Article

Analysis of chatter marks damage on the Yankee dryer surface of a tissue machine

Center for Industrial Diagnostics, Universitat Politècnica de Catalunya, Av. Diagonal 647, 08028 Barcelona, Spain
Engineering Failure Analysis (Impact Factor: 1.03). 04/2012; 23(23):44-54. DOI: 10.1016/j.engfailanal.2012.02.003

ABSTRACT

a b s t r a c t A tissue machine suffering from Yankee chatter marks has been experimentally investi-gated. A series of vibration measurements during normal operation at various Yankee speeds on both the creping and the cleaning blade holders have been carried out. The anal-ysis in a frequency range up to 20 kHz has permitted to identify speed dependent fre-quency peaks and broadband high frequency vibration content on the creping zone. Hence, an experimental modal analysis of the creping blade and holder has been carried out with the machine stopped to identify its natural frequencies. As a result, resonance conditions have been identified due to the gearbox excitation originated by the meshing process. The study of the corresponding mode shapes has permitted to understand the vibration behavior and its relationship with the damage. To solve the problem, the creping blade holder structure has been redesigned to detune the resonances. Since this overhaul, comparable measurements have confirmed a significant reduction of vibrations and high frequency noise. The appearance of chatter marks has been minimized.

Full-text

Available from: Eduard Egusquiza
(This is a sample cover image for this issue. The actual cover is not yet available at this time.)
This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
Page 1
Author's personal copy
Analysis of chatter marks damage on the Yankee dryer surface
of a tissue machine
Xavier Escaler
, Oscar de la Torre, Eduard Egusquiza
Center for Industrial Diagnostics, Universitat Politècnica de Catalunya, Av. Diagonal 647, 08028 Barcelona, Spain
article info
Article history:
Received 23 December 2011
Accepted 25 February 2012
Available online 8 March 2012
Keywords:
Chatter marks
Yankee dryer
Creping blade
Vibrations
Modal analysis
abstract
A tissue machine suffering from Yankee chatter marks has been experimentally investi-
gated. A series of vibration measurements during normal operation at various Yankee
speeds on both the creping and the cleaning blade holders have been carried out. The anal-
ysis in a frequency range up to 20 kHz has permitted to identify speed dependent fre-
quency peaks and broadband high frequency vibration content on the creping zone.
Hence, an experimental modal analysis of the creping blade and holder has been carried
out with the machine stopped to identify its natural frequencies. As a result, resonance
conditions have been identified due to the gearbox excitation originated by the meshing
process. The study of the corresponding mode shapes has permitted to understand the
vibration behavior and its relationship with the damage. To solve the problem, the creping
blade holder structure has been redesigned to detune the resonances. Since this overhaul,
comparable measurements have confirmed a significant reduction of vibrations and high
frequency noise. The appearance of chatter marks has been minimized.
Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction
The drying section of a tissue machine consists of a large-scale Yankee cylinder internally heated with pressurized steam
that removes the water content of the previously formed and pressed wet fiber web at the end of the paper making line. For
that purpose, the Yankee drier is equipped on one side with a pressure roll that attaches the sheet on its surface and a doctor
blade that removes and crepes it at the opposite side thus creating the tissue (see Fig. 1). In this process it is necessary to
apply an adhesive substance to coat the Yankee surface in order to well attach the sheet and to improve the blade perfor-
mance at the blade–sheet–coating interface. A good coating layer should present a hardness gradient along its thickness with
higher values next to the metal Yankee surface and lower ones next to the sheet [1]. In this way, the Yankee surface is well
protected against blade friction or penetration and at the same time it facilitates adhesion at the pressure roll. In order to
remove contamination build-up on the surface of the Yankee both inside and outside the web, a cleaning blade is also ap-
plied just after the creping and before the coating shower. The blades are supported by aluminum beams, commonly referred
to as blade holders, which in turn are connected to a positioning system.
