See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/327281591

Damage Analysis and Refurbishment of Hot-Gas-Path Components in

Industrial Gas Turbines

Conference Paper · March 2002

CITATIONS READS

0 389

3 authors, including:

Giampiero Antonelli Claudia Rinaldi

Independent consultant Ricerca Sistema Energetico
14 PUBLICATIONS 74 CITATIONS 25 PUBLICATIONS 314 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Giampiero Antonelli on 29 August 2018.

The user has requested enhancement of the downloaded file.

 EPRI Int. Conf. On “Advances in Life Assessment and Optimization of Fossil Power Plants”;
March 11-13, 2002; Orlando, Florida (USA)

DAMAGE ANALYSIS AND REFURBISHMENT OF HOT GAS PATH

COMPONENTS IN INDUSTRIAL GAS TURBINES

G. Antonelli
V. Bicego
C. Rinaldi
CESI SpA
Via Rubattino 54, 20134 Milan, Italy

Abstract

The application of the miniature Small Punch method is presented, as a direct tool to measure
material residual mechanical properties, specifically for the areas of the components where the
largest, most damaging combinations of temperature and stress levels are concentrated. This
method was recently developed up to full industrial use for steam plant components; work is
ongoing to extend the technique to the complex shapes of gas turbine critical parts.

Recently improved methods of blade repair are appearing for components made of the traditional
equiaxed alloys, able to provide mechanical properties in the repaired zones as close as possible
to original values (typical targeted values are 80% of the base material’s value). This is nowadays
becoming common practice of some repairers for blade critical positions where such performance
is required by the high operating stresses. Moreover advanced repair techniques are described,
presently under development for the more critical single crystal Ni based superalloys.

A novel electromagnetic NDE method for Al content measurements in the beta phase bond coat
layer of metallic and ceramic coatings was recently developed. This advanced method can be
utilised as a damage assessment tool for residual life prediction of service exposed blades.

Introduction

Competition and deregulation in the electricity market require striving for the optimisation of
fundamental aspects of power production, such as continuity of operation and availability, in
order to exploit market opportunities and avoid the potentially severe profit penalties for prime
time outages. In combined cycle units the principal causes of forced and scheduled outages are
traceable to the gas turbine hot-section components. Advanced technologies for improving cost
effectiveness of maintenance operations and of service performance of industrial gas turbines are
considered a key factor by providers of maintenance and life management services to the
electrical utilities.

1
 EPRI Int. Conf. On “Advances in Life Assessment and Optimization of Fossil Power Plants”;
March 11-13, 2002; Orlando, Florida (USA)

The critical components of a plant are those for which the cost of replacemennt is the largest, the
on line or off line control is the most difficult, the impact on plant efficiency the largest, and the
danger for avalanche ruptures affecting the nearby parts the largest. The first raw moving blades
in a gas turbine are therefore generally considered the most critical parts. Thermomechanical
fatigue, coating degradation and creep are the main damage processes operating during service.
The rate of developments of such types of damage in the critical components dictate the service
inspection intervals, the possibilty/convenience of refurbishment or replacement actions, and the
whole gas turbine availability.

In this paper, three particular technologies which have been recently developed or are still under
development at the company of the authors are presented, namely

• a special technology, the so called Small Punch (SP) method of testing, originally developed
for characterising aged materials in steel components using miniature sampling taken almost
non invasively, which is presently under exploration for application to hot gas path GT
components;

• some recent progresses in the field of repair by laser cladding, aimed at extending the blade
portion volume suitable to be repaired, and to develop new repair capabilities also for the
newest DS and SC alloys,

• the use of a special electromagnetic tool for assessing the integrity and damage status of the
metallic coatings as well as of the bond coat of TBC.

Material Characterization And Damage Analysis

The Small Punch (SP) miniaturised test technique, firstly developed at MIT for radiation
1
embrittlement studies , is a potent method capable of providing direct values for mechanical
properties of materials in contrast with indirect measurements using X rays, cavitation, ultrasonic
or magnetic techniques. The SP technique utilises a small disk specimen, typically 3 to 10 mm in
diameter and 0.25-0.5 mm in thickness, clamped around its circumference and indented by a
spherical punch up to failure. Monotonic load-displacement records are used to derive estimates
of tensile and fracture toughness parameters and, from tests at several temperatures, of FATT
(Fracture Appearance Transition Temperature). Creep tests at elevated temperatures in inert
environments enable creep rates and rupture strengths to be measured.

