Evaluation and qualification of diffusion braze repair techniques for superalloy gas turbine components

Publisher’s version / Version de l'éditeur: Laboratory Technical Report, 1991-09-09

READ THESE TERMS AND CONDITIONS CAREFULLY BEFORE USING THIS WEBSITE. https://nrc-publications.canada.ca/eng/copyright

Vous avez des questions? Nous pouvons vous aider. Pour communiquer directement avec un auteur, consultez la

première page de la revue dans laquelle son article a été publié afin de trouver ses coordonnées. Si vous n’arrivez pas à les repérer, communiquez avec nous à PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca.

Questions? Contact the NRC Publications Archive team at

PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca. If you wish to email the authors directly, please see the first page of the publication for their contact information.

NRC Publications Archive

Archives des publications du CNRC

For the publisher’s version, please access the DOI link below./ Pour consulter la version de l’éditeur, utilisez le lien DOI ci-dessous.

https://doi.org/10.4224/21271773

Access and use of this website and the material on it are subject to the Terms and Conditions set forth at

Evaluation and qualification of diffusion braze repair techniques for

superalloy gas turbine components

Stoute, P.; Manente, D.; Immarigeon, J.-P.

https://publications-cnrc.canada.ca/fra/droits

L’accès à ce site Web et l’utilisation de son contenu sont assujettis aux conditions présentées dans le site LISEZ CES CONDITIONS ATTENTIVEMENT AVANT D’UTILISER CE SITE WEB.

NRC Publications Record / Notice d'Archives des publications de CNRC: https://nrc-publications.canada.ca/eng/view/object/?id=412c98eb-2f2f-4c1d-b966-89d990c6de08 https://publications-cnrc.canada.ca/fra/voir/objet/?id=412c98eb-2f2f-4c1d-b966-89d990c6de08

1+1

National Research Council Canada Consell national de recherches Canada

taC·CtaC

UNCLASSIFIED UNLIMITED

INSTITUTE FOR AEROSPACE RESEARCH

LABORATORY TECHNICAL REPORT

LTR- ST- 18·39

EVALUATION AND QUALIFICATION OF DIFFUSION BRAZE REPAIR

TECHNIQUES FOR SUPERALLOY GAS TURBINE COMPONENTS

P. STOUTE, D. MANENTE AND J-P. IMMARIGEON

RAPPORT TECHNIQUE DE LABORATOIRE

INSTITUT DE RECHERCHE AEROSPATIAL.E

9 SEPTEMBER 1991 OTTAWA, CANADA

:::_:!::-/.iii

ii)jjj!f!{!f!/j;;! .. :··

ﾷ ｾ ｾｮＧｨＧｾﾷＬ ﾷ ｴ ｾ ﾷＧＡＧ Ｇ Ｂ Ｚ＠:. L,; ｴｾＺｾｴＮ＠ · ·:: · · · ·

. .

. ; : t

_{ＧｪＺｾＡＧｾＺｩ｜ＱｾＧＩＬ［｛}

_Ａ＠

;\'1\\, \"( .: ,

: r·

Arecv

INSTITUTE FOR AEROSPACE RESEARCH

ｾ ｾ＠

INSTITUT DE

irl

t

.J

RECHERCHE AEROSPATIALE 13 Report Pages : _ _ _ _ _ _ _

REPORT

RAPPORT

Rapport L TR-ST -1839 Fig.

Diag. _ _ 2_0 _ _ _ _ _ _{STRUCTURES AND MATERIALS} LABORATORY Date _ _ 9_S__:ep:..._t_em_be:._r_1.:..99::....:1_ For Pour Reference Reference LABORATOIRE DES STRUCTURES ET DES MATERIAUX Presentation at the Canadian Gas Association Symposium on Industrial Application of Gas Turbines Prepared under DND FE 47788 FACCG

and IAR Project JHG 00 in collaboration with Vac-Aero International

L TR-ST-1839

EVALUATION AND QUALIFICATION OF DIFFUSION BRAZE REPAIR TECHNIQUES FOR SUPERALLOY GAS TURBINE COMPONENTS

Submitted by

Presente par W. Wallace

Laboratory Head/Chef de laboratoire

P. Stoute* D. Manente* Author J-P. lmmarigeon Unclassified Unlimited Auteur _ _ _ _ _ _ _ _ _ _ Approved Approuve G.F. Marsters