One potential problem associated to the creping process is the so called blade chatter which can cause defects in the
sheet as well as damage to the Yankee surface. Blade chatter takes place when the tip of the blade moves out-of-plane
due to vibration and a stick–slip motion between its surface and the coating layer develops [2]. If the forces involved are
strong enough, the blade can eventually mark the metallic Yankee surface as shown in Fig. 2. In this case, it is said that
1350-6307/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engfailanal.2012.02.003
Corresponding author.
E-mail address: escaler@mf.upc.edu (X. Escaler).
Engineering Failure Analysis 23 (2012) 44–54
Contents lists available at SciVerse ScienceDirect
Engineering Failure Analysis
journal homepage: www.elsevier.com/locate/engfailanal
Page 2
Author's personal copy
Yankee chatter marks appear. The formation of marks on the Yankee supposes a high cost to the manufacturer because it
must be reground too often.
Chatter has been extensively investigated in the field of machining process involving cutting and/or grinding. It is under-
stood as a very complex vibrational phenomenon that requires the use of various types of sensors such us accelerometers for
its detection in industrial conditions [3]. Moreover, the chatter may cause abnormal wear and friction. In this sense there are
investigations that support the use of acoustic emissions or high frequency vibrations as a diagnostic method to evaluate
lubrication conditions in engineering applications [3–5]. However, few investigations relative to chatter in tissue machines
have been reported up to now. Moreover, there is a lack of published information on the application of vibration based con-
dition monitoring methods to such machines.
The paper machine under study has a Yankee drier made of cast iron that is driven by two electric engines coupled to
a double shafted gearbox. The creping and cleaning blades are made of steel and must be replaced periodically once
their tips are worn. The dimensions of the Yankee cylinder are a face length of 6180 mm and a diameter of
4572 mm. The Yankee lineal speed at its periphery is variable and it can range from 1150 m/min to 2000 m/min depend-
ing on the type of paper and its grade. The machine can produce napkins, kitchen towel and toilet tissue from virgin
pulp.
An experimental investigation has been carried out on this particular machine because it has been suffering from Yankee
chatter marks since its initial start-up. In order to ascertain the possible causes of this problem, the vibration behavior of the
creping zone has been analyzed. For that purpose, a series of measurements had to be carried out during operation and with
the machine stopped.
Fig. 1. Outline of the creping area of a tissue machine.
Fig. 2. Photograph of the chatter marks on the Yankee drier surface (the creping blade appears at the bottom).
X. Escaler et al. / Engineering Failure Analysis 23 (2012) 44–54
45
Page 3
Author's personal copy
2. Experimental setup and measurements
The sensors used for the measurements were high frequency accelerometers, industrial accelerometers, an optical
tachometer and an instrumented impact hammer. The accelerometers were fixed on the creping and cleaning components
with cemented studs. They were orientated in the directions of a cylindrical coordinate system with origin on the Yankee
cylinder axis: tangential (T), radial (R) and axial (A). The optical tachometer was mounted on the Yankee shaft to monitor
its rotation. The impact hammer was used to excite the system.
A digital recorder with 16 simultaneous channels and a signal conditioning unit were used for the data acquisition. Fur-
ther data processing and analysis was carried with the help of LabVIEW and ME’scopeVES softwares. The frequency content
of vibrations was investigated in terms of acceleration and velocity. Three frequency bands were considered ranging from
1 Hz up to 1600 Hz, 5000 Hz and 20,000 Hz.
The vibrations during steady machine operation were measured simultaneously on creping (CR) and cleaning (CL) blades
on operator (OS) and drive (DS) sides. On top of Fig. 3, the positions and orientations of the accelerometers are indicated with
arrows. The high frequency accelerometers (indicated with HF) were only mounted in OS on both holders (see photograph on
the left of Fig. 4 corresponding to OS of creping system). The vibrations were recorded at different Yankee lineal speeds from
a minimum of 1300 to a maximum of 1650 m/min, in steps of 50 m/min.