During the last decade, the attention of SP researchers, particularly in the US, was focussed on
conventional steam generation plant, the potential problem of temper embrittlement in certain
2
ferritic steels and the assessment of residual plant life . The key problem was how to remove
samples from existing major components, with perhaps 20 or more years service life, in such a
way that plant integrity was not compromised nor were the samples rendered unusable due to
excessive residual stresses arising from the removal process. FAA developed, from an earlier bolt
2
sampler, a small sample removing tool; Ssam-2 is the current version . This employs a liquid
cooled, hemi-spherical, cup-shaped cutter which slowly feeds into the parent material removing a

2
 EPRI Int. Conf. On “Advances in Life Assessment and Optimization of Fossil Power Plants”;
March 11-13, 2002; Orlando, Florida (USA)
part spherical segment of material. Use of the Ssam-2 has spread to Europe both through licensed
sales of the sampler and the provision of sampling services. A spark erosion sampling system was
3
developed by CESI, with the cutting spark erosion electrodes shaped in the form of a ring sector .
Generally, with an allowed depth of 3 mm and a diameter of the sampled zone as large as 30 mm,
about 8 SP specimens, 8 mm in diameter and 0.5 mm thick, can be obtained from each spherical
segment. A typical sampling operation in situ, for getting 20-25 SP samples, takes 2-3 days of
work (during plant outage). Sampling from service-exposed thick components (typically made of
steels) is effectively non-invasive in the sense that material procurement does nor require post-
sampling repair nor reduced operating conditions for subsequent service life.

Material Sample for Small

Punch tests

Figure 1
The Small Punch portable sampler device developed by CESI

Material characterization by SP testing in the sub-creep range

At the authors company a considerable experience has been accumulating in the last years with
the SP method utilised to measure low temperature properties of steels: application on GT blades
superalloys have been started only recently. In fig. 2 an example is given of the capacity of the SP

3
 EPRI Int. Conf. On “Advances in Life Assessment and Optimization of Fossil Power Plants”;
March 11-13, 2002; Orlando, Florida (USA)
method to provide accurate estimates of a material characteristic which, if the normal test
technique had to be used, would be only applicable to retired (and also massive) parts. The
outcome in this figure is actually the result of a special method of analysis which was developed
by CESI during an activity consisting of (a) SP test programmes on running components as well
as on retired parts, in which case the traditional destructive tests were also possible for
comparison, (b) checking the validity of several existing methods of SP data analysis, for several
classes of steels (i.e. generally beyond the materials class considered by the proposers of those
methods), and (c) exploring new routes for the SP data analysis.

700
Yield strength, actual (Mpa)

600
± 5% error in estimate

500

Boiler drum steels:

400 WSB 62
CuNi52Mo
Marrel AM60
300
300 400 500 600 700
Yield strength, predicted (Mpa)

Figure 2
The prediction of yield strength values of different steels, uning a unique correlation of
measurement results from non invasive types of test (SP, hardness and replicas).

Traditionally, tensile properties are estimated from SP quantities taken at appropriate points of SP
4
load-deflection records, and using empirically derived correlations . This method works well for
some materials (nuclear vessel steels), but fairly poor for others (rotor, steam pipe, boiler drum
5,6
steels) . A different approach, based on finite element modelling of SP tensile curves for
derivation of Ramberg-Osgood parameters and values of Kic (the so called “innovative” EPRI SP
method, including video recording for crack initiation detection) provided similar controversial
outcome. Similarly, FATT estimates by the traditional linear correlations to the Tsp values (the
ductile to brittle transition temperature as determined from SP tests) resulted in small errors for
6
CrMoV rotor steels, and large errors (even above 30°C) for CrNi52Mo steels . Therefore in

