jet-to-wire speed ratio is then expressed by the following expression, where index o denotes
minimum orientation effects.

q
uj
T
=
1 + u22 R
(26)
uw
0

This means that, if a paper with the least possible fiber orientation anisotropy is desired on a
twin-wire roll former, a jet-to-wire speed ratio greater than 1 must always be used. This differs in
principle from the fourdrinier case, where the least fiber orientation is attained at a speed ratio
equal to 1.
If the jet velocity is higher than that given by Eq. 26 for "isotropic" conditions, the mix speed
will also be higher than the wire speed during dewatering, and vice versa.
It follows from Eq. 26 that the jet/wire speed ratio for minimum fiber orientation anisotropy
increases with decreasing machine speed, and this is exemplified in Table 5:
Table 5. Jet/wire speed ratio for minimum fiber orientation anisotropy at different wire
speeds with T = 6 kN/m and R = 0.8 m.
Wire speed
100
200
500
1 000
2 000
m/min
Jet/wire
2.53
1.53
1.10
1.027
1.007
speed ratio
It should be pointed out that when the jet decelerates in the twin-wire gap, the conservation of
Papermaking Part 1, Stock Preparation and Wet End - Page 28

Chapter 6 Web forming

mass will require a relative increase in mix thickness equal to the jet/wire speed ratio given in
Table 5. Too high an expansion, however, is not possible while maintaining stable flow
conditions; therefore, a roll former cannot be run at speeds below approximately 200 m/min.
It should also be pointed out that the absolute speed difference rather than the speed ratio
between mix and wires is physically more relevant in determining final paper properties. Due to
the simplicity of Eq. 26, it is however easier to carry out the principal discussion above based on
speed ratio.
As described above, in roll forming the pressure is constant during a large part of the
dewatering process. This means that it is not possible to introduce local disturbances, such as
with a foil element in a fourdrinier wire section. The only way to improve formation then is to run
with some speed differential between mix and wires. Within a limited range, this will have a
positive effect on formation through oriented shear, as will be explained further below.

Paper properties
As already mentioned, it has proved positive to use a headbox with a high nozzle contraction
ratio. Figure 39 shows how the fiber orientation anisotropy, evaluated as the tensile stiffness
anisotropy, depends on the speed difference between the mix and the wires in twin-wire roll
forming.
Figure 39. The influence on tensile stiffness anisotropy of the speed difference between the
decelerated mix jet and the wires at different headbox nozzle contraction ratios. Twin-wire roll
forming, 700 m/min, TMP furnish, FEX trials33. Squares = low contraction, filled circles = medium
contraction, unfilled circles = high contraction.
The traditional result (see Fig. 32) is shown by the square symbols representing a headbox
with a nozzle contraction ratio of 7. Based on the minimum point, the anisotropy in the jet itself
can be estimated to be about 1.6. Anisotropy increases rapidly with increasing speed difference
between mix and wires. The filled circles represent a headbox with nozzle contraction 17 and a
jet anisotropy around 2.4. Also in this case, the anisotropy of the paper increases with increasing
speed difference. The unfilled circles represent a headbox with nozzle contraction ratio 30, where
orientation anisotropy remains practically unchanged at different speed differences. Already in
the jet, the anisotropy is above 4 and only a small decrease results when running with a
high-speed difference between mix and wires. The conclusion is that, at a sufficiently high
contraction ratio in the headbox nozzle, the jet develops such a high degree of fiber orientation
anisotropy that it cannot be further increased during the dewatering process.
The experiments described above were performed using twin-wire roll dewatering. Later
investigations have compared identical headbox designs using twin-wire roll dewatering and
fourdrinier dewatering respectively67. It was then concluded that anisotropy effects corresponding
to those shown in Fig. 39 were obtained only with twin-wire roll dewatering and not with
fourdrinier dewatering. This demonstrates the importance of the basic difference between the two
dewatering principles. In twin-wire roll dewatering, the flow instabilities in the jet are dampened
due to the enclosure between the two wires, and the dewatering is relatively fast. In fourdrinier
dewatering, on the other hand, flow instabilities can grow in the open mix layer on the wire.
Combined with longer drainage times, this allows for considerable fiber reorientation.
Figure 40 shows the influence of nozzle contraction ratio on the small-scale formation of
paper, i.e., within a wavelength range of 0.33 mm.
Figure 40. The influence on small-scale formation of the speed difference between the
decelerated mix jet and the wires at different headbox nozzle contraction ratios. Forming data
and legends as in Fig. 3933.
Papermaking Part 1, Stock Preparation and Wet End - Page 29