An experimental modal analysis (EMA) of the cleaning blade and holder system was also carried out with the Yankee
cylinder completely stopped and the blade being applied to its surface. For that, high frequency accelerometers were
mounted in radial direction along the blade tip (see photograph on the right of Fig. 4) and industrial accelerometers, with
a lower bandwidth, were mainly mounted along the blade holder in both radial and tangential directions. The system was
excited by an instrumented impact hammer in different locations and the signals were recorded simultaneously. For the
cleaning system, only a few tests could be carried out. The sensor locations for the EMA are indicated at the bottom of
Fig. 3.
The paper machine typical operating range corresponds to Yankee tangential speeds, V, from around 1300 to 1850 m/
min. Given the Yankee radius R of 2286 mm and the number of tooth of the wheel on the gearbox output shaft Z of 137,
the Yankee rotating frequency (f
n
) can be calculated as well as the gear mesh frequency between the output shaft
(Yankee) and the intermediate shaft, GMF, as described with the following expressions where
x
is the angular speed
in rad/s:
Fig. 3. Measuring positions on creping (a) and cleaning (b) holders. Measuring positions for experimental modal analysis of creping (c) and cleaning (d)
holders.
46 X. Escaler et al. / Engineering Failure Analysis 23 (2012) 44–54
Page 4
Author's personal copy
f
n
¼
x
2
p
¼
V
2
p
R
;
GMF ¼ Zf
n
For example, the values of f
n
and GMF for a minimum speed of 1300 m/min are of about 1.51 Hz and 206.6 Hz respectively;
and for the maximum speed of 1650 m/min are of about 1.92 Hz and 262.2 Hz respectively.
3. Experimental results
The power spectra up to 300 Hz of the gearbox vibration expressed as the root mean square value (RMS) of the vibration
velocity are plotted in Fig. 5 as a waterfall graph for the various operating Yankee speeds. The main frequency peak that ap-
pears on the right hand side of the graphs corresponds to GMF. Its amplitude shows a maximum close to 4 mm/s at 1550 m/
min.
Fig. 4. (Left) photograph of creping holder operator side with the sensors mounted. (Right) photograph of a high frequency accelerometer mounted on the
creping blade during the modal tests.
Fig. 5. RMS vibration velocity power spectra measured in vertical direction on gearbox at various Yankee lineal speeds.
X. Escaler et al. / Engineering Failure Analysis 23 (2012) 44–54
47
Page 5
Author's personal copy
The power spectra of the vibrations measured on the OS of the creping holder are presented in Fig. 6 for radial (top),
tangential (middle) and axial (bottom) directions. The higher frequencies up to 1600 Hz have been omitted because no
significant amplitudes are found. It is clearly observed that the predominant peak at certain Yankee speeds correlates with
Fig. 6. RMS vibration velocity power spectra measured in radial (top), tangential (middle) and radial (bottom) directions on operator side of creping holder
at various Yankee lineal speeds.
48 X. Escaler et al. / Engineering Failure Analysis 23 (2012) 44–54
Page 6
Author's personal copy
the GMF. The maximum amplitude of this frequency is observed in tangential direction with a value of about 1.3 mm/s at
1500 m/min. In radial direction, the maximum amplitude is of about 0.8 mm/s at 1300 m/min. In axial direction, the ampli-
tudes of the GMF peak are significantly lower than in the other two directions. The results for the cleaning holder have also
been omitted since their amplitudes are in general terms lower than in the creping holder. The vibration signature is similar
between both holders with the main vibration occurring at GMF. In axial direction on the cleaning holder the levels are also
lower than in radial and tangential directions.
To verify the common origin of the main frequency peaks, the coherence levels between pairs of signals have been
calculated from the transmissibility functions. The highest coherence levels of about 1 have been found in a wide frequency
range around the GMF peak in all directions in the OS of the creping holder. Similar levels of coherence are also found
between the OS and the DS vibrations. This has permitted to confirm that the origin of these high responses in all directions
in both holders must obviously correspond to the excitation originated during the meshing process in the gearbox which is
transmitted to both holders through the structure.
