Journal of System Vol. 2, No.

1, 2008

Design and
Dynamics
Analysis and Measurement of Damping
Characteristics of Integral Shroud Blade for
Steam Turbine*
Yasutomo KANEKO** and Hiroharu OHYAMA***
**
Takasago Research & Development Center, Mitsubishi Heavy Industries, Ltd.,
2-1-1 Shinhama Arai-Cho Takasago, Hyogo, 676-8686, Japan
E-mail: yasutomo_kaneko@mhi.co.jp
***
Takasago Machinery Works, Mitsubishi Heavy Industries, Ltd.,
2-1-1 Shinhama Arai-Cho Takasago, Hyogo, 676-8686, Japan
E-mail: hiroharu_ohyama@mhi.co.jp

Abstract
To improve the reliability and the thermal efficiency of LP Low pressure end blades
of steam turbine, new standard series of LP end blades have been developed. The
new LP end blades are characterized by the ISB (Integral Shroud Blade) structure.
In the ISB structure, blades are continuously coupled by blade untwist due to
centrifugal force when the blades rotate at high speed. This paper, first, explains the
effect of the contact condition of the shroud on the vibration characteristics of the
ISB structure, using the analysis method developed by authors. Second, the
calculated natural frequency and blade damping are compared with the measured
ones. From these results, the effect of the shroud contact condition on the vibration
characteristics of the ISB structure is clarified, and the validity of the analysis
method is confirmed.

Key words: Blade, Vibration, Friction, Damping

1. Introduction
New standard series of LP (Low Pressure) end blades of steam turbine have been
developed to improve the thermal efficiency and the reliability. The new LP end blades are
characterized by the ISB (Integral Shroud Blade) structure(1). In the ISB structure, blades
are continuously coupled by blade untwist due to centrifugal force when the blades rotate at
high speed as shown in Fig. 1. Therefore, the number of the resonant vibration modes can
be reduced by virtue of the vibration characteristics of the circumferentially continuous
blades, and the resonant stress can be decreased due to the additional friction damping
generated at shrouds and stubs. In order to develop the ISB with high reliability, it is
indispensable to predict the natural frequency of the circumferentially continuous blade
structure and the blade damping caused by friction in blade design.

Shroud
Shroud Untwist
Untwist
deformation
deformation

Stub
Stub

Gap
Gap

Friction
Friction
damping
damping

*Received 24 Aug., 2007 (No. 07-0406)

[DOI: 10.1299/jsdd.2.311]

Top
Top view
view of
of blade
blade tip
tip
Fig. 1 ISB structure of steam turbine
311
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics

The vibration characteristics of the ISB structure with friction contact interface should
be evaluated by carrying out the forced response. Analysis of the forced response of the
bladed disk with friction contact interface is usually performed in the frequency domain
owing to the high computational efficiency of this approach. Transformation of the
formulation of the problem into the frequency domain can be carried out by representing
each steady-state displacement as a sum of harmonic components, or even as one harmonic
only(2)(3). This approach is known as harmonic balance method, and widely used in
designing a bladed disk with friction contact interface like the ISB structure. The authors
also developed the analysis methods based on the harmonic balance method(4)(5), and have
applied them for the development of the ISB structure.
Recently, because a LP end blade of steam turbine has increased its length and
decreased its thickness for the purpose of increasing the thermal efficiency, the blade
deformation due to centrifugal force has increased remarkably. Therefore, in the ISB
structure, it is possible that the shroud and the stub do not get in contact with each other in
the whole contact surface as expected in blade design. As a result, it seems that the vibration
characteristics may be different from those expected in blade design due to the difference of
the contact condition.
From this point of view, in this paper, the effect of the contact condition of the shroud
on the vibration characteristics of the ISB structure is studied, using the analysis method
developed by authors. Next, the calculated natural frequency and blade damping are
compared with the measured ones for the world’s largest steam turbine blade of the 3600
rpm 45 inch titanium blade. From these results, the effect of the shroud contact condition on
the vibration characteristics of the ISB structure is clarified. And it is confirmed that the
vibration characteristics of the ISB structure can be evaluated with practical accuracy by
analysis method proposed here.
2. Analysis Method
When the bladed disk system of the ISB structure is rotating in the circumferentially
non-uniform flow, the equation of motion for the bladed disk system shown in Fig. 2, can be
expressed by Eq.(1).

