Proceedings of ASME Turbo Expo 2012

GT2012
June 11-15, 2012, Copenhagen, Denmark

GT2012-6

Prediction of Clearance in Industrial Gas Turbine Validated by Field

Operation Data

Bordo Luca Bruzzone Silvia

Ansaldo Energia Ansaldo Energia
Product Development Engineering Product Development Engineering
Via Nicola Lorenzi 8, Genoa, Italy Via Nicola Lorenzi 8, Genoa, Italy

Perrone Andrea Traversone Laura

Ansaldo Energia Ansaldo Energia
Product Development Engineering Product Development Engineering
Via Nicola Lorenzi 8, Genoa, Italy Via Nicola Lorenzi 8, Genoa, Italy

ABSTRACT rotor discs and the connecting tie-rod. Particular care has been
put in the definition of the boundary conditions which are: the
During load operations the gas turbine is characterized by centrifugal loads, the tierod pre-tightening, the thermal
cyclic transients and long dwell times, resulting clearance transient conditions. To take into account 3D effect of airfoil
variation occurs. twisting and bending, the 3D models of the assembly blade-
Tight clearances in both the compressor and the turbine disk are built. The 3D FE displacements of the tip blades are
section is one key to high component efficiencies. However, combined with the 2D FE displacements of the complete rotor
under any operating condition, some clearance must be in the evaluation process. 3D casing FEM model are built to
maintained in order to avoid contact between rotating and calculate the casing distortion due to temperature difference
stationary parts. between cylindrical shell and horizontal flange. The procedure
The optimization of the clearances during the running is one evaluates the axial and radial clearance combining the
of the most relevant key to improve the engine performance, a rotor/stator displacements.
Hydraulic Clearance Optimization (HCO) system is installed The clearance and temperature values are measured and filed
on the Ansaldo Energia gas-turbine models fleet in order to thanks to tip timing capacitive sensors and thermocouples
control turbine blade-tip clearances by appropriate shifting of sensors installed on same stages.
the rotor against the flow direction and consequently reducing The proposed procedure gives results in good agreement with
the radial gaps above the turbine blade tips. the measured clearances in all the transient operation
In particular the proposed procedure takes into account the conditions.
HCO effect on the blade clearances and their optimizing
during the operation conditions. NOMENCLATURE
AEN Ansaldo Energia
The proposed procedure is based on a combination of 2D/3D HCO Hydraulic Clearance Optimization
FEM model of complete rotor-casing assembly and field α duct slope of vane carrier
operation data. U displacement.
The computational approach is based on a 2D axi-symmetric C clearance
finite element model of the engine rotor, including the stacked TVC Turbine Vane Carrier

1 Copyright © 2012 by ASME

CVC1 1st Compressor Vane Carrier The procedure described in this paper allows the clearance
CVC2 2nd Compressor Vane Carrier evaluation in all transient condition and in all relevant position
CVC3 3rd Compressor Vane Carrier of the whole engine (Figure 1):
HC_LE Hot Clearance Leading Edge • radial clearance between tip rotor blade and casing;
HC_TE Hot Clearance Trailing Edge • radial clearance between vane and disks in sealing
CC Assembly Cold Clearance cavity;
Vane Stationary blade • axial clearance between vane and disks in sealing
Blade Rotating blade cavity.

Subscripts
r radial
ax axial
C cold
H hot

Superscripts
cv casing and vane
asu axial sealing upstream
asd axial sealing downstream
rs radial sealing
rt radial tip Figure 1: Clearance calculated positions
HCO Hydraulic Clearance Optimization
When the gas turbine is sufficiently heated, HCO system
works, because during transient operations radial clearances
INTRODUCTION are lower than base load condition. The clearance evaluation
showed in this paper relates to base load condition.
The heavy-duty Ansaldo Energia gas Turbine is made using The computational approach is based on a combination of 2D
the splined-disk rotor concept which features a single central and 3D analysis for the axial and radial clearance calculation
tie-rod assembling the individual bladed disks together by during engine operation. Particular care has been put in the
radial facial serration known as Hirth-teeth. The assembly of definition of the boundary conditions which are: the
the rotor is shown in Figure 2. The basic design principle is centrifugal loads, the pre-tightening load of the tie rod, the
that the central tie-rod without transmitting any torque, holds thermal conditions on all surface of rotor, blades and casing.
the multi-disks rotor tightly together by an adequate Results simulation by FEM have been compared and validated
compressive prestress force. The full torque developed by the with the field operation data in terms of: temperatures,
power turbine is transmitted to the compressor and then to the displacements and clearances.
coupling by the Hirth teeth which connect adjacent disks.