Chapter 6 Web forming

The square symbols represent the traditional case with comparatively low nozzle contraction
ratio, and the formation is initially improved with increasing speed difference between mix and
wires. This formation improvement has earlier been described as being the result of oriented
shear. A further increase in the speed difference leads to a poorer formation.
The figure also shows that, at a contraction ratio of 30 (unfilled circles), the situation is
changed. The best formation is instead obtained at a speed difference of zero. This must be due
to an improved state of flocculation already in the headbox jet, at high nozzle contraction ratio.
The formation is then worsened by the application of oriented shear through a speed difference
during dewatering.
It has also been shown that, with high nozzle contraction ratio, large-scale formation that is,
formation in the size range of 330 mm showed a rather wide optimum around zero speed
difference in roll forming33.
From these results, the conclusion can be drawn that for a TMP-containing product and
where a high level of anisotropy is desired, a headbox with high nozzle contraction ratio should
be used. It should be operated without speed difference between mix and wires, which would
give optimum small-scale as well as large-scale formation. The anisotropy should be controlled
by the nozzle contraction ratio, i.e., through the size of the lip opening.
A comparison of the effect of mix-to-wire speed difference on paper formation can be made
between fourdrinier (Fig. 32) and twin-wire (Fig. 40) dewatering respectively. It could be observed
that in fourdrinier dewatering much lower speed differences are allowed before deterioration of
formation is initiated. This can be explained by the constant dewatering pressure in the roll
former, which continuously will keep the forming webs compressed against the wires. In
fourdrinier dewatering, on the other hand, dewatering is intermittent, and therefore the web will be
uncompressed when dewatering pressure is not applied. During such times, the wet web
resistance against shear forces from the mix will be low, the web may loosen from the wire, and
the formation may deteriorate.
Analysis of fiber orientation anisotropy at different levels in the thickness direction of a paper
sample was made according to the method described above 34. This technique was applied to
investigate the influence on anisotropy by mix to wire speed difference and by headbox nozzle
contraction ratio respectively (see Fig. 41).
Figure 41. Fiber orientation anisotropy at different thickness levels of a paper sample. FEX
twin-wire roll forming experiments, 60 g/m 2, hardwood/softwood-mixture, 700 m/min. Headbox
contraction ratio 8.5. Different mix-wire speed differences: 0 m/min (squares), +45 m/min
(circles), 75 m/min (triangles)35.
From the figure it can be concluded that, with a low headbox nozzle contraction ratio and no
speed difference between mix and wires, there is a low degree of anisotropy throughout the
paper sample.
Fiber orientation anisotropy can be obtained in two ways:
- By using a low contraction ratio nozzle and a positive or negative speed difference during
dewatering.
- By using a high contraction nozzle and no speed difference during dewatering
At sheet center, a slight anisotropy minimum appears in the speed difference case. This
could be explained by the presence of fiber flocs in the mix, which would from purely geometrical
reasons be squeezed toward the center during dewatering, and on which oriented shear would
have less directional effects than on freely moving fibers.
The lower orientation toward the paper surfaces has been suggested to depend on the short
Papermaking Part 1, Stock Preparation and Wet End - Page 30

Chapter 6 Web forming

drainage time for the layers close to the wire and thus on the short times for the oriented shear to
align the fibers34. However, the degree of orientation anisotropy follows the same pattern without
speed difference, at a high nozzle contraction ratio. As previously discussed in connection with
Fig. 17 the lower degree of anisotropy toward the paper sides could therefore alternatively be an
effect of local flow disturbances when the jet hits the wires, and the evolving dewatering pressure
decelerates the mix between the wires. There could also be less initial fiber orientation anisotropy
on the two sides of the mix jet, as a result of boundary layer effects along the upper and lower
headbox lips.
The assumption of a constant dewatering pressure during roll forming resulted in the jet
speed required for minimum fiber orientation anisotropy listed in Table 5. As already mentioned,
the pressure in the initial dewatering phase only gradually reaches the static value assumed in
Table 5. This means that, in reality, there are difficulties to determine what jet speed to actually
aim for to generate minimum anisotropy. Furthermore, the exact value of the jet speed is in
practice not possible to accurately measure in a gap former. For experiments on the FEX
machine, a special method has therefore been developed to find the "effective mix speed" giving
minimum fiber orientation effects during forming. The mix flow rate and headbox parameters are
kept at constant values, and instead the wire speed is gradually changed. At each speed, paper
samples are analyzed for anisotropy (ultrasonic measurements of orientation distribution in
elastic modulus), and the wire speed giving minimum anisotropy then equals the "effective mix
speed." It is evident that some differences in mean basis weight will result from this way of
evaluating mix speed, but most paper properties can be normalized for basis weight differences
within a limited range.
6.3.5.2 Blade forming
During the 1960s, twin-wire dewatering principles based on stationary dewatering elements were
also developed. In the Black Clawson Vertiforma design, stationary elements were mounted on
both sides of the wires, initially opposing each other but later positioned in a staggered mode
(see Fig. 42).
Figure 42. Black Clawson Vertiforma twin-wire former.
Paper scientists tried to theoretically describe the interrelationship between forming
conditions and dewatering pressure. The energy equation (Bernoulli equation) was applied to the
mix jet flowing into a predefined space between the forming wires, generally with linearly
decreasing cross section. This was an unsuccessful approach since the primary variable in
reality is the local wire curvature, which determines dewatering pressure according to Eq. 22. It is
then obvious that no dewatering pressure could exist across a straight, free wire.
In the Beloit Bel Baie design, a curved slotted forming shoe was placed on one side of the
wires. As late as in the middle of the 1970s, the basic dewatering principle for such formers was
explained as being the same as that in roll forming68. The dewatering pressure over the forming
shoe was calculated according to Eq. 22, and the radius R of mounting of the shoe element ends
was inserted for the radius of wire curvature. This radius was one order of magnitude larger than
that of a roll former. The hypothetically lower dewatering pressure was suggested to be the main
reason for the improved formation typical for Bel Baie forming in comparison with twin-wire roll
forming.
Figure 43. Reaction force F keeping a deflected, tensioned wire in position56.
However, Norman emphasized the pulsating nature of dewatering using forming shoes in
197756. The local deflection angle 2 of an outer wire of tension T over the individual forming
Papermaking Part 1, Stock Preparation and Wet End - Page 31