4
 EPRI Int. Conf. On “Advances in Life Assessment and Optimization of Fossil Power Plants”;
March 11-13, 2002; Orlando, Florida (USA)
order to try to overcome the difficulties in all existing approaches (poor accuracy in the estimated
properties, limited extrapolation to other material classes), a novel method was recently attempted
at CESI. Namely, the component sampled material is subject to SP testing, metallographic
replication (for measurement of ASTM grain size) and Vickers hardness measurement. Estimates
of y, uts, FATT, and Kic are then estimated by multiparametric correlations with the measured
values of Tsp, Hv and dASTM. In the figure 2 the results by this method for the yield strength
estimation for 3 different types of steels used in boiler drums are shown. This SP data analysis
methodology is rather new and experience with other (rotor, pipe) steels is growing, however
current indications suggest that the typical errors are as low as 3% for the estimates of y and
uts, 20% for Kic and 12°C for the FATT. It should be noticed that the method utilises parameters
that come from non invasive test techniques, and that general correlations are used, i.e. not
needing complex “personalisations” for each specific class of alloy.

Current activities to extend the SP method

in high temperature characterizations of Nickel base components

A widely used technique to assess creep damage in high temperature components is to perform
iso-stress creep testing. This approach consists in carrying out creep tests at a stress level
comparable with the service stress, and at several temperatures higher than service. The failure
times are then extrapolated to the service temperature in order to get a creep life prediction.

The geometry of the rotating blades in GT’s is generally complex (also for the presence of
cooling holes) and the amount of material available for post exposure testing is therefore very
small. Moreover the damage accumulation (thermomechanical fatigue, coating degradation and
creep are the most important mechanisms) is larger at the component surface than in the bulk.
The use of SP discs is meant to provide advantages over the normal uniaxial (cylindrical)
specimens, for tensile and creep testing, not only because the material required is limited, but,
above all, the SP specimens removed from surface are representative of the most damaged layer.

In CESI the SP creep technique was firstly applied and validated for low alloy steels in the frame
7
of an European Program . The SP creep results in temperature accelerated iso-load tests have
generally demonstrated a good agreement with the slopes of iso-stress creep rupture curves
derived from uniaxial specimens and with the value of activation energy. More recently SP creep
8
tests have been performed on the IN939 Ni base superalloy, currently employed in first stages
vanes and nozzles in gas turbines. In parallel, in order to validate the SP approach, a creep
resistance curve with cylindrical uniaxial specimens has been obtained, with the purpose of
verifying a possible extension of the SP creep technique to the damage assessment of hot section
components of gas turbines.

Experimental

The execution of SP creep tests requires a special constant load creep machine in wich a
hemisperical punch is compressed onto the disc surface. A ceramic sphere (2.5 mm diameter) is
mounted on the punch and the disc is positioned on a die and clamped around the edge.

5
 EPRI Int. Conf. On “Advances in Life Assessment and Optimization of Fossil Power Plants”;
March 11-13, 2002; Orlando, Florida (USA)
The machine is equipped with a cylindrical three zone zone furnace; a K thermocouple is placed
at the die, as close as possible to the disc; the temperature control is maintained within ±1°C.
An environmental chamber around the punch-disc assembly allows to perform the SP test in an
inert gas atmosphere, avoiding the superposition of a massive oxidation effect on the disc surface;
at the beginning of the test a void pump evacuates the chamber that is subsequently filled with
argon; during the whole test duration argon is fluxed into the chamber.

The deflection measurement system is formed by a couple of capacitance transducers connected

to the load axis, immediately above the punch, out of the environmental chamber. This solution
shows good reproducibility, fast initial setting, and low sensibility against accidental movements
during the loading phase. An acquisition data system allows fast acquisitions both during loading
phase and towards the end of the test. In comparison with the SP creep machine already
developed to perform tests on steels, the SP creep machine for testing superalloys is equipped
with a furnace able to reach T= 1000° C; moreover, the punch-die assembly and the hottest
section of the load axis have been made in ODS PM3030 superalloy.

A crucial element in order to obtain reliable creep informations from SP creep tests is the
definition of a load value FSP equivalent to the stress value of the uniaxial test, i.e. the load FSP
should provide the same failure time of the uniaxial specimen subject to the stress . This
definition should ideally be the same for all materials and size independent: several correlations
between FSP and have been proposed, among them the one based on a membrane stress
3
concept . According to this correlation the FSP / ratio can be defined as follows:

FSP/ = [ 2•t (R + t/2 ) sin ] sin ] (1)

Here t is the SP disk thickness, R the radius of the punch, and sin is a measure of SP deflection
during the main part of the SP test; sin is an empirical parameter, possibly a constant for each
material class where is the angle between the load axis and the normal to the specimen surface
at the points of surface inflection.