[M T ]{UT } + [K T ]{U T } = {PT } + ∑ {Fl k } (１)

k ,l

i=1 i=0
i=2 i=N-1
Shroud

Blade
kd

Rotation　Ω
Disk

µN * Fl k

(a) Bladed disk system of ISB (b) Mechanical model of damper

Fig. 2 Bladed disk system and damper model of ISB

312
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics
Where, [MT], [KT], {UT}, and {PT} are the mass matrix, stiffness matrix, displacement
vector and excitation force vector of the whole bladed disk system, and can be expressed by
Eq.(2).
[M ] 0 
 [M ] 
[M T ] =   (Cyclic matrix)
 % 

 0 [M ]
 [K 1 ] [K 2 ] [0] " [K 2 ]T 
 
[K 2 ] [K1 ] [K 2 ]
T
 (Cyclic matrix)
[K T ] =  0 [K 2 ]T [K1 ] 
 
 % 
 [K1 ] 


{U T }T {
= {U 0 } {U 1 } ""{U N −1}
T T T
}
{PT }T {
= {P}T cos ωt {P}T cos(ωt − α H )""{P}T cos(ωt − ( N − 1)α H ) } (2)

In Eq.(2), [M] and [K1] are the mass matrix and the stiffness matrix of a substructure
containing one blade, and [K2] is the stiffness matrix determined by the boundary condition
between the adjacent blades, and {Ui} is the displacement vector of the i-th blade. {P} is the
excitation force acting on the blade, ω is the angular frequency, and αH is the phase angle
caused by the rotation of the bladed disk system. Let the angular speed of the rotor be Ω, the
excitation harmonic (engine order) be H, and the number of the blades in the bladed disk
system be N, then ω and αH can be expressed by Eq.(3)
2πH
ω = HΩ, α H = （3）
N
In Eq.(1), Fl { }
k
is the friction force vector between shrouds, and the superscript k
and the subscript l denote the position of the damper element and the direction of slip,
respectively. The damper element of the shroud is modeled by use of the stiffness, kd
and Coulomb friction force, µN * as shown in Fig. 2(b). According to the concept of the
harmonic balance method, the friction force acting on the damper element is expressed as
Eq.(4).
Fl k = Fclk cos(ωt − φlk ) + Fslk sin(ωt − φlk ) (4)

Where, φl is the phase angle between the excitation force and the friction force.
k

When calculating the frequency response, the friction force of the damper element,
µN , the excitation force, {P}, and the excitation frequency, ω are given as known
*

variables. Next, the nonlinear frequency response analysis of Eq.(1) is carried out, using the
modal analysis method and the harmonic balance method. The modal damping of the bladed
disk system without friction contact interface is also given as a known value. A detailed
procedure and the theoretical basis are provided in the reference (5).
Figure 3 shows the procedure of calculating the frequency response of the ISB
structure.

stub
Eigenvalue analysis of shroud/stab Mechanical model of damper element
unconstrained mode (Fig. 2(b))

Frequency response of bladed disk Formulation of damper force

(Damper force is treated as and displacement
unknown excitation force)

Compatibility condition
for damper force and displacement
(Harmonic balance method)

Nonlinear frequency response

Peak frequency and damping

Fig. 3 Procedure of calculating nonlinear frequency response of ISB

313
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics
3. Measurement of Damping Characteristics by Actual Loading Test
To verify the reliability of a newly developed steam turbine blade, it is essential to
measure the vibratory stress under actual operating condition. Because the load and the
rotational speed of the turbine cannot be changed freely in the field test, the vibratory stress
of the blade is measured by use of the test turbine facility where the actual loading test can
be carried out.
Figure 4 shows the system diagram of the test turbine facility. The steam supplied from
the boiler is adjusted to the inlet condition of the actual machine through the pressure
regulator and temperature regulator. The steam is condensed to water by the condenser after
passing through the turbine. The capacity of this test facility is a world leading 400 ton/hour
steam flow rate. Under high load, the operating conditions of the model turbine such as the
flow rate, the condenser vacuum, and the rotational speed are adjusted to those of the actual
machine by use of the water dynamometer connected to the end of the rotor shaft. In the test
of the low load condition, the drive turbine, which is connected to the end of the rotor shaft,
is used to adjust the operating speed of the model turbine. This test facility has been used
effectively to develop the steam turbine blades with high reliability and performance(1).