Chapter 6 Web forming

shoe elements was suggested to be the basis for the generation of a dewatering force F (see Fig.
43).
The dewatering force F was calculated according to Eq. 27.
F = 2T sin (27)
The shape of the pressure pulse producing the forceF depends on several factors such as
wire speed, wire separation, and wire tension.
It was further suggested that the pulsating nature of the dewatering pressure generates shear
in the mix between the wires, thus improving the formation of the web produced.
For several experimental results in this chapter, reference is made to the FEX machine65,
69,70, which is a pilot paper machine started up at STFI in Stockholm in 1982. The new ideas
about pulsating dewatering pressure were applied in the design of the wire section of the FEX
machine, and the term "blade" was introduced for the deflector elements composing a "blade
former." Several fixed blades mounted together constitute a "blade shoe."
Figure 44. Wire movement with deflector blades on one side (top) and on both sides (bottom).
Solid lines: Original wire positions. Broken lines: Doubled distance between the wires.
If all blades are placed on the same side of the wires, the pressure pulses are practically
unchanged if the separation between the wires is changed since the external wire is freely mobile
in the thickness direction. This means that the dewatering pressure, like in a roll former, is
self-adjusting at, e.g., changed dewatering properties of the mix or at a change in basis weight.
However, if one blade is positioned on the opposite side of the wires, the deflection angles
across the individual blades will become strongly dependent on the separation between the wires
(see Fig. 44).
At the original wire separation, the wires could pass straight between the blades, whereas a
doubling of the wire separation means strongly increased deflection angles. This means that the
position of the "opposite" blade is extremely critical, and necessarily has to be adjustable on-line
to accommodate with the actual running conditions. In the traditional twin-wire blade formers
(Vertiforma, Bel Baie), this however was not possible.
Figure 45. Pressure pulses at four consecutive blades in a Bel Baie former. Basis weight 48.8
g/m2, wire tension 7 kN/m, and speed 920 m/min71.
Measurements of the shape of the pressure pulses were first made by Beloit, using a static
pressure probe with a thin metal tube extending from within the headbox into the mix between the
wires (see Fig. 45).
Figure 46. Pressure distribution along a triangular blade, with a 2.3 degree top angle, for two
different deflection angles around the downstream edge. FEX experiment, machine speed 1 200
m/min, wire tension 6 kN/m, wire separation circa 2 mm73.
Kerekes and Zhao72 made experimental evaluation and mathematical modeling of the
one-dimensional blade pressure event along a flat blade.
Figure 46 demonstrates pressure pulse shapes measured through pressure tappings drilled
along a triangular shaped blade.
The figure clearly demonstrates that the pressure pulse generated by the deflection at the
middle of the blade is developed almost completely before the local change of blade angle. For a
flat blade in a blade-former, with wire deflection across the front edge, the pressure pulse will
thus fall almost completely ahead of the blade. This explains why approximately the same
Papermaking Part 1, Stock Preparation and Wet End - Page 32

Chapter 6 Web forming

amount of dewatering can be obtained through both wires in a blade former where all blades are
placed on the same side of the wires.
Zahrai74 extended the theoretical analysis of blade pressure pulses to the two-dimensional
case by also considering gradients between the two wires.
It should be pointed out that a pressure pulse is also generated when a blade doctors off a
water layer on a wire (see Eq. 21), forcing water back through the wire.
There is a general tendency that blade formers with pulsating dewatering pressure, when
compared with roll formers with more uniform dewatering pressure, produce sheets with superior
formation. However, the retention level for fines and fillers will be lower.
The fiber orientation anisotropy cannot be manipulated to the same degree through variation
of the speed difference between jet and wires in a blade former as in a roll former, since the
blade pressure pulses also have an additional effect on anisotropy.
6.3.5.3 Hybrid forming
During the 1970s, initial fourdrinier forming followed by twin-wire forming was an alternative
for new paper machine installations such as the Valmet SymFormer. During the 1980s, this
principle was also applied for the rebuild of fourdrinier machines for printing paper production.
The first design was the Dynaformer from Dominion Engineering 75.
Figure 47. Dominion Engineering Dynaformer75.
Due to basic differences mentioned earlier, regarding the speed difference between jet and
mix on a fourdrinier in comparison with a twin-wire machine, the conditions for dewatering of the
two web sides differ in hybrid forming. To produce a web with a low degree of fiber orientation
anisotropy, jet and wire speeds should be equal on the initial fourdrinier wire section when the
bottom side of the web is formed. However, this also means that mix speed will equal wire
speeds at the entrance to the twin-wire nip, which according to Eq. 26 will result in a deceleration
to a mix speed inside the twin-wire nip, lower than wire speed. Some increase in fiber orientation
on the topside of the web, through oriented shear during dewatering, will then take place. It is
thus not possible to optimize the dewatering conditions for both web sides in a hybrid former.
During the 1980s, major paper machinery manufacturers rebuilt a large number of printing
paper machines, using principles similar to that shown in Fig. 47. Sometimes also some blade
forming was included. In the Black Clawson Top-Flyte former, a blade arrangement similar to that
in the Vertiforma (Fig. 42) was applied.
6.3.5.4 Roll-blade forming
Also during the 1980s, new forming combinations of dewatering blades, forming rolls, and blade
shoes (forming shoes) were introduced.
Figure 48. Slotted blade shoe with radius R (broken line) and blade tip lengthl = L/2. Left: Blade
tip curvature R. Right: Blade tip curvature r = R/3 66.
If a blade shoe is designed as a curved surface with parallel slots, pressure pulses will be
generated due to the wire deflections at the blade edges (see Fig. 48).
To avoid local wire deflections, the radius of the blade tips must be decreased in proportion to
the relative blade length, l / (l + L). In reality, flat blade tips may be used if the wire deflections are
small. To avoid too large local wire deflections and for excessive wire friction wear reasons, the
tips may be polygon shaped so that the total deflection is distributed on smaller, individual
deflections.
It should be pointed out that, due to the dynamic effects, the upstream pressure pulses will
Papermaking Part 1, Stock Preparation and Wet End - Page 33

Chapter 6 Web forming

appear in front of the blades (see Refs. 72 and 73).