The uniaxial creep tests were performed at =200MPa in the same temperature range
T=850÷940°C, aiming to obtain test durations up to about one thousand hours. The SP discs and
the uniaxial cylindrical specimens were obtained from equiaxed casted bars of the Ni superalloy
IN 939 (dia = 12.7 mm, length = 120 mm); as stresses on a SP disc are plane biaxial, the discs
were machined orthogonally to the bar axis; they were then finished with a surface precision
grinding in order to avoid any uneveness at the surfaces.

Results

The deflection-time curves obtained with IN939 SP discs show a shape that is, at least
qualitatively, similar to the uniaxial strain-time curves. An initial deflection is infact produced at
loading (primary stage), the deflection rate then decreases to a steady regime; towards the end of
the SP test the deflection rate rapidly increases (tertiary stage) up to macrocracking and final

6
 EPRI Int. Conf. On “Advances in Life Assessment and Optimization of Fossil Power Plants”;
March 11-13, 2002; Orlando, Florida (USA)
fracture of the disc. Some scatter of the data at the lowest temperatures has been however
observed as shown in figure 3 (time to rupture at 850°C is sligtly higher in comparison with
865°C).

Figure 3
Deflection-time behaviour in SP creep tests.

The uniaxial creep results have been interpolated by a Manson-Haferd type law (linear
dependence between temperature and log (tr), where tr = time to rupture) as shown in fig. 4: here
the SP results have also been reported for comparison. Even if the two sets of data are rather
limited, a similar trend may be observed. An estimation of the apparent activation energy Qc
from the SP creep tests, using typical analytical approaches, provides a value of 540 KJmol-1 .
that is comparable with the value obtained from the uniaxial tests ( 550 KJmol-1).

Repair Of Blade Materials

Conventional Equi-Axed Superalloys

On damaged zones of rotating blades made of conventional Ni-based superalloys it is common

use to repair with the use of IN625 alloy or Nimonic Alloy 263, which are ductile and weldable
solution strengthened alloys. This way to solve the problem of cracking in the heat affected zone

7
 EPRI Int. Conf. On “Advances in Life Assessment and Optimization of Fossil Power Plants”;
March 11-13, 2002; Orlando, Florida (USA)
(due to the presence of the ’ strengthening phase in the blade material) produces repaired regions
with low creep resistance; for this reason the method is applicable to very reduced blade regions,
only tips or unstressed edges of the root.

Figure 4
Comparison between SP creep tests and uniaxial test results for IN939 Nisuperalloy.

To extend repair to more critical regions of the blades the requirement to have a creep resistance
not less than 80% of the base material must be fulfilled, using fillers containing the strengthening
9-11
gamma prime phase . Recently interesting achievements have been obtained, and proprietary
fillers more weldable than the base material have been developed, with reduced amounts of
11,12
gamma-prime forming elements . The advantage of these fillers is that they let to use
traditional welding techniques, so that the repair process is not too expensive.

Among the possible approaches to avoid liquation cracking at the grain boundaries in the heat
affected zone, one is the use of self similar filler material. Both very low heat inputs are necessary
and high temperature preheating (T> ’ solvus) in the region of the component where the repair
process must be applied. Induction heating and a laser cladding process (a very low heat input
13,14
technique) are successfully used .
The system used in the authors’ company is a 1kW Nd:Yag laser interfaced with a 5 axis CNC
system to manage complex geometry; an automatic temperature control system was also

8
 EPRI Int. Conf. On “Advances in Life Assessment and Optimization of Fossil Power Plants”;
March 11-13, 2002; Orlando, Florida (USA)
developed: the temperature signal is continuously measured by a pyrometer or a termocouple and
is sent to a computerized control system which provides on line regulation of an induction heating
furnace to keep the temperature of the zone under repair at the required preheating value.

Claddings of a good quality can be obtained without any crack or oxide inclusion, even inside a
45° V groove with width and height of about 8 mm ( see figure 5). After the cladding process and
a full heat treatment (solution and ageing typical of the IN738 substrate), the hardness values are
comparable with that of the base material (HV10 = 405±12 Kg/mm2) both in the clad and in the
HAZ (heat affected zone) regions (HV10 = 410±9 and 417±5 Kg/mm2 respectively).