The evaluation of rotor-stator radial and axial clearances

during operation is one of the most important topics in
turbomachinery for heavy duty gas turbine power and
efficiency.
The minimum value of clearance doesn’t always occur during
base load condition, in order to avoid rubbing between rotating
blades and casing clearance calculation must be performed in
all transient conditions.
AEN gas turbines are provided with a Hydraulic Clearance
Optimization system, which reduces the turbine tip blade
clearance in order to increase power and efficiency. The gain
in efficiency and the increase in power output, that are
obtained from HCO system, can be estimated in around
Figure 2: AE94.2 Ansaldo Energia Rotor Assembly
0.2% and 1% respectively.

2 Copyright © 2012 by ASME

HCO DESCRIPTION 2D axisymmetric rotor model

A gain in efficiency and an increase in power output can be The 2D axi-symmetric finite element model is shown in Figure
obtained from Gas Turbines when the rotor is shifted against 4.
the direction of flow. Through this action the gaps between
turbine blades and the casing become narrower, and the gaps
in the compressor are increased accordingly but the sum of the
two effects produce an increased GT efficiency and power
output.
The shifting may be performed only when the GT is
sufficiently heated.
HCO system (Figure 3) is located on compressor bearing and
allows through hydraulic pistons an axial shift of the rotor
against the direction of flow.
Figure 4: FEM model of the rotor and mesh particular

During the generation of the FE model the following

assumptions and simplifications have been applied.

• Axisymmetric parabolic quadratic elements are used

for the description of rotor.
• The pre-tightening load/stress in the tie-rod during
the assembly phase has been simulated by means
special FE elements applied in the tie rod.
• The centrifugal loads generated by the rotating blades
have been simulated by a distributed load applied at
the top of the disc
• The two threads at the ends of the tie-rod have been
modeled by axi-symmetric geometries.
• The coupling between adjacent discs have been
Figure 3: AE94.2 Ansaldo Energia compressor bearing simulated by constraining the displacements in axial
system with HCO direction.
• FE calculation has been performed considering linear
The axial shift of the rotor is the key of HCO design, this value elastic materials behavior
must be optimized in order to avoid rubbing and reduce turbine
clearance. For this reason a detailed FEM simulation is The applied boundary conditions on the rotor are tierod
required. pretightening, centrifugal loads and heat transfer
coefficients with bulk temperature, applied both on the
GAS TURBINE FULL FEM MODEL outside flow path and on the inner surfaces.

All relevant clearances have been evaluated using different 2D axisymmetric and plane stress casing model
FEM models; in particular:
• 2D axisymmetric rotor model The 2D axi-symmetric and plane stress finite element
• 2D axisymmetric and plane stress casing model model is shown in Figure 5.
• 3D casing and blade carrier model
• 3D compressor and turbine blade and disk models
• 3D turbine vane model
All thermo-mechanical calculations have been performed by
Ansys® v11.
The procedure for calculation of component thermal loads for
running clearances of heavy-duty gas turbines is shown in Ref.
2.

3 Copyright © 2012 by ASME

 During the generation of the FE model the following
assumptions and simplifications have been applied:

• Brick parabolic elements are used for the description

of casing and blade carrier.
• In order to simulate the influence of the combustion
chamber on casing deformation it has been simulated
with a simplified approach through shell quadratic
elements.
• The internal parts of combustion chamber (i.e. burner,
Figure 5: Axisymmetric FEM model of assembly casing and
mesh particular dome, flame tube...) have not been modeled, so, in
order to take into account their weight a corrected
During the generation of the FE model the following density has been applied to the shells.
assumptions and simplifications have been applied : • In the thermal analysis on combustion chamber an
• Axisymmetric parabolic quadratic elements are used average temperature has been applied.
for the description of casing. • Compressor and turbine vanes have not been
• Plane stress parabolic quadratic elements are used for modeled, so, in order to take into account vanes
the description of compressor vanes. weight a corrected density has been applied to the
blade carriers.
• The coupling between different casings and vanes
have been simulated by constraining the • The coupling between casing, blade carriers and
displacements in axial and radial direction. combustion chamber have been simulated by
constraining the displacements in all necessary
• FE calculation has been performed considering linear
directions.
elastic materials behavior.
• FE calculation has been performed considering linear
The thermal boundary conditions have been disregarded in elastic materials behavior.
the stationary vanes in 2D model, as the rigorous
evaluation of the clearances will be done, taking into The same boundary conditions have been applied on the
account them in a full 3D model. 3D and 2D models. The gravity load has been considered
The applied boundary conditions on the casing are only in 3D model.
pressure loads and heat transfer coefficients with bulk
temperature on the flow path and on the air extractions. 3D compressor and turbine blade and disk models
As the outer casing is thermally insulated adiabatic
boundary conditions are set in the FEM model. The 3D compressor and turbine blade and disk finite
element model are shown in Figure 7.
3D casing and blade carrier model

The 3D casing and blade carrier finite element model is shown

in Figure 6.

Figure 7: FEM model of compressor and turbine blade

assembled on the disk

During the generation of the FE model the following

assumptions and simplifications have been applied :
Figure 6: FEM model of casing and blade carrier

4 Copyright © 2012 by ASME

 • Brick and tetrahedric parabolic elements are used for
the description of blade and disk.
• The coupling between blade and disk have been
modeled through contact surface to surface elements.
• FE calculation has been performed considering linear
elastic materials behavior

The applied boundary conditions on the blade and disk

assembly are pressure loads, centrifugal load and heat transfer
coefficients with bulk temperature on the blade and on the
disk.

3D turbine vanes models

Figure 9: Computational workflow
The 3D turbine vane finite element model is shown in Figure
8. In Figure 9 the computational workflow is shown; using FEM
models described before and applying all mechanical and
thermal boundary conditions needed displacements in all
positions are obtained.
Then, through an in-house tool, all displacements are
combined in order to obtain the clearance in all needed
locations.
The developed procedure has been automated in order to
update the clearance evaluation against a boundary conditions
modification or a new operating condition. To achieve this goal
the following topics have been implemented:
• all FEM models have been filed in structured folders;
• the displacements of the relevant locations of the
Figure 8: Turbine vane temperature distribution models are evaluated and filed for any operating
conditions (Figure 10).
During the generation of the FE model the following • an APDL macro has been developed to catch and
assumptions and simplifications have been applied : collect the needed displacement in a text file.

• Brick and tetrahedric parabolic elements are used for The previous described procedure has been implemented in a
the description of vane and stator ring. in-house tool that reads text files with displacement and
• The coupling between vane and stator ring has been evaluates the axial and radial clearances in all positions,
simulated by constraining the displacements in all starting from the assembly clearances.
necessary directions. Moreover, in the in-house tool it is possible to introduce an
• FE calculation has been performed considering linear HCO effect in order to define an optimum value for the axial
elastic materials behavior shift. Due to this effect, turbine blade tip clearances get
reduced and compressor blade tip clearances increase, so a
The applied boundary conditions on the vane and stator compromise between these two opposite effects must be found:
ring assembly are pressure loads and heat transfer using the described procedure this optimization is really fast.
coefficients with bulk temperature on the vane and the As shown in Figure 9 a field experimental campaign has been
ring. developed in terms of temperatures measurements, absolute
axial displacements of the casing, relative axial displacement
GENERAL APPROACH DESCRIPTION between casing and rotor and clearance measurements. The
validation of the FEM results is of primary importance to
This paper presents a complete clearance evaluation approach obtain reliable values of the clearances even in the positions
for the whole engine, using, as mentioned before, 2D and 3D where no direct measurements are possible.
models.