To increase inward dewatering, vacuum is often applied in blade shoes. The vac-uum will
deform the web and inner wire with a radius of curvature according to Eq. 22. It has been
suggested that the outer wire will also follow that shape, which would generate increased
pressure pulses at the blade ends76. This is unrealistic, however, since the pressure drop will be
almost completely absorbed by the wet web. The pressure drop across the outer wire is therefore
quite insignificant, and it will travel along straight lines between the individual blades.
Figure 49. Bel Baie III twin-wire former by Beloit.
Beloit designed a former with initial blades followed by a blade shoe and roll dewatering (see
Fig. 49).
Valmet applied the opposite philosophy and used initial roll dewatering, followed by blade
shoes (see Fig. 50). The C-former from Black Clawson also combined roll and blade forming.
Figure 50. Valmet Speedformer HS with initial roll forming followed by forming shoes. 112
denotes positions for collection of white water.
In high basis weight board forming, which takes place at relatively low speeds, it is difficult to
avoid flow instabilities in a roll nip. In the early 1980s, the Drries Company (now part of the Voith
Sulzer Group) developed a new method of twin-wire board forming by placing a top wire on a
fourdrinier wire, where blades located below the wires could be pressed upward with individually
adjustable loads77. This was the first twin-wire arrangement where the amplitude of the
dewatering pressure along the forming zone could be controlled on-line. From the patent, it can
be concluded that the pulsating nature of the dewatering event was not recognized. In all
dewatering segments, the free wire is straight, which means that no dewatering pressure can
exist. In reality, the wires move in a zigzag pattern similar to that in Fig. 44.
Figure 51. Forming part of Voith top-wire former Duoformer-D with individually adjustable
blades78.
Figure 52. The STFI-former: Initial dewatering at constant pressure followed by adjustable
pressure pulses.
Voith modified the Drries design to the Duoformer-D version (Fig. 51).
A combination of initial constant dewatering pressure followed by adjustable pressure pulses
was implemented on the FEX-machine in 1991, to form the STFI-former (Fig. 52)70.
The headbox and the outer wire lead-in roll are movable around the center of the forming roll
and can be repositioned during operation, thereby adjusting the roll cover angle. Downstream of
the forming roll, there are blade arrangements where the right-hand blades are mounted along a
vertical line. To the left, there are six movable blades with position indicators and individually
adjustable pressure forces F.
This arrangement, with symmetrical dewatering in all stages, makes it possible to optimize
the distribution of the dewatering between constant and pulsating pressure events and to adjust
the amplitude of the individual pressure pulses. By the choice of a suitable blade shape, it is also
possible to obtain the desired combination of pulsating and constant dewatering pressure along
the individual blades. The main aim of the STFI former was to improve the possibilities to form
multilayered printing papers.
Contrary to the complete dewatering in early roll formers, a comparatively small wire cover

Papermaking Part 1, Stock Preparation and Wet End - Page 34

Chapter 6 Web forming

angle (minimum 25 degrees) over the roll is utilized in the STFI former. It is important to avoid a
too large sheet deposition at this stage, since enough undewatered mix has to remain between
the forming webs at the blade section. This makes it possible to improve the formation by the
application of suitable blade pulses.
Figure 53. Roll-blade, Duoformer CFD forming unit by Voith.
Figure 54. Pressure event along a forming roll of diameter 1 635 mm. Wire wrapping angle 30
degrees, wire tension 7 kN/m, and machine speed 800 m/min 53. FEX measurements in
co-operation with Valmet.
A similar development as that at STFI was carried out within the Voith Corporation. Figure 53
demonstrates a Voith design from the early 1990s.
The three large paper machine manufacturers Voith Sulzer, Valmet, and Beloit now apply
the basic principle used in the STFI-former/Duoformer CFD*.
Figure 54 demonstrates the pressure event along the forming roll, using a rather small
wire-wrapping angle along the roll periphery.
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 (C).
It is clear that the dewatering pressure does not reach the level T/R predicted by Eq. 22 until
the final stage of the roll dewatering. The pressure event during the initial part of roll forming has
not yet been modeled theoretically.
It should also be pointed out that, by using a limited cover angle on the roll, the separation
between wires and roll surface might be critical. The local vacuum generated at the separation
point by "the table roll effect" can have a negative influence on the comparatively thin webs
already formed on the two wires.
6.3.6 Multi-ply/multilayer forming
Multi-ply forming has traditionally been applied to board forming for two main reasons:
- Dewatering resistance increases exponentially with mean basis weight, which means large
practical problems to form a high basis weight board product as one single web.
- Economic use of different fiber raw materials requires selective positioning of different fiber
qualities in the thickness direction of a board product.
The forming of multi-ply board products using separate forming units for each ply therefore
has a long history.
Multilayer tissue products are produced in a simultaneous forming process, in which a
multilayered jet is formed by using a multilayer headbox and is dewatered in a twin-wire former.
This type of forming was developed to take advantage of different furnish compositions, without
the costly installation of several dewatering sections.
Simultaneous forming is sometimes called "multilayer forming," in contrast to the term
"multi-ply" forming for the case when separate forming units are used. These definitions,
however, are not generally accepted; neither are the terms self-evident. It is therefore preferable
to use the terms "separate" and "simultaneous" to discriminate between the two forming
principles.
6.3.6.1 Separate forming
The traditional board former machine was the vat former. A vat unit consisted of a wire-covered,
rotating open roll, partly submerged into a trough filled with mix. A wet web was gradually built up
during the rotation of the roll in the mix. The top of the roll was pressed against a felt, on the
Papermaking Part 1, Stock Preparation and Wet End - Page 35