OPTICAL MACROGRAPH 400 µm

Figure 5
Sound repair obtained on IN738 alloy by laser cladding with a Nd:YAG laser and self similar
powder delivered by a coaxial nozzle; preheating at high temperature avoids crack occurrence.
Actually the size and distribution of the gamma prime phase is very similar to that of the base
materials both in the HAZ and in the clad, as shown in figure 6.

9
 EPRI Int. Conf. On “Advances in Life Assessment and Optimization of Fossil Power Plants”;
March 11-13, 2002; Orlando, Florida (USA)

This result is obtained with the only gas shielding coming out of the coaxial nozzle. To obtain
sound deposits with larger volumes a further inert gas shielding system should be used.

Figure 6
Gamma prime phase distribution in the heat affected zone (left) and in the weld produced by laser
cladding on IN738 with self similar powder.

Epitaxial Growth By Nd:YAG Laser Cladding On Single Crystal Superalloys

At the moment repair and refurbishment processes seem to be more difficult and still under
development, for new Single Crystal (SC) Nickel base superalloys. For example special final
heat treatments have to be studied, to restore the microstructure of the strengthening gamma
prime phase, avoiding recristallisation processes; these processes can occur also in the base
15,16
material , if it is solution treated in presence of the residual stresses induced by operation
and/or by the repair process itself. On the other hand problems are encountered applying
traditional repair techniques to these materials as the requirement must be fulfilled that the
crystallographic orientation of the repaired region be the same of the component, to obtain
acceptable creep resistance. This means that the repair process must be able to produce an
epitaxial growth with a filler similar in composition to the substrate.
17,18
Several methods are under development both studying special brazing techniques and using
9,19
dedicated laser cladding techniques .

Using a coaxial powder feeding nozzle developed at CESI for the 1 kW Nd:YAG laser, the
possibility to produce epitaxial growth on a CMSX-4 substrate was verified (see figure 7).

10
 EPRI Int. Conf. On “Advances in Life Assessment and Optimization of Fossil Power Plants”;
March 11-13, 2002; Orlando, Florida (USA)

Figure 7
Deposit of CMSX-4 epitaxially grown on a CMSX-4 sample (top left and right metallography);
left center and bottom coloured misorientation maps detected by EBSD before and after HT.
20
Simple models found in the literature let to calculate the solidification rate and the temperature
gradient at the solid-liquid interface of the liquid pool created by the laser beam. If the sample to

11
 EPRI Int. Conf. On “Advances in Life Assessment and Optimization of Fossil Power Plants”;
March 11-13, 2002; Orlando, Florida (USA)
be cladded is big enough it can work as a heat sink and the very large thermal gradient values are
obtained, as required to produce columnar growth during solidification.

Starting from such a model a user friendly software was prepared, suitable to calculate in which
regions of the interface the solidification process will produce the columnar structure typical of
the epitaxial growth in the case of a complex alloy. Results are shown on easily understandable
pictures. With the support of this tool, parameters like the laser/substrate relative speed and
powder feed were correctly adjusted to produce the epitaxial growth by the laser cladding process
in the specific configuration of CESI system. Further experiments let to optimize operating
conditions to avoid the occurrence of small toe cracks in the substrate and find out the powder
feed values required during the growth of deposits, with some millimeters in height. To check
the epitaxial nature of the cladding, disorientation maps were produced in a dedicated scanning
19
electron microscope, with the EBSD technique . An example of such a deposit is shown in
figuue 7, top left. The metallographic analysis shows a very well defined columnar growth,
figure 7 right; the EBSD technique let to check that the cladding orientation differs less than 10°
from that of the substrate (see the disorientation map in left center and bottom of figure 7); only
the top of the cladding is policrystalline, as expected due the high solidification rate here.

As shown in figure 8 the technique can be easily transferred onto the top of a “full size “ sample
(a single crystal CMSX-4 plate, 100x 40x20mm). Epitaxial deposits as high as 2.6mm can be
obtained with the required crystallographic orientation in the first 2.2mm; the width is at the
moment of 2mm.