The analysis of the RMS values of the vibration acceleration in a frequency band up to 5000 Hz also shows that in radial
and tangential directions the frequency peak due to the GMF is the dominant one. It can be seen for example in the waterfall
plot on top of Fig. 7 for radial direction on OS of creping holder. In the DS, the amplitudes are lower than in the OS. In the
cleaning holder, the levels are in general terms also lower than in the creping holder.
The generation of chatter marks is associated to mechanical impacts or increased friction between the blade tip and the
Yankee surface. Based on this assumption, the high frequency content of the vibrations has been measured with miniature
accelerometers in both holders up to 20,000 Hz. The results corresponding to the radial acceleration of vibration on the OS of
the creping holder are shown on the bottom of Fig. 7. These sensors also detect the GMF as well as some excitations at very
Fig. 7. RMS vibration acceleration power spectra measured with the industrial sensor (top) and the high frequency sensor (bottom) in radial direction on
operator side of creping holder at various Yankee lineal speeds.
X. Escaler et al. / Engineering Failure Analysis 23 (2012) 44–54
49
Page 7
Author's personal copy
high frequencies above 12,000 Hz. In general terms, the baselines of the spectra present higher levels in radial than in
tangential directions. Moreover, the amplitudes are higher on the creping holder than on the cleaning one.
A modal analysis has been carried out from the frequency response functions (FRF) calculated between the impact
hammer impulsive excitation and the simultaneous responses measured with the accelerometers. The analysis has been
restricted to a frequency band from 200 to 270 Hz which corresponds to the range where the GMF is expected to appear.
As a result, the main modes of vibration of the creping system including the blade and the holder have been well char-
acterized in terms of natural frequencies and corresponding mode shapes. In particular, three possible natural frequencies
have been detected approximately at 207, 245 and 260 Hz. The first one at 207 Hz corresponds to the 2nd bending mode
of a simple beam fixed on both edges as it can be seen on the top of Fig. 8. The displacement mainly occurs on radial
direction and it affects the blade and the beam simultaneously. Another radial mode shape of higher order, possibly
the 4th one, corresponds to 260 Hz. On the contrary, the mode at 245 Hz shows a significant deformation in tangential
direction as it can be observed on the bottom of the same figure. In this case, due to the radial orientation of the sensors
mounted on the blade tip, the tangential displacement cannot be detected on the blade tip. This mode resembles a 4th
order bending mode.
4. Discussions
The Yankee speeds at which the GMF risks to coincides with the beam natural frequencies are 1300, 1500 and 1650 m/
min since in these cases the GMF are of about 207, 238 and 262 Hz respectively. Therefore, resonance condition can be
achieved between the gear mesh excitation transmitted to the creping holder and its modal response. In fact, the sudden
vibration amplification in certain directions that occurs at those speeds confirms such prediction. The evolution of the radial
and tangential amplitudes of the GMF peak as a function of speed in terms of vibration velocity and acceleration is shown in
Fig. 11 (dashed lines). It can be observed that the highest radial vibrations occur at 1300 and 1650 m/min when the radial
modes are excited, and that maximum tangential vibration levels occur at 1500 m/min when the tangential mode is excited.
At this point it is has to be reminded that there is no lineal relationship between the amplitude of the GMF peak measured
directly on the gearbox and the creping blade amplitudes. For instance, the maximum amplitude of GMF frequency on the
gearbox occurs at a different speed of about 1550 m/min. Consequently, it is confirmed that the vibrations at these speeds
are amplified by the resonance conditions.
Another consideration to take into account is that the EMA has been carried out with the machine stopped. Therefore, the
detected natural frequencies might be slightly shifted when the machine is operating. This should be due to changes in the
boundary conditions of the system caused for instance by the interaction of paper sheet and the existence of different loads
on the mechanical arms that support and position the creping system.
To consider the high frequency vibration contents, the RMS values of the time signals filtered in various frequency bands
have been computed. As an example, for the band from 16,000 to 18,000 Hz, the resulting acceleration broadband levels in
Fig. 8. Snapshots of the animated vibration mode shapes for the natural frequencies at 207 and 245 Hz of the creping holder and blade.