-Director General

THIS REPORT MAY NOT BE PUBLISHED WHOLLY OR IN PART WITHOUT THE WRITTEN CONSENT OF THE INSTITUTE FOR AEROSPACE RESEARCH

*Vac-Aero International

CE RAPPORT NE DOlT PAS ETRE REPRODUIT, Nl EN ENTlER PARTIE SANS UNE AUTORISATION ECRITE DE L'INSTITUT DE RECHERCHE AEROSPATIALE

. .

. . . . . .

. __ : ,; . ＭＭＭ［ｾＭＺＭ ＭＭｾｾｻｾ＠

- Ｚ ｟ ｾ ﾷ＠ ｾ Ｎ＠

EVALUATION AND QUALIFICATION OF DIFFUSION BRAZE REPAIR TECHNIQUES FOR SUPERALLOY GAS TURBINE COMPONENTS

P. STOUTE·, D. MANENTE·, AND J-P. IMMARIGEON .. VAC-AERO INTERNATIONAL

'137'1 SPEERS ROAD OAKVILLE, ONTARIO

L6L 2X5

INSTITUTE FOR AEROSPACE RESEARCH NATIONAL RESEARCH COUNCIL OF CANADA

OTIAWA, CANADA

CANADIAN GAS ASSOCIATION

SYMPOSIUM ON INDUSTRIAL APPLICATION OF GAS TURBINES

ABSTRACT

BANFF, ALBERTA OCTOBER '16- '18, '199'1

This paper reports on the development of Diffusion Braze Repair (DBR) techniques for service damaged turbine engine nozzle guide vanes (NGVs), made from cobalt and

nickel-base superalloys. A description of the work undertaken by Vac-Aero in

collaboration with the Institute for Aerospace Research of the National Research Council on qualification of the repair techniques for aero engine applications is also provided. This qualification work included bench testing of material test coupons as well as accelerated endurance testing of repaired aero engine parts under simulated service conditions in a burner rig. The results of stress rupture and burner rig tests indicate that Diffusion Braze Repair is an attractive and viable alternative to component replacement.

. ｾ＠ .. ::::. ｾ＠ . ... ..

... .. :·· ···

1.0 INTRODUCTION

. . thermally-induced damage in

Superalloy components in g.as turbine

ｾｮｧｬｮｾｳ＠

Incur n ultimately lead to crackin

service that gradually degrades their

ｳｴｲｵ｣ｴｾｲ｡ｬｬｮｴ･ｧｮｴｹ＠

an.d ca (NGVs) are

･ｸｰ･ｮｳｩｶｾ＠

and rejection of the components f:om

ｳｾｲｾｬ｣･

｟Ｎ＠ ｾｯｺｺｬ･＠

ｧｵｬ､ｾ＠

Ｚｾｾｾｾｳ･ｦｵｬ＠

lives and reduce

to replace but can be braze repa1red w1th1n ilm1ts to exten engine operating costs.

. . . R ir (DBR) techniques have been

At Vac-Aero International, D1ffUs1on Braze epa .

developed in collaboration with the Institute for Aerospace Research

Ｈｉｾ

Ｑ

ｒＩ＠

Ｐ

ｾｾｾ＠

ｎｾｾｮ｡ｬ＠

Research Council (NRC) for cobalt-base as well as

ｮｩ｣ｫ･ｬＭ｢｡ｳｾ＠

supera oy . s. ese

• ' .

1• t' The work 1nvolved extens1ve bench

repa1rs are now offered for serv1ce app 1ca 1ons. .

testing, the results of which were partly reported at the 1987 NRCC Symposium on

Industrial Applications of Gas

ｔｵｲ｢ｩｮ･ｳＨ

Ｑ

ｾ Ｎ＠

In the

ｰｲｾｳ･ｮｴ＠

paper, the latest

､･ｶ･ｬｯｰｭｾｮｴ＠

work on qualification of these repairs, with emphasiS

ｯｾ＠

accelerated endurance test1ng

of repaired aero engine components, is outlined and discussed.