Figure 8
Epitaxial laser cladding with the Nd:YAG laser on the top of a CMSX-4 single crystal plate.
Further development of the methodology is ongoing, to transfer the process to a diode laser
recently installed; the wavelength of this laser is even better absorbed by materials than the

12
 EPRI Int. Conf. On “Advances in Life Assessment and Optimization of Fossil Power Plants”;
March 11-13, 2002; Orlando, Florida (USA)
Nd:YAG one, so that deposits of larger section and height will be feasible in a few months.

Thickness And ß-Phase Depletion Measurement

Of Protective Coatings Applied On Gas Turbine Blades

In modern machines, the protection of the blades from high-temperature oxidation and corrosion
relies on high-efficiency coatings, whose service degradation is recognised as the main factor that
determines the necessity for blade refurbishment. Because of the high capital cost of a row of
blades, maintenance actions are based on a viable compromise between fully exploiting the useful
life of the coating and keeping the threat of a failure low. Founding this compromise on
quantitative data on coating condition allows sound optimisation of the refurbishment cycles and
early recognition of possible unexpected and accelerated coating deterioration for timely reaction.

The Frequency Scanning Eddy Current system (F-SECT) has been developed at CESI for non-
destructive condition assessment of new and serviced high-temperature coatings applied on the
hot gas-path components of modern gas turbines. The system (figure 9) is capable of providing
quantitative data for the coating thickness and ß-aluminide depletion: this latter being a quantity
strictly related with the expended life of the MCrAlY coatings extensively used as protecting
systems, either alone or as components of thermal barrier and duplex coating systems. In critical
applications, where high sensitivity and stability is required - i.e. MCrAlY coatings-, it is
necessary to carefully position the probe on the sample by means of a specifically designed
manipulator.

The instrument adopts a frequency scanning eddy current technique to characterise a multiple-
layer test material (coating and substrate) on the basis of the small difference in electrical
conductivity and magnetic permeability of the different layers. The data collected by the system
are analysed in terms of a model of the interaction between a plane electromagnetic wave and a
layered material. The analysis simultaneously yields the thickness and the conductivity to
permeability ratio (electromagnetic parameter) characterising each layer.

This model-based analysis allows calibration-free measurement of the thickness of coatings

before being put in service, during quality/acceptance tests on new or refurbished blades., figure
10. On the other hands,condition assessment of service exposed coatings requires knowledge of
the relation between -aluminide depletion and the electromagnetic parameter for the specific
coating-substrate system under examination, which can be readily obtained through destructive
analysis performed on a scrap blade, figs.11 and 12.

13
 EPRI Int. Conf. On “Advances in Life Assessment and Optimization of Fossil Power Plants”;
March 11-13, 2002; Orlando, Florida (USA)

Figure 9
The Frequency Scanning Eddy Current system (F-SECT): electronic unit connected to the
portable computer, and to the probe on the manipulator.

pressure side suction

side

100
distance from platform (mm)

80

60

40

20
-80 -60 -40 -20 0 20 40 60 80
distance from leading edge (mm)

Figure 10
Percentage of residual ß-phase measured on an Amdry995/Udimet520 1° stage blade, serviced
for 15,000 equivalent hours.

14
 EPRI Int. Conf. On “Advances in Life Assessment and Optimization of Fossil Power Plants”;
March 11-13, 2002; Orlando, Florida (USA)

300

250
F-SECT estimates (microns)

200

150

100

50
Equivalent beta thickness of serviced coating
Thickness of new coating
0
0 50 100 150 200 250 300
Destructive measurements (microns)

Figure 11
Measurement of a new coating thickness, and estimation of the -phase thickness of a service
exposed coating

Conclusions

The first results from SP creep tests on the Ni base superalloy IN939 have been presented and
compared with uniaxial creep tests. The shape of the SP deflection-time curves and SP strain-
time curves is qualitatively similar to uniaxial creep curves. Moreover an estimation of the
apparent activation energy Qc from the SP creep tests is in agreement with the Qc value the from
uniaxial tests.