Chapter 6 Web forming

bottom of which all the individual vat plies were successively couched. To improve the speed
potential, the trough was reduced in size, and eventually designed as a headbox, in which the
bottom lip was the rotating roll surface. The Bristol former and the BRDA former have such
designs, and the limitation is the deviation from perfect circular roll shape. This will create
periodic variations in headbox slice opening, and thus also in basis weight.
The first twin-wire board forming unit was the Inverform machine, developed during the 1950s
(see Fig. 36). This was later modernized to the Beloit Bel Bond unit (see Fig. 55) several of which
were placed along a fourdrinier wire to form a multi-ply product. In these machines, each
additional layer is formed against the previously formed sheet.
Figure 55. Beloit Bel Bond board forming unit.
The Voith Duoformer-D twin-wire unit (see Fig. 51) makes it possible to control the individual
dewatering pressure pulses along the former. This principle has now also been applied in the
Symformer-MB by Valmet and the Bel Bond CB with Counter Blades by Beloit. In all these
designs, additional layers are formed against the existing sheet.
An alternative way of board forming is to use a separate fourdrinier section for each ply, and
then couch the individual plies together on a common bottom wire. This principle improves the
possibilities to separately optimize each individual ply, but is a more costly solution.
Recently, also modern twin-wire principles (roll and blade formers) have been applied by
separately forming the individual plies, and couching them onto a common bottom wire. An
example is the Voith Duoformer Top, see Fig. 56, which forms the bottom ply.
A basically similar unit, the Duoformer Base, is used to form the bottom ply.
A critical board property is the ply bond strength between the individual plies, and this is
mainly a problem when separately formed plies are couched together. Earlier it was thought that
web dryness at couching was the critical parameter, but it has now been shown that fines content
at ply interface is a decisive parameter. If individual plies are formed with one-sided drainage, the
wire side will normally have a lower fines content. A wire section design should therefore be
avoided in which plies are couched together with two wire sides facing towards each other. To
improve ply bond, it may sometimes be necessary to spray a starch solution or white water
containing fines between two layers before couching them together.
Figure 56. Voith Duoformer Top board forming unit.
6.3.6.2 Simultaneous forming
For the manufacture of tissue products, two and three layer headboxes were introduced during
the 1980s. From a single headbox, a layered jet is then delivered with different furnish in the
different layers. This provides better possibilities of optimizing both the creping process on the
yankee drying cylinder and the softness of the final product.
Test liner, that is liner made of recycled fibers, is another product where two-layer headboxes
are applied industrially. It is then possible, e.g., to improve the printing properties by using
different mix composition in the surface and bottom plies respectively.
With a layered structure, there would be a great potential for raw material savings or quality
improvement in the manufacture of layered printing papers. It would then be possible to use
different raw materials in the surface layers and middle layer, or alternatively to fractionate a
given stock such as a mechanical pulp, and to place the different fractions at different positions in
the thickness direction. It would also be possible to use different types and amounts of filler on
the surfaces and in the middle of the web. To obtain material and filler distribution effects in the
z-direction, it is also possible to selectively add different amounts and types of retention aids to
Papermaking Part 1, Stock Preparation and Wet End - Page 36

Chapter 6 Web forming

the surfaces and the center.


Considerable developmental work is being carried out to form a three-layer printing paper.
The basic problem is still the headbox design, with high demands on jet layering with good layer
separation and smooth jet surfaces.
6.3.7 Forming and paper strength
It is sometimes suggested that improved paper formation also implies a higher strength level.
Unfortunately, no such universal relationship exists. It is well known that a decrease in forming
concentration improves both paper formation and strength. Therefore, a standard laboratory
sheet, formed at about 0.1 g/L, yields both better strength and formation when compared to
paper formed on a paper machine at a concentration of, e.g., 5 g/L. It is further well known, that
with a delay time between final mixing and dewatering in a laboratory sheet former, fiber
flocculation will take place, which will result in a deterioration of paper formation as well as of
paper strength.
The situation is however quite different if formation is improved by pressure pulses during
twin-wire forming. In a new linerboard machine of the twin-wire blade former type, a high final
pressure pulse improved the overall linerboard formation but, at the same time, the burst value
deteriorated79. Large flocs might have been broken apart thus improving large-scale formation,
but this was accompanied by some deterioration of small-scale formation. The broken flocs then
contributed to some small-scale variations and, because of insufficient bonding between the floc
fragments, strength properties might have been reduced.
Later investigations have confirmed that improvement of large-scale flocculation by twin-wire
pressure pulses does reduce paper strength 33. As an example, the delamination strength of
paper is an increasingly important property, e.g., at high running speeds of printing presses and
in board products. It was found that in twin-wire forming, the introduction of pressure pulses
during the later stages of dewatering has a considerable negative effect on delamination
strength. This can be caused by breakup of fiber network elements in an oriented shear field at
web center.
As mentioned in an earlier section, fourdrinier forming has the potential to develop better
strength properties than twin-wire forming in the plane of the paper, due to the larger self-healing
effect on fiber distribution with one-sided dewatering. One-sided dewatering also has an
advantage regarding the delamination strength. In two-sided dewatering there will be a reduced
amount of fines at the dewatering center plane, which will create a locally low delamination
strength.
Excessive addition of retention chemicals did improve paper formation; however, the
formation improvement was not accompanied by an improvement in strength, which mainly
remained unchanged80. One possible explanation of the formation improvement is the effect of
the retention aid polymers to increase the elongational viscosity of the draining fluid 81. This could
reduce fiber interaction and thus reduce fiber flocculation.
Fredlund82 summarized paper strength in relation to forming and concluded that the potential
for paper strength is determined by the properties of the mix in the headbox jet. Manipulations
during the dewatering event, either in a fourdrinier machine or in a twin-wire machine, can
improve formation, but this is always accompanied by a reduction of paper strength.
*) In 1999 these were called Duoformer TQ, OptiFormer and BelBaie RCD respectively.

References
1. Parker, J., "The Sheet-Forming Process," STAP No. 9, TAPPI, 1972.
2. Norman, B., "Overview of the physics of forming," 1989 9th Fundamental Research
Symposium Notes, Mechanical Engineers Publication Ltd, London, Vol. 3, p. 73.
Papermaking Part 1, Stock Preparation and Wet End - Page 37