(a) Whole view of test facility

Desuperheater
Cooling
Silencer tower
MSV
Boiler GV
Water dynamometer
Test
turbine
Boiler feed Drive
pump turbine Cooling water pump

Condenser

Weight Steam flow

tank
Pit Grand Condensate Water flow
condenser pump Sea water Sea water flow

(b) Main system diagram

Fig. 4 Actual loading test facility for steam turbine

As for the measurement of the blade vibration, the non-contact measurement system
using laser sensors is applied(6), in addition to the conventional slip ring system and
telemetry system. In the measurement with the slip ring system or telemetry system, the
strain gages are attached on the blade surface, and the vibration signal measured with strain
gage is transferred from the rotor to the stationary side. In the measurement with the
non-contact measurement system, laser sensors are installed on the casing and the vibration
amplitudes, the vibration modes, and the frequencies of all blades are detected.

314
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics
Figure 5(a) shows the whole view of full-scaled model turbine of the 3600 rpm 45 inch
titanium blade. In the actual loading test, after adjusting the turbine load, the rotor speed is
gradually changed to resonate the blade and to measure the blade frequency, the resonant
stress, and the damping. Figure 5(b) shows the blade with strain gauge.

（a） Whole view of full-scaled model turbine

(b) Test blade with strain gauge

Fig. 5 Full-scaled model turbine of 3600 rpm 45inch titanium blade

4. Calculated and Measured Results

4.1 Calculated Results of Damping Characteristics of ISB structure

Figure 6 shows the FEA model of the 3600 rpm 45 inch titanium blade, and the shroud
contact condition used in the analysis. When calculating the damping characteristics of the
blade, first, the eigenvalue analysis of the freestanding blade was carried out by use of the
commercial FEA code NASTRAN, in which the shroud and the stub do not contact with
adjacent blades. Next, using the natural frequencies and vibration modes up to 20th mode of
the freestanding blade, the frequency response analysis was carried out according to the
procedure in Fig. 3. Modal logarithmic decrement of the freestanding blade is assumed to be
0.01 for all modes, and the friction coefficient between the shrouds and the stubs is assumed
to be 0.5.
Shroud

Shroud contact surface

Stub

Shroud full contact

Shroud upper contact

Shroud side contact

Fig. 6 FEA model and shroud contact condition

315
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics
Figure 7 shows the nonlinear frequency response for the 3rd engine order (H=3) of the
1st vibration mode, where the blade tip is excited with the excitation force of 10 / 3 N
in the x, y, and z directions, respectively. Namely, the resultant excitation force on the blade
tip is 10 N. In the analysis, the reaction force of the shroud and the stub is changed so that
the ratio of the reaction force between the shroud and the stub is same as the design value.
In Fig. 7, the parameter, PSW denotes the amplification factor of the reaction force of the
shroud and the stub to the design value.
0
10

Psw=0.001 Psw=5
-1
10

Amplitude [mm/N]
Psw=0.2
Psw=0.005
-2
Psw=0.05
10

-3
10

Psw=0.01

-4
10
50 100 150 200

Excitation frequency [Hz]

Fig. 7 Frequency response of 1st mode (H=3)

(Change of reaction force of shroud and stub)

Figure 8 shows the peak frequency and the logarithmic decrement of the 1st mode
excited with the 3rd engine order (H=3), which is obtained from the frequency response of
Fig. 7. The abscissa of Fig. 8, ＱSW is the ratio of the total reaction force of the shroud and
the stub to the modal excitation force of the 1st mode with 3 nodal diameters, which is
normalized by the design value. Although the ISB structure with friction contact interface
does not show the characteristic of logarithmic damping, the logarithmic decrement of the
ISB in Fig. 8 was calculated by applying the half power method to the frequency response
curve of Fig. 7, in order to make a comparison with the linear system easily.
180
Peak frequency [Hz]