5 Copyright © 2012 by ASME

 The above equations describe the radial clearances between
vanes and rotor; displacements required come from 2D rotor
model, 2D casing model, 3D casing model and 3D turbine
vanes model.
The turbine vane carrier tends to ovalize during operation
because of uneven temperature and stiffness distribution in the
circumferential direction. This is mainly due to split-plane
flange area, which is thick in comparison to the rest of the
components. This phenomenon is less pronounced in the
compressor vane carrier where the flange is thinner and the
temperature distribution is uniform.
Moreover, vane turbine radial displacements are not uniform
Figure 10: Components of nodes defined in the in-house
due to a complex temperature distribution (Figure 8), that can
application
only be taken into account using a full 3D vane model.
Due to the considerations above described, the evaluation of
Radial clearance between tip rotor blade and casing
the radial sealing clearances in the compressor and in the
turbine have been carried out with a different approach. For
c H = c Crt + (urcasin g − u rblade ) ⋅ cos(α ) +
rt
compressor clearances a 2D casing and vane model has been
+ (uax
ca sin g
− u ax
rb
+ u axHCO ) ⋅ sin(α ) (1) used, for turbine clearance a complete 3D casing and vane
model has been used in order to take into account the
u =u
rb
ax
blade
ax +u rotor
ax
(2) ovalization and vane temperature distribution.

The above equations describe the radial clearances between Axial clearance between vane and disks in sealing cavity
blade tip and casing; displacements required come from 2D
= cC + (u ax − u ax − uaxHCO )
asu asu cv rotor
rotor model, 3D blade and disk model and 3D casing model. In cH (5)
particular the procedure analyses all displacements of the
casing in all circumferential position, in order to take into
= cC + (−u ax + u ax + u ax
asd asd cv rotor HCO
account the effect of casing ovalization. cH ) (6)
Assembly clearance (evaluated in the condition of minimum
= u ax + u ax
cv ca sin g vane
tolerance) and axial shift due to HCO must be provided by the uax (7)
designer.
The above equations describe the axial clearances between
vanes and rotor; displacements required come from 2D rotor
model, 2D casing model, 3D casing model and 3D turbine
vanes model.
Not uniform temperature distribution of turbine vanes
determinates a rotation in the meridian plane, that can only be
taken into account using a full 3D vane model. This effect is
very important because produces a reduction/growing of axial
clearances in upstream and downstream sealing cavity.

RESULTS ASSESSMENT

Vane carrier ovalization

Figure 11: Scheme of evaluated clearance
Figure 12 shows that the turbine vane carrier tends to ovalize
Radial clearance between vane and disks in sealing cavity during operation due to flange thickness and temperature
distribution. The ovalization determines the increasing or
reduction of clearance in the circumferential direction; in the
= cC + (u rcv − u r
rs rs rotor
cH ) (3)
calculation the minimum value is used.

u rcv = u r + ur
ca sin g vane
(4)

6 Copyright © 2012 by ASME

 The proposed procedure allows to quickly evaluate the
optimum clearance in base load condition in order to
maximize power and efficiency and to avoid blade rubbing.
In the following figures clearance radial blade tip variation in
compressor and turbine sections is shown; the dotted lines
represent the values with HCO effect.

Figure 12: Vane carrier radial displacement

Clearances assessment

The radial clearance evaluation carried out for the compressor

clearance is presented below as an example of the application
Figure 14: Blade tip compressor radial clearances shift due
of the proposed procedure.
to HCO

Using the same value of HCO shift, as shown in Figure 14 and

Figure 15, the variation of clearances is greater in turbine
section (gap decreased) than in compressor one (gap
increased). This effect is due to turbine duct slope, that is much
greater than compressor one.

Figure 13: Blade tip compressor radial clearances

In Figure 13 black line represents assembly nominal clearances

(CC), magenta line represents trailing edge hot clearances
(HC_TE) and light blue line represents leading edge hot
clearances (HC_LE).
Clearance difference between leading and trailing edge of tip
blades is due to: Figure 15: Blade tip turbine radial clearances shift due to
• blade untwist; HCO
• casing deformation;
• rotor axial shift; In the example shown in the previous figures clearances must
be optimized giving special attention to last turbine blade.
• different radial blade elongation on the tip chord.

Clearances optimization with HCO system FIELD VALIDATION

HCO system permits an axial shift of the rotor against flow In order to validate the proposed procedure, it’s been necessary
direction to reduce turbine clearances. During transient compare FEM results to experimental measurements.
Direct clearance measurement has been carried out on 2nd
operations radial clearances are lower than base load
condition, to avoid rubbing this system is used only when the compressor blade; a dedicate geometry sensor has been used in
gas turbine is sufficiently heated. order to fit existing casing holes (factory inspection clearance).