Chapter 6 Web forming

3. Taylor, G.I., "The spectrum of turbulence," Proc. Roy. Soc. Lond. A, 164:476 (1938).
4. Norman, B. and Wahren, D., Svensk Papperstid. 75(20):807 (1972).
5. Meyer, R. and Wahren, D., Svensk Papperstid. 67(10):432 (1964).
6. Soszynski, R. and Kerekes, R., Nordic Pulp Paper Res. J. 3(4):172 (1988).
7. Kerekes, R. and Schell, C., J. Pulp Paper. Sci. 18(1):32 (1992).
8. Thaln, N. and Wahren, D., Svensk Papperstid. 67(7):259 (1964).
9. Wgberg, L., "Adsorption of polyelectrolytes and polymer-induced flocculation of cellulosic
fibers", PhD thesis, Paper Technology, KTH, Stockholm 1987.
10. Beghello, L., "The tendency of fibers to build flocs", PhD thesis, bo Akademi University,
1998.
11. Kerekes, R., "Perspectives on fibre flocculation in papermaking," 1995 International Paper
Physics Conference Proceedings, CPPA, Montreal, p. 23.
12. Kerekes, R., Soszynski, R., Tam Doo, P., "The flocculation of pulp fibres," 1985 8th
Fundamental Research Symposium Notes, Mechanical Engineers Publication Ltd, London, Vol 1,
p. 265.
13. Komppa, A. and Ebeling, K., "Correlation between the areal mass and optical densities of
paper," 1981 7th Fundamental Research Symposium Notes, Mechanical Engineers Publication
Ltd, London, Vol. 2, p. 603.
14. Daunais, R. and Garner, R., "The NUI formation tester an evaluation, modification and
comparison with other techniques," 1987 International Paper Physics Conference Proceedings,
CPPA, Montreal, p. 43.
15. Kallmes, O. and Ayer, J., "Light scanning system provides qualitative formation
measurement," 1987 International Paper Physics Conference Proceedings, CPPA, Montreal p.
209.
16. Ambertec, Beta formation tester, Ambertec OY, Espoo, Finland, 1990.
17. Norman, B. and Wahren, D., Svensk Papperstid. 77(11):397 (1974).
18. Johansson, P.. and Norman, B., "Methods for evaluating formation, print unevenness and
gloss variations developed at STFI," TAPPI 1996 Process and Product Quality Conference
Proceedings, TAPPI PRESS, Atlanta, p. 140.
19. Wrist, P., "Dynamics of sheet formation on the fourdrinier machine," 1961 2nd Fundamental
Research Symposium Notes, Technical Section of the British Paper and Board Makers
Association, London, Vol. 2, p. 839.
20. Haglund, L., Norman, B., Wahren, D., Svensk Papperstid. 77(10):362 (1974).
21. Norman, B., Sjdin, U., Alm, B., Bjrklund, K., Nilsson, F., Pfister, J.-L., "The effect of
localised dewatering on paper formation," 1995 International Paper Physics Conference
Proceedings, CPPA, Montreal, p. 55.
22. Lucisano, M. and Norman, B., "The forming and properties of quasi-random laboratory paper
sheets," 1999 International Paper Physics Conference Proceedings, TAPPI PRESS, Atlanta, p.
331. Submitted to J. Pulp Paper Sci.
23. Hiertner, M., "Basis weight varies faster than on-line systems can measure," 1998 Control
Systems '98 Proceedings, Finnish Society of Automation, Helsinki, p. 307.
24. Ferguson, H., Pulp and Paper 71(10):75 (1997).
Papermaking Part 1, Stock Preparation and Wet End - Page 38

Chapter 6 Web forming

25. Trufitt, A., Design Aspects of Manifold Type Flowspreaders, TAPPI Pulp and Paper
Technology Series, Vol. 32, TAPPI PRESS, Atlanta, 1975.
26. Parker, J. and Hergert, R., Tappi 51(10):425 (1968).
27. Wolf, K. and Egelhof, D., Paper Tech. 15(2):77 (1974).
28. Bubik, A. and Christ, A., "Step diffuser the hydraulics of Escher Wyss headboxes," TAPPI
1976 Engineering Conference Proceedings, TAPPI PRESS, Atlanta, Vol. A, p. 149.
29. Ilmoniemi, E., Aroviita, L., Yli-Kokko, E., Appita 39(1):36 (1986).
30. Sanford, L., "Characteristics of flow from a two-dimensional nozzle with multiple-tube inlet,"
TAPPI 1976 Engineering Conference Proceedings, TAPPI PRESS, Atlanta, Vol. A, p. 159.
31. Parsheh, M., "Nozzle flow," Tech. lic. thesis, Dep. of Mechanics, KTH, Stockholm, 1999.
32. Niskanen, K., "Distribution of fibre orientation in paper," 1989 9th Fundamental Research
Symposium Notes, Mechanical Engineers Publication Ltd, London, Vol. 1, p. 275.
33. Nordstrm, B., "Effects of headbox design and dewatering conditions on twin-wire forming of
TMP," Ph.D. thesis, Pulp and Paper Chemistry and Technology, KTH, Stockholm, 1995.
34. Erkkil, A.-L., Pakarinen, P., Odell, M., "Sheet forming studies using layered orientation
analysis," 1996 CPPA Annual Meeting Notes, CPPA, Montreal.
35. Jansson, M., "Fibre orientation anisotropy variations in the z-direction" (in Swedish), M.Sc.
thesis, Pulp and Paper Chemistry and Technology, KTH, Stockholm, 1998.
36. Sderberg, D. and Alfredsson, H., Eur. J. Mech. B/Fluids 17(5):689 (1998).
37. Aidun, C., TAPPI J. 81(5):159 (1998).
38. Kerekes, R. and Koller, E., Tappi 64(4):104 (1981).
39. Westmeyer, W., Das Papier 41(11):591 (1987).
40. Schultz, H.-J., "The use of standard units for headbox design," 1992 4th International New
Available Techniques and Current Trends Conference Proceedings, SPCI, Stockholm, Vol. 2, p.
6.
41. Begemann, U., Das Papier 47(10A):V149 (1993).
42. Pantaleo, S., Tappi J. 78(11):89 (1995).
43. Nyberg, P. and Malashenko, A., "Dilution control headbox choices, threats and solutions,"
1997 83rd CPPA Annual Meeting Notes, CPPA, Montreal, Vol. A, p. 17.
44. Lloyd, M. and Norman, B., Tappi J. 81(11):195 (1998).
45. Grundstrm, K.-J., Meinander, P.-O., Norman, B., Reiner, L., Waris, T., Tappi J. 59(3):58
(1976).
46. Sandgren, B., "New developments in high consistency forming," 1996 5th International New
Available Techniques Conference Proceedings, SPCI, Stockholm, Vol. 2, p. 616.
47. Norman, B., Mller, K., Ek, R., Duffy, G., "Hydrodynamics of papermaking fibres in water
suspension," 1977 Fundamental Research Symposium Notes, Technical Division, The British
Paper and Board Industry Federation, London, Vol. I, p. 195.
48. Svensson, O. and sterberg, L., Svensk Papperstid. 68(11):403 (1965).
49. Herzig, R. and Johnson, D., "Investigation of thin fibre mats formed at high velocity," TAPPI
1997 Engineering and Papermakers Conference Proceedings, TAPPI PRESS, Atlanta, Vol. 1, p.
109.
Papermaking Part 1, Stock Preparation and Wet End - Page 39