160

140

120 Free-standing Continuous ring

100

80
0.0001 0.001 0.01 0.1 1 10

Normalized Reaction force/Excitation force, QSW

Normalized blade damping

100

10

0.1
0.0001 0.001 0.01 0.1 1 10

Normalized reaction force/Excitation force, QSW

Fig. 8 Peak frequency and damping of 1st mode (H=3)

316
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics
From Fig. 7 and Fig. 8, it is shown that the larger the reaction force of the shroud and
the stub is, the higher the peak frequency is, and the lower the logarithmic decrement is,
because the shroud and the stub become completely stuck. Conversely, the smaller the
reaction force of the shroud and the stub is, the lower the peak frequency and the
logarithmic decrement are, because the shroud and the stub become completely slipped. The
optimal damping is achieved when the reaction force is between them. In Fig. 8, the
maximum damping is gained aroundＱSW = 0.03. In addition, when the reaction force
becomes extremely small, the vibration characteristic approaches that of the freestanding
blade, because the shroud and the stub become essentially free.
The reaction force between shrouds is determined by the twist back moment of the
blade due to centrifugal force, the initial gap between shrouds at standstill, the contact
surface angle of the shroud, and so on. Therefore, in the mechanical design of the ISB
structure, these design parameters are adjusted to achieve the optimal vibration
characteristics.
As shown in Fig. 8, the most noticeable vibration characteristic of the ISB structure is
the fact that the peak frequency converges to the upper limit essentially beyond ＱSW＝0.5,
while the logarithmic decrement does not converge to the lower limit, unless Ｑ SW
becomes more than 5. This characteristic means that the peak frequency can be predicted by
the linear FEA, in which the shroud and the stub are completely stuck, and the blade
damping several times as large as the blade without friction damping can be gained, if the
design value of ＱSW is set in the range of 0.5 to 1.0. This characteristic is very useful in the
design of the ISB structure.
Figure 9 and Fig. 10 show the nonlinear frequency response for the 3rd engine order
(H=3) of the 1st vibration mode, where the stub is completely coupled at the whole contact
surface by making the reaction force very large. The only reaction force of the shroud is
changed, in order to study the effect of the shroud contact condition on the vibration
characteristics. The excitation force is the same as that of Fig. 7. In the analysis of Fig. 9,
the shroud contacts with the adjacent ones at the whole contact surface as shown in the
figure. The parameter in Fig. 9, PS denotes the ratio of the reaction force of the shroud used
in the analysis and the design value. On the other hand, the frequency response analysis in
Fig. 10 is carried out under the condition that the shroud contacts with the adjacent ones at a
part of contact surface as shown in the figure.

-1
10
Shroud full contact Ps=5.7
Ps=0.006
Ps=0.46
Amplitude [mm/N]

Ps=0.11
-2
10 Ps=0.03

-3
10

-4
10
50 100 150 200

Excitation frequency [Hz]

Fig. 9 Frequency response of 1st mode (H=3)

(Change of shroud reaction force）

317
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics

-1
10
Shroud side contact Ps=5.7
Ps=0.006
Ps=0.46

Ps=0.11

Amplitude [mm/N]
-2
10 Ps=0.03

-3
10

-4
10
50 100 150 200

Excitation frequency [Hz]

Fig. 10 Frequency response of 1st mode (H=3)

(Change of shroud reaction force）

Figure 11 shows the peak frequency and the logarithmic decrement of the 1st mode
excited with the 3rd engine order (H=3), which is obtained from the frequency response of
Fig. 9 and Fig. 10. The abscissa of Fig. 11, ＱS is the ratio of the reaction force of the
shroud to the modal excitation force of the 1st mode with 3 nodal diameters, which is
normalized by the design value. From the results of Fig. 9 to Fig. 11, it can be said that the
difference of the contact area of the shroud has little effect on the peak frequency and the
blade damping for the vibration mode with the small number of the nodal diameters.