The results since now obtained show therefore an SP creep behaviour comparable with uniaxial
creep. In order that a definite SP load equivalent to uniaxial creep stress may be successfully
proved and standardised, there is however a need for an international, multi-laboratory study. An
SP creep activity is presently underway in the frame of the European Pressure Equipment
21
Research Council , with the aim to check lab comparability on a CrMoV steam turbine rotor
steel. It is felt that more efforts should be done in the field in SP creep modeling, to develop
suitable methodologies of comparability to conventional uniaxial creep quantities, and in the field
of SP creep testing of Ni base alloys (and coatings). To fully validate the SP creep technique for
damage assessment on gas turbine hot section components, post exposure disc testing is also
being planned on first stage rotating blades retired from service.

15
 EPRI Int. Conf. On “Advances in Life Assessment and Optimization of Fossil Power Plants”;
March 11-13, 2002; Orlando, Florida (USA)

140

120

thickness (micron)
100

80

60
Equivalent

40

20

0
0 5000 10000 15000 20000 25000 30000
E O H (o p . h o u rs + 1 0 x s ta rtu p s )

Figure 12
Reduction of the protective coating thickness vs equivalent operating hours (EOH), for the most
critical area on the suction side of an Amdry995/Udimet520 1st stage blade. Coating end of life
corresponds to thickness = 0 mm, from which coating life is estimated about 24,000 EOH.

Among the advanced repair techniques for rotating blades the laser cladding by a Nd:YAG laser
with a coaxial nozzle feeding powder (self similar to the substrate) has been presented, being
useful both for conventional equiaxed superalloys and for the advanced single crystal alloys. On
the conventional alloys this method could be more expensive with respect to traditional welding
techniques using proprietary fillers, developed by companies deeply involved in refurbishment of
gas turbines.

On the other hands, this repair method seems very promising on the advanced single crystal
superalloys. On these materials the technique here described is able to produce epitaxial growth
on the substrate in the z direction, generating a columnar microstructure with disorientation less
than 10-15° with respect to the base materials. The component mass itself creates the heat sink
necessary to obtain the high temperature gradients at the liquid-solid interface required to
generate columnar growth. Actually this technique is entering the world of gas turbine
refurbishment9.

The qualification and quality control of the coated components is generally made by dummy and
sacrificial spot check testing. A novel electromagnetic NDE method for Al content measurements
in the beta phase bond coat layer was recently proposed. The method here presented can be

16
 EPRI Int. Conf. On “Advances in Life Assessment and Optimization of Fossil Power Plants”;
March 11-13, 2002; Orlando, Florida (USA)
utilised as a damage assessment tool for residual life prediction of service exposed blades, under
conditions where cyclic oxidation is the major dominant mechanism of damage. When also
thermomechanical fatigue is important, these two terms provide the basis for the damage kinetics
algorithms in the codes for theoretical estimates of the blades residual lives (Life Management
System).

References

1. M.P. Manahan, A.S. Argon, and O.K. Harling. “The Development of a Miniaturized Disk
Bend Test for the Determination of Postirradiation Mechanical Properties”, Journalof Nuclear
Materials, 103 & 104, North-Holland Publishing Company, 1981, pp. 1545-1550.
2. J. R. Foulds, C,W. Jewitt, L.H. Bisbee, G.A. Whicker, and R. Viswanathan, “Miniature
Sample Removal and Small Punch Testing for In-Service Component FATT”, Proceedings,
Robert I. Jaffee Memorial Symposium on Clean Materials Technology, ASM, 1992, pp.101-
109.
3. V. Bicego, R.D. Lohr. “Mechanical Testing on Miniature Specimens by the Small Punch
Method”, in Proc. Int. Symp. Mat. Ageing and Life Management , Kalpakkam (2000), Eds.
Baldev Raj et al., p.1445.
4. X. Mao, and H. Takahashi. “Development of a Further-Miniaturized Specimen of 3mm
Diameter for TEM Disk ( 3 mm) Small Punch Tests”, Journal of Nuclear Materials,150,
North-Holland, 1987, pp. 42-52.
5. V. Bicego, E. Lucon, and R. Crudeli. “Integrated technologies for life assessment of primary
power plant components”, Proc. of Int. Symp. on Materials ageing and component life
extension, Eds. Bicego, Nitta and Viswanathan, 1995, EMAS, Vol. I, 295-305.
6. V. Bicego, E. Lucon, C. Sampietri. “The Small Punch Technique for Evaluating Quasi Non-
destructively the Mechanical Properties of Steels”, European Conf. on Fracture, Proc. of
ECF12, 14-18 Sep. 1998.
7. F. Dobes, K. Milicka, B. Ule, T. Sustar, V. Bicego, S. Tettamanti, R.H. Kozlowski, J.
Klaput, M.P. Whelan, K. Maile, C. Schwarzkopf. “Miniaturised disk-bend creep test of
heat-resistant steels at elevated temperatures”, Engineering Mechanics, ISSN 1210-2717,
Vol.5, 1998, N°3, 157-160.
8. C. Guardamagna, N. Ricci, S. Tettamanti. “Small punch creep test application to IN939 Ni
superalloy, paper presented at the Symposium on High Temperature Materials for energy
generation”, EUROMAT 2001, Rimini, Italy, 2001.
9. Stefan Krause. “Laser Powder Build-up Welding”, Sulzer Technical review 4/2001.
10. J.T.Stover and D.W. Gandy. “Status of Weld Repair Technology for Ni-based Superalloy
Gas Turbine Blading”, EPRI Report TR-108272, May 1997.
11. D.W. Gandy, G.J. Frederick, A.J.Peterson,R. Viswanathan, J.T. Stover. “Development
of a Laser-based/High Strength Weld Filler Process to extend Repair Limits on IN-738 Gas
Turbine Blades”,, in Welding and Repair Technology for Power Plants. Fourth
International EPRI/RRAC Conference, Marco Island, Florida. June 7-9, 2000.
12. R. Curtis et al. “Development and Implementation of Repair Processes for Refurbishment