Chapter 6 Web forming

50. Ingmansson, W. and Andrews, B., Tappi 46(3):150 (1963).


51. Radvan, B., Forming the web of paper, The raw materials and processing of papermaking,
Rance, Elsevier, London, 1976, p. 165.
52. Vomhoff, H. and Schmidt, A., Nordic Pulp Paper Res. J. 12(4):267 (1997).
53. Martinez, M., J. Pulp Paper Sci. 24(1):7 (1998).
54. Wahlstrm, B. and OBlenes, G., Pulp Paper Mag. Can. 63(8):T405 (1962).
55. Parker, J. and Epton, J., Tappi 60(10):90 (1977).
56. Norman, B., "Basic theories of twin-wire sheet forming," 1977 XVII EUCEPA Conference
Proceedings, EUCEPA, Paris, Vol. 2, p. 79.
57. Schmidt, W., Wochenbl. Papierfabr. 109(11/12):371 (1981).
58. Kiviranta, A., "Table activity on the fourdrinier: Its characterization and its effect on formation,"
Ph.D. thesis, Laboratory of Paper Technology, HUT, Helsinki, 1993.
59. Neun, J., "Performance of high vacuum dewatering elements in the forming section," TAPPI
1993 Engineering Conference Proceedings, TAPPI PRESS, Atlanta, Vol. 2, p. 1207.
60. Risnen, K., "Water removal by flat boxes and a couch roll on a paper machine wire
section," Ph.D. thesis, Laboratory of Paper Technology, HUT, Helsinki, 1998.
61. Danielsson, R. and Steenberg, B., Svensk Papperstid. 50(13):301 (1947).
62. Holik, H. and Weisshuhn, E., "Influence of headbox flow conditions on paper properties and
their constancy," TAPPI 1987 Engineering Conference Proceedings, TAPPI PRESS, Atlanta, Vol.
2, p. 469.
63. Webster, D., U.S. Pat. No. 3,056,719 (1962).
64. Attwood, B. and Lawrence, L., Paper Tech. 1(5):T195 (1960).
65. Norman, B., Svensk Papperstid. 82(11):330 (1979).
66. Norman, B., Nordic Pulp Paper Res. J. 3(Spec. issue):39 (1987).
67. Nordstrm, B., SCA Research, FEX contract experiments, 1997.
68. Gustavsson, D., "Water removal from the Bel-Baie former," 1975 International Water
Removal Symposium Notes, The British Paper and Board Industry Federation, London, Vol. 1, p.
234.
69. Rding, S. and Norman, B., Tappi J. 69(5):94 (1986).
70. Nordstrm, B. and Norman, B., Nordic Pulp Paper Res. J. 9(3):176 (1994).
71. Brauns, R., "Wet end development," 1986 CPPA 72nd Annual Meeting Notes, CPPA,
Montreal, Vol. A, p. 275.
72. Kerekes, R. and Zhao, R., "Pressure distribution between forming fabrics in blade gap
formers; Thin blades," 1994 CPPA 80th Annual Meeting Notes, CPPA, Montreal, Vol. A, p. A31.
73. Norman, B., Paperi Puu 78(67):376 (1996).
74. Zahrai, S., Bark, F., Norman, B., J. Pulp Paper Sci. 23(9):452 (1997).
75. Malashenko, A., Deperis, G., Lindstrm, R., "Dominion Dynaformer," 1982 CPPA 68th Annual
Meeting Notes, CPPA, Montreal, Vol. A, p. 139.
76. Green, S., Tappi 82(9):136 (1999).
77. Drries, Controllable dewatering patent DE 35 03 242 (1986).
Papermaking Part 1, Stock Preparation and Wet End - Page 40

Chapter

78. Baumann, W.-D., "Duoformer-D a new approach to top wire forming," TAPPI 1988 Annual
Meeting Notes, TAPPI PRESS, Atlanta, p. 75.
79. Reiner, L., Influence of blade position on the burst value of linerboard, Paper Structure and
Properties, K. Bristow, Marcel Dekker, New York, 1986, p. 147.
80. Terland, O., Influence of paper chemicals on paper formation and strength, STFI, Stockholm,
1986.
81. Lee, P. and Lindstrm, T., Nordic Pulp Paper Res. J. 4(2):61 (1989).
82. Fredlund, M., Unpublished summary of FEX investigations, 1996.

Papermaking Part 1, Stock Preparation and Wet End - Page 41

Figure 1. Mobile fibers (top) and a fixed fiber (bottom).

Figure 2. Jacquelin-flocs.

Figure 3. Threshold concentration as a function of fiber slenderness. The


solid line corresponds to Eq. 5 with n = 3. Soszynski and Kerekes 6.

Figure 4. Beta radiographs for two kraft paper samples, softwood kraft, 60
g/m2, twin-wire blade forming. Left: Forming concentration 4 g/L. Right:
Forming concentration 10 g/L.

Figure 5. Local basis weight variations along a newsprint sample using 0.1
mm measurement resolution.

Figure 6. Modified STFI formation spectra for the two samples in Fig. 4.
Solid line: Concentration 4 g/L, F=13.3%, F(0.3-3)=9.3%, F(3-30)=9.5%.
Broken line: Concentration 10 g/L, F=19.1%, F(0.3-3)=9.8%, F(330)=16.4%.