160
[Hz]

155
frequency [Hz]
Peakfrequency

150

Ｓｈｒｏｕｄ　full contact
Shroud upper contact
145
Peak

Shroud side contact

140
0.001 0.01 0.1 1 10

Normalized Reaction
Normalized force/Excitation
Reaction force/Excitation force, Qs
force, Qs
damping

100
blade damping

10
Normalized blade
Normalized

1 Shroud full contact

Shroud upper contact
Shroud side contact

0.1
0.001 0.01 0.1 1 10

Normalized Reaction
Normalized force/Excitation
Reaction force/Excitation force, Qs
force, Qs

Fig. 11 Peak frequency and damping of 1st mode (H=3)

318
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics
In order to study the effect of the shroud contact condition on the vibration
characteristics for the vibration mode with the large number of the nodal diameters, the
same analysis is carried out for the vibration mode with 20 nodal diameters, as shown in
Fig. 12 and Fig. 13.
-1
10
Ps=0.006 Shroud full contact Ps=5.7

Amplitude [mm/N]
-2 Ps=0.46
10

Ps=0.11

Ps=0.03
-3
10

-4
10
100 150 200 250 300

Excitation frequency [Hz]

Fig. 12 Frequency response of 1st mode (H=20)

(Change of shroud reaction force）

0
10
Ps=5.7 Shroud side contact

-1 Ps=0.006 Ps=0.46
10
Amplitude [mm/N]

Ps=0.11

-2 Ps=0.03
10

-3
10

-4
10
100 150 200 250 300
Excitation frequency [Hz]

Fig. 13 Frequency response of 1st mode (H=20)

(Change of shroud reaction force）

Figure 14 shows the peak frequency and the logarithmic decrement of the 1st mode
excited with the 20th engine order (H=20), which is obtained from the frequency response
of Fig. 12 and Fig. 13. As shown in Fig. 12 to Fig. 14, the peak frequency and the blade
damping is remarkably different depending on the shroud contact condition for the vibration
mode with the large number of the nodal diameters. Especially, when the shroud contact
with adjacent shrouds at the only side of the contact surface, the peak frequency and the
blade damping remarkably decrease, comparing with the condition of the whole contact.
This is because the larger the number of the nodal diameters of the vibration mode is, the
larger the constrained force between the adjacent shrouds is. As a result, the sensitivity of
the shroud contact condition on the peak frequency becomes larger for the vibration mode
with the large number of the nodal diameters. The sensitivity of the shroud contact
condition on the blade damping is also larger for the vibration mode with the large number
of the nodal diameters, because the friction damping is affected by the frequency difference
between the shroud slip condition and the shroud stuck condition(7).

319
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics
350
Shroud full contact
Shroud upper contact

[Hz]
300 Shroud side contact

frequency [Hz]
frequency
250

200

Peak
Peak
150

100
0.001 0.01 0.1 1 10
Normalized Reaction
Normalized Reactionforce/Excitation
force/Excitation force,force,
Qs Qs
1000
damping
blade damping

100
Normalizedblade

10
Normalized

Shroud full contact

1
Shroud upper contact
Shroud side contact
0.1
0.001 0.01 0.1 1 10

Normalized
Normalized Reactionforce/Excitation
Reaction force/Excitation force,force,
Qs Qs
Fig. 14 Peak frequency and damping of 1st mode (H=20)

4.2 Comparison of Calculated and Measured Results

Figure 15 shows the example of the Campbell diagram and the circumferential
vibration mode resonating with the 3rd engine order measured by the non-contact
measurement method. In Fig. 15, the frequency calculated by the linear FEA is also plotted,
where the shroud and the stub are assumed to be completely stuck. As shown in Fig. 15, the
calculated frequency shows a good agreement with the measured one, and the vibration
mode with the same number of nodal diameters as the excitation engine order is only
excited.

Calculated freq.