50 X. Escaler et al. / Engineering Failure Analysis 23 (2012) 44–54
Page 8
Author's personal copy
radial direction on OS of creping holder are presented on the bottom of Fig. 7. It can be observed that the highest amplitudes
are found at 1300 and 1650 m/min and that the levels are lower in tangential direction than in radial although they follow
the same trend. From that, it can be concluded that the intensity of the friction between the creping blade and the Yankee is
Fig. 9. RMS vibration velocity power spectra measured in radial (top), tangential (middle) and radial (bottom) directions on operator side of the new
creping holder at various Yankee lineal speeds.
X. Escaler et al. / Engineering Failure Analysis 23 (2012) 44–54
51
Page 9
Author's personal copy
significantly enhanced when the radial modes of vibration are excited by the GMF. On the contrary, the resonance with
dominant tangential deformation does not induce any significant increase of the high frequency content. In this sense, it
is also observed that the high frequency vibrations are always lower in tangential than in radial direction.
Given the existence of resonance conditions inside the operating range of the tissue machine, firstly it was tried to isolate
the cleaning blade from the source of excitation which is the gearbox. For that, the support where the gearbox was attached
was rigidified and reinforced, but this method did not succeed. As a second attempt, it was decided to detune the system by
acting on the creping blade holder. Therefore, a new holder was manufactured with different structural properties. The most
relevant change was an enlargement of the cross section as it can be observed in the outlines of Fig. 13. Since the shape of the
section is almost kept unchanged and the material is brought away from the neutral axis, the second moment of inertia of
the beam is increased. In turn, the beam stiffness is also increased and the natural frequencies are moved away to higher
values.
The comparison of the vibration levels before (old) and after (new) the beam modification confirms the correctness of the
action. The results with the new holder can be observed in Figs. 9 and 10. The GMF peak has almost disappeared or is present
with very low amplitudes. Moreover, the vibration acceleration levels have been significantly reduced in the whole baseband
from low frequencies to the highest ones. Now, the vibrations levels in different directions present similar amplitudes inde-
pendently of the Yankee speed, with the exception of the high frequency content that still indicates a slight difference be-
tween radial and tangential directions. In fact, the natural frequencies have been displaced outside the excitation range of
the gear mesh process and the machine exhibits a normal vibration signature and severity. This is also appreciated in Figs.
11 and 12 where the amplitudes of the GMF peak and of the high frequency broadband are well below the previous ones and
show a more stable trend in reference to Yankee speed.
Fig. 10. RMS vibration acceleration power spectra measured with the industrial sensor (top) and the high frequency sensor (bottom) in radial direction on
operator side of the new creping holder at various Yankee lineal speeds.
52 X. Escaler et al. / Engineering Failure Analysis 23 (2012) 44–54
Page 10
Author's personal copy
Fig. 11. Amplitude of the GMF frequency peak in terms of vibration velocity (top) and acceleration (bottom) in radial and tangential directions on operator
side of the original and the new creping holder at various Yankee lineal speeds.
Fig. 12. Overall RMS vibration acceleration levels in the frequency band from 16,000 to 18,000 Hz in radial and tangential directions on operator side of the
original and the new creping holder at various Yankee lineal speeds.
X. Escaler et al. / Engineering Failure Analysis 23 (2012) 44–54
53
Page 11
Author's personal copy
5. Conclusions
The generation of chatter marks in the Yankee dryer surface of a tissue machine can be enhanced by forced vibrations due
to resonance conditions on the beams that hold the creping/cleaning blades. The gearbox that couples the electric engine
with the Yankee shaft is a source of excitation due to the meshing process. In our case, the monitoring of vibrations on cre-
ping holder in terms of velocity in the low frequency range (up to 1600 Hz) and of acceleration in the medium frequency
range (up to 5000 Hz) during machine operation have shown the gear mesh frequency peak with high levels at certain oper-
ating speeds. Moreover, the high frequency vibration content (up to 20,000 Hz) also has exhibited large amplitudes, which is
an indicator of the friction process between the blade tip and the Yankee surface. An experimental modal analysis carried out
with the machine stopped has permitted to identify several natural frequencies in the range of the GMF. The modes of vibra-
tion responsible of the vibration amplification are those with the main deformations occurring in radial or tangential
directions. Nevertheless, only the radial modes appear to induce high frequency vibrations up to 20,000 Hz. The detuning
of the natural frequencies by changing the cross section of the blade holder has been a reliable solution to solve this problem.