Figure 7. Formation spectra at 60 g/m2 for a laboratory sheet of softwood


fibers, F=10.2%, and the corresponding random sheet, F=14.1%21.

Figure 8. Strength of single- and double-layered laboratory sheets


respectively 21. Laboratory sheets with 50/50 mixture of short/long fibers.

Figure 9. Gray scale picture (original in color) of basis weight of full width
paper web. Newsprint quality, with each level curve indicating a deviation of
1 g/m2(23).

Figure 10. Feed pipe for mix and cross-section of jet from headbox (not to
scale).

Figure 11. CD distributor with main flow Q, discharges D Q, and overflow


q.

Figure 12. Air-cushion headbox with two perforated rolls.

Figure 13. Beloit Converflo headbox nozzle with separation vanes.

Figure 14. Flow pattern in Escher Wyss step diffuser headbox.

Figure 15. Velocity profiles at nozzle outlet for different slice openings d030.

Figure 16. Polar fiber orientation distribution. Left: Isotropic laboratory


sheet. Right: Anisotropic, machine made sheet.

Figure 17. Fiber orientation anisotropy at different levels in a paper sample,


using low (8.5:1, squares) and high (30:1, triangles) headbox nozzle
contraction ratio respectively. FEX twin-wire roll forming experiments, 60
g/m2, hardwood/softwood-mixture, 700 m/min. Minimum velocity
difference between mix and wires 35.

Figure 18. Outflow from a headbox with a sloping top lip and a vertical
slice.

Figure 19. Outflow from a hydraulic headbox.

Figure 20. Transverse flow components at local lip opening reduction.

Figure 21. Radial flow design for CD distribution.

Figure 22. Voith principle for addition of dilution water.

Figure 23. Flow instability with a layered headbox nozzle44.

Figure 24. Headbox for high-concentration forming (STFI).

Figure 25. Dewatering through filtration (left) and thickening (right)1.

Figure 26. Principles for improving the evenness of fiber distribution: (a)
turbulence, (b) elongational flow, (c) dewatering and (d) oriented shear.

Figure 27. 0.2, 4, 7, and 10 g/m2 fiber mats on a triple layer fabric49.

Figure 28. Web concentration as a function of compression pressure52.

Figure 29. Basic characteristics of a fourdrinier paper machine.

Figure 30. Comparison between the suction pulses of a table roll (above)
and of a foil element (below) at 660 m/min19.

Figure 31. Reaction force F per unit width on wire with speed u, generated
by the deflection of a water layer of thickness h at an angle b.

Figure 32. Sheet anisotropy (breaking length) and formation as a function of


jet to wire speed difference48.

Figure 33. The relative mix velocity vector urel as a function of the
difference between jet and wire speeds, uj- uw and cross flow ucross. Left: ujuw> 0, Center: uj- uw= 0, Right: uj- uw < 0.

Figure 34. Basic principle of two-sided dewatering.

Figure 35. Roll former according to Websters U.S. patent63.

Figure 36. Inverform unit for dewatering of a board ply.

Figure 37. Twin-wire roll former with two sided dewatering.

Figure 38. Maximum wire speed u for stable dewatering conditions as a


function of suspension radial thickness h and wire tension T66.

Figure 39. The influence on tensile stiffness anisotropy of the speed


difference between the decelerated mix jet and the wires at different headbox
nozzle contraction ratios. Twin-wire roll forming, 700 m/min, TMP furnish,
FEX trials33. Squares = low contraction, filled circles = medium contraction,
unfilled circles = high contraction.

Figure 40. The influence on small-scale formation of the speed difference


between the decelerated mix jet and the wires at different headbox nozzle
contraction ratios. Forming data and legends as in Fig. 3933.

Figure 41. Fiber orientation anisotropy at different thickness levels of a


paper sample. FEX twin-wire roll forming experiments, 60 g/m2,
hardwood/softwood-mixture, 700 m/min. Headbox contraction ratio 8.5.
Different mix-wire speed differences: 0 m/min (squares), +45 m/min
(circles), -75 m/min (triangles)35.

Figure 42. Black Clawson Vertiforma twin-wire former.

Figure 43. Reaction force F keeping a deflected, tensioned wire in


position56.

Figure 44. Wire movement with deflector blades on one side (top) and on
both sides (bottom). Solid lines: Original wire positions. Broken lines:
Doubled distance between the wires.

Figure 45. Pressure pulses at four consecutive blades in a Bel Baie former.
Basis weight 48.8 g/m2, wire tension 7 kN/m, and speed 920 m/min71.

Figure 46. Pressure distribution along a triangular blade, with a 2.3 degree
top angle, for two different deflection angles around the downstream edge.
FEX experiment, machine speed 1 200 m/min, wire tension 6 kN/m, wire
separation circa 2 mm73.

Figure 47. Dominion Engineering Dynaformer75.

Figure 48. Slotted blade shoe with radius R (broken line) and blade tip
length l = L/2. Left: Blade tip curvature R. Right: Blade tip curvature r =
R/3 66.

Figure 49. Bel Baie III twin-wire former by Beloit.

Figure 50. Valmet Speedformer HS with initial roll forming followed by


forming shoes. 1-12 denotes positions for collection of white water.

Figure 51. Forming part of Voith top-wire former Duoformer-D with


individually adjustable blades78.

Figure 52. The STFI-former: Initial dewatering at constant pressure


followed by adjustable pressure pulses.

Figure 53. Roll-blade, Duoformer CFD forming unit by Voith.

Figure 54. Pressure event along a forming roll of diameter 1 635 mm. Wire
wrapping angle 30 degrees, wire tension 7 kN/m, and machine speed 800
m/min 53. FEX measurements in co-operation with Valmet.

Figure 55. Beloit Bel Bond board forming unit.

Figure 56. Voith Duoformer Top board forming unit.

Anda mungkin juga menyukai