17
 EPRI Int. Conf. On “Advances in Life Assessment and Optimization of Fossil Power Plants”;
March 11-13, 2002; Orlando, Florida (USA)
of W501F, Row 1 Turbine Blades”, Proc. of ASME TURBO EXPO 2001 June 4-7, New
Orleans, Louisiana, Materials and Manufacturing - Paper ASME TURBO 2001-GT-0569.
13. F. Goodwater et al., Chromalloy U.S. Patent “ Interactive laser welding at elevated
temperatures of superalloy articles”, Patent WO 95/06540.
14. C. Rinaldi et al. “ Automatic Refurbishment of Gas turbine Components by CO2 Robo-
laser“, Proceedings of the Conference on Joining and Repair of Gas Turbine Components,
Indianapolis, U.S.A., ed. by D. Tillack and R. Thompson, ASM Int., 1997, pp. 109-118.
15. M. Okazaki et al. “Unfavorable Effect of Local Cellular Transformation in
Microstructurally-controlled Ni-base Superalloys subjected to Previous Damage on High
Temperature Fatigue Strength and the Prevention- For Recoating and Refurbishment
Technology”, paper presented at the Symposium on High Temperature Materials for
energy generation, EUROMAT 2001, Rimini, Italy, 2001.
16. M. Konter. “Single Crystal Technologies in Power Generation Equipment”, paper
presented at the Symposium on High Temperature Materials for energy generation,
EUROMAT 2001, Rimini, Italy, 2001.
17. W. M. Miglietti and R.C. Pennefather. “ Microstructure, Mechanical Properties and
Coatability of Diffusion Brazed CMSX-4 Single Crystal”, Proceedings of the Conference
on Joining and Repair of Gas Turbine Components, Indianapolis, U.S.A., ed. by D. Tillack
and R. Thompson, ASM Int., 1997, pp. 61-77.
18. R. W. Broomfield. “ Development of Brazing Techniques for the Joining of Single Crystal
Components”, PARSONS 2000: Advanced Materials for 21st Century Turbines and Power
Plant, Proc. Of the 5th Int. Clarles Parsons Turbine Conference, 3-10 July 2000, Cambridge,
UK, ed. by A. Strang et al., Book 736, IOM communications Ltd, 2000, pp. 741-754.
19. M. Gaumann et al. “ Epitaxial Laser Metal Forming: analysis of microstructure formation”
Mater. Sci. Eng. A271 (1999) pp. 232-241.
20. M. Picasso et al. “ A Simple but Realistic Model for Laser Cladding”, Metall. and Mater.
Trans. B, vol. 25B, 1994, pp. 281-291.
21. EPERC (European Pressure Equipment Research Council (EPERC), Technical Task Force
5, Service Integrity During Operation, EPERC DOC 200009.007/5, 2001.

18

View publication stats