The resulting increase of stiffness has displacement the natural frequencies away from the range of excitation. Consequently,
the forced response of the blade holder has been mitigated and the risk of chatter marks has been reduced.
References
[1] Furman GS, Su W. A review of chemical and physical factors influencing Yankee dryer coatings. Nord Pulp Pap J 1993;08(1):217–22.
[2] Archer S, Grigoriev V, Furman G, Bonday L, Su W. Chatter and soft tissue production process driven mechanisms. In: Tissue World Americas
Conference, Miami, Florida, 11th March; 2008.
[3] Kuljanic E, Sortino M, Totis G. Multisensor approaches for chatter detection in milling. J Sound Vib 2008;312:672–93.
[4] Fan Y, Gu F, Ball A. Modelling acoustic emissions generated by sliding friction. Wear 2010;268:811–5.
[5] Inasaki I. Application of acoustic emission sensor for monitoring machining processes. Ultrasonics 1998;36:273–81.
Fig. 13. Old and new cross sections of the creping holder beam (at the same scale).
54 X. Escaler et al. / Engineering Failure Analysis 23 (2012) 44–54
Page 12
  • Source
    • "In the past few decades, the issue of vibration with chatter marks has also become of great concern in machining processes. Many papers emphasize on vibration with chatter marks as part of metal cutting tool performance [3, 5, 14] and have focused on the mechanism of chatter mark forming. Traditional vibration-based analysis methods of chatter marks involve spectrum and waveform analysis. "
    [Show abstract] [Hide abstract] ABSTRACT: Detecting chatter mark vibration in rolling mills operating under normal working conditions is difficult. A novel characteristic recognition method of chatter mark vibration, where the non-dimensional parameters are calculated with time varying signals and kurtosis under normal rolling mill operating conditions, is presented in this paper. The character of the chatter mark vibration signal is obtained by calculating the kurtosis value of each vibration signal segment obtained by subdividing the raw time varying vibration signal. The probability density function is utilized to reveal obvious differences between signals with respect to the normal and chatter conditions. The method overcomes the limitation of traditional spectrum analysis, which is sensitive to working conditions. Numerical simulation and experimental results show that the proposed method has better recognition capability than traditional spectrum analysis.
    Full-text · Article · Jun 2014 · Journal of Mechanical Science and Technology
  • Source
    • "The issue of vibration with chatter marks in machining processes has elicited a considerable amount of concern in the past few decades. Many studies focused on vibration with chatter marks as part of metal cutting tool performance [3] [5] [14] and on the mechanism of chatter mark forming. Traditional vibration-based analysis methods for chatter marks involve spectrum and waveform analyses. "
    [Show abstract] [Hide abstract] ABSTRACT: Chatter vibration generated during the steel rolling process not only seriously limit the mill performance, but also degrade the surface quality of the steel plate. By analysing the vibration signals from a steel plate mill, Sendzimir 20-high mill, a frequency modulation phenomenon can be observed in the mill chatter vibration signal. In order to monitor and control the chatter vibration during the steel rolling process, an algorithm based on the second order cyclic autocorrelation demodulation is proposed to extract the mill chatter vibration signal. The algorithm is the coupling of frequency analysing, band-pass filtering, second order cyclic autocorrelation calculation and envelope analysing method, and the results of chatter mark vibration can be shown as the waveform in sliced into time domain (SITD) and sliced into frequency domain (SIFD). The experimental results have shown that the modulation frequency is close to the rotational frequency of the intermediate rolls, and it is also shown that mill chatter mark vibration could be evaluated by monitoring the amplitude of the demodulated component whose frequency is close to the rotational frequency of the second intermediate rolls.
    Full-text · Article · Jan 2011 · International Journal of Design Engineering