7 Copyright © 2012 by ASME

Clearance measurements in all interested positions are experimental data can be observed; globally the mismatch is
expensive, difficult and time consuming. To obviate this issue, around 10%.
secondary parameters (that control clearance) have been
measured: Clearance
• temperatures;
• axial displacements of rotor and casing. Clearance measurements are carried out in power station, the
The experimental tests are described in detail in Ref. 3. sensor are positioned in the inspection holes of the 2nd
compressor blade (see Figure 17).
Axial rotor and casing displacements Figure 18 shows a satisfactory agreement between calculated
and experimental data; it is interesting to notice the measured
The axial measurements are carried out in power station, the clearance scatter decrement with the increasing of rotation
sensors are positioned near the turbine bearing. speed. This is mainly due to sensor sensitivity, which increases
In the following figure a satisfactory agreement between FEM because of reduced distance between sensor and blade.
results and experimental data can be observed.

Figure 16: Difference between measured and calculated

axial displacement

Temperatures

Figure 18: Difference between measured and calculated

clearance

At nominal speed the mismatch between calculated and

experimental data is around 11%.

CONCLUSIONS

1. The proposed procedure allows a fast evaluation of

whole engine clearances in every location and in all
operating conditions.
2. In order to simulate the full gas turbine model 2D and
Figure 17: Difference between measured and calculated 3D FEM models have been developed.
flow temperature 3. To evaluate the axial and radial clearances in the
relevant locations, displacements resulting from 2D
Temperature measurements are carried out in power station, and 3D FEM models are used.
the thermocouples are positioned in the inspection holes. 4. HCO effect on radial tip clearances can easily be
Figure 17 shows the flow temperature distribution along the taken into account, regarding performance and
compressor length: blue line represents calculated flow rubbing. Due to different compressor and turbine duct
temperature and magenta line represents measured flow slopes a rotor axial shift produces a variation of
temperature. A satisfactory agreement between calculated and clearances, which is greater in turbine section than in
compressor one.

8 Copyright © 2012 by ASME

 5. FEM results showed a significant ovalization in TVC:
this phenomenon determines the increasing or
reduction of radial clearance in the circumferential
direction.
6. The FEM results have been analyzed and compared
with satisfactory agreement with experimental data
coming from tests performed on site, i.e. axial rotor
and casing displacements, temperatures and
clearance.

ACKNOWLEDGMENTS

Authors would like to acknowledge the precious support of

Engineers of PDE team, V. Brizzolara, F. Malavasi and S.
Patrone for experimental section and L. Bozzi for the section
relating thermal calculations. Moreover, authors would like to
show gratitude to HGP Engineers for the data provided.

REFERENCES

[1] Meissonnier, F.T., Stoisser, C.M., “Compressor Rubbing

Risk Analyses for Combustion Turbine Using
Thermomechanical and Dynamical FE Modelling”, ASME
Paper GT2006-90835, Proc. of ASME Turbo Expo 2006. May
8–11, 2006. Barcelona, Spain.
[2] Bozzi, L., Giacobone, L., “Procedure for calculation of
component thermal loads for running clearances of heavy-duty
gas turbines”, ASME Paper GT2012 – 68184, Proc. of ASME
Turbo Expo 2012. June 11–15, 2012. Copenhagen, Denmark.
[3] Aguglia, F., Avanzino, M., Brizzolara, V., Malavasi, F.,
Patrone, S., “Heavy-Duty Gas Turbine “Two Shifting”
Operating Mechanical Integrity Compressor On-Line
Monitoring”, ASME Paper GT2012-69685, Proc. of ASME
Turbo Expo 2012. June 11–15, 2012. Copenhagen, Denmark
[4] Arkhipov, A.N., Karaban, V.V., Putchkov, I.V., Filkorn, G.,
Kieninger, A., “The Whole-Engine Model for Clearance
Evaluation”, ASME Paper GT2009-59259, Proc. of ASME
Turbo Expo 2009. June 8–12, 2009. Orlando, Florida, USA.

9 Copyright © 2012 by ASME