Fig. 15 Campbell diagram of 3600 rpm 45 inch titanium

blade measured by non-contact measurement method

320
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics
Figure 16 shows the comparison of the measured and calculated vibration
characteristics of the 3600 rpm 45 inch titanium blade. The abscissa of Fig. 16 is the
normalized excitation force, and the ordinate is the peak frequency and the blade damping
(logarithmic decrement) for the 1st mode resonance with the 3rd engine order (H=3). The
peak frequency and the blade damping are measured with strain gauge in the actual loading
test. In the test, the rotor speed is gradually changed to resonate the blade and to measure
the peak frequency and the damping. The excitation force on the blade is adjusted by
changing the flow rate step by step.
As shown in Fig. 16, the calculated result shows a good agreement with measured one,
and the larger the excitation force is, the larger the blade damping is, while the peak
frequency is kept constant. Under the high loading condition, the blade damping is around 5
times as large as the blade without the friction damping as expected in the blade design.

frequency
1.2 Calculated
Calculated
peak frequency

1.0
Normalized peak

0.8
Measured
Measured
0.6
Normalized

0.4 Continuous ring

Continuous ring Free-standing
Fee-standing
0.2
0.01 0.1 1 10 100 1000

Normalized Excitation
Normalized force
excitation force

100
damping
blade damping

10 Measured
Measured
Normalizedblade
Normalized

1
Calculated
Calculated

0.1
0.01 0.1 1 10 100 1000

Normalized Excitation
Normalized force
excitation force

Fig. 16 Peak frequency and damping of 1st mode (H=3)

5. Conclusion
In this paper, first, the effect of the shroud contact condition on the vibration
characteristic of the ISB structure was studied, by applying the analysis method developed
by authors. Next, the calculated peak frequency and damping of the 3600 rpm 45 inch
titanium blade are compared with those measured in the actual loading test. From these
results, it is concluded that the contact area of the shroud has little effect on the peak
frequency and the blade damping for the vibration mode with the small number of the nodal
diameters. However, it is also confirmed that the contact area of the shroud has the large
effect for the vibration mode with the large number of the nodal diameters. Namely, in these
vibration modes, it is probable that the peak frequency and the blade damping can be

321
Journal of System Vol. 2, No. 1, 2008
Design and
Dynamics
drastically changed if the shroud gets in contact with adjacent shrouds at a part of contact
surface. And also it is confirmed that the peak frequency and the blade damping can be
predicted with practical accuracy by the analysis method proposed here.
It is worth noting that the fretting fatigue strength is remarkably affected by the shroud
contact condition for the vibration mode with the small number of the nodal diameters, even
if the effect on the peak frequency and the blade damping is small. This is because the local
distribution of the vibratory stress around the contact edge of the shroud changes depending
on the shroud contact condition. The fretting fatigue strength should be evaluated by other
proper method(8).

References
(1) Watanabe, E., et al., Development of New Advanced Low Pressure End Blades for High
Efficiency Steam Turbine, JSME International Journal, Series B, Vol. 45, No. 3, 2002,
pp.552-558.
(2) Chen, J.J. and Menq, C.H., Prediction of periodic response of blades having 3D shroud
constraints, ASME 99-GT289, 1999, pp.1-9.
(3) Petrov, E.P., A method for use of cyclic symmetry properties in analysis of nonlinear
multiharmonic vibration of bladed discs, ASME GT-2003-38480, 2003, pp.1-11.
(4) Kaneko, Y., et al., Vibration Analysis of Friction Damper Blade (in Japanese), Transactions
of the Japan Society of Mechanical Engineers, Series C, Vol. 59, No. 559, 1993,
pp.672-677.
(5) Kaneko, Y., et al., Vibration Analysis of Damper Blade by Substructure Synthesis Method
(in Japanese), Transactions of the Japan Society of Mechanical Engineers, Series C, Vol.
60, No. 570 , 1994, pp.399-405.
(6) Kaneko, Y., et al., Development of Non-contact Measurement Method of Turbomachinery
Blade Vibration (in Japanese), Transactions of the Japan Society of Mechanical Engineers,
Series C, Vol. 67, No. 658, 2001, pp.1846-1852.
(7) Griffin J. H., Friction Damping of Resonant Stress in Gas Turbine Engine Airfoils,
Transactions of the ASME, Journal of Engineering for Gas Turbine and Power, Vol.102,
1980, pp.329-333.
(8) Kaneko, Y., et al., Analysis of Fretting Fatigue Strength of Integral Shroud Blade for Steam
Turbine, International Conference on Power Engineering-2007, ICOPE-2007, 2007.

322