2.0 REPAIR DEVELOPMENTS

In the previously reported work(1_{l, efforts were devoted to. ident.ifying optimum}

brazing parameters for achieving braze joints with good structural 1ntegnty.

Joints brazed with conventional Ni-Cr-B-Si high strength nickel braze alloys (eg.

AMS 4775, AMS 4777, and AMS 4779), utilizing Band Si as

ｭ･ｬｾｩｮｧ＠

ｰｯｩｾｴ＠

suppressants,

are prone to the formation of brittle boride and silicide chains wh1ch prov1de

ｾ＠ ｲ･ｾ､ｹ＠

path

for cracking. Diffusion braze alloys are silicon-free and rely on a post-braz:

､ＱｦｦｾｳＱｯｾ＠

heat

treatment to reduce formation of brittle boride chains in the joint by forc1ng diffUSIOn of boron into the base material. The repairs discussed in this paper utilize diffusion braze alloys .

2.1 Cobalt-base Superalloy Components

In addition to metallographic evaluation of braze joints, the mechanical properties

of the joints were evaluated through stress rupture testing(1). Braze joint strengths of X-40

stress rupture samples brazed with Ni-base diffusion braze alloys were found

ｴｾ＠

?e

excellent. The specimens used were from investment cast bars with a brazed butt JOint at the middle of the gauge section . Stress rupture testing was performed in accordance with base alloy (X-40) specification AMS 5382. The basic requirements are as follows:

Test Temperature Test Load

Minimum Rupture Time Minimum Elongation

1500 ± 3 ° F (815 ± I . 7 ° C)

30,000 psi (207 MPa) 15 hours

6 percent

X-40 stress rupture samples brazed with Ni-base diffusion braze alloy VA-N I had

ｲｵｰｴｾｲ･＠

lives of between 25 and 35 hours and elongation of approximately 15 - 25 percent (1.8·

stress rupture properties greater than the minimum required properties for base X-40).

It should also be noted that these specimens failed outside of the b · · t · h

material. raze JOin , 1n t e base

Complete Diffusion ｾｲ｡ｺ･＠ Repair (DBR) schemes (see Table 1) have been

developed for several unserviceable cobalt-base NGVs, and more specifically for second stage NGVs (NGV2) (P /N 6871 .352) from Allison T56-A-7B (Series II) and first stage NGVs

(NGV1) (P /N ＶＸＹＶＸＹｾＩ＠ from All1son T56-A-14LFE/-15 (Series Ill) engines. Figure 1 shows

ｾｨ･＠ ｣ｯｾｰｯｮ･ｮｴｳ＠ of Interest. Unserviceable components from canadian Forces (CF)

1nventones were used as test articles.

. The ｾ･ｲｩ･ｳ＠ II NGV2 (P /N 6871352) is made from 'X-40 material and is a solid (no 1nternal cooling), uncoated component.

. The series I

if

NGVI (P /N 6896896) is also made from X-40 material but is a hollow

Internally cooled component. This

ｰ｡ｾ＠

was originally Alpak coated but th is coating

ｨ｡ｾ＠

recently ｢･･ｾ＠ ｲ･ｰｬｾ｣･､＠ by AEP-100 (AII1son Electrophoretic coating for Co-base materials)

and part des1gnat1on number has been changed from P /N 6896896 to p /N 23031745.

2.2 Nickel-base Superalloy Components

. The successful repair of super.aii?Y components requires that oxides formed during serv1ce , on .external surfaces and w1th1n cracks , be removed prior to brazing to ensure

proper wett1ng ｡ｾ､＠ flow of the braze material. For cobalt-base superalloy parts, heating

1n a hydrogen-nch atmosphere (hydrogen reduction) is successful in this regard(1l.

ｾｯｾ･ｶ･ｲＬ＠ hydrogen reduction is not effective in reducing the oxides of aluminum and t1tan1um, forn:ed on . . nickel-base superalloy parts in service , owing to their high therm?dynam1c s.tab1ilty. Hydrogen fluoride reduction has been shown capable of

reduc1ng these ox1des from the surfaces of Ni-base parts(2•3l. An Activated Fluoride Oxide

Reduction (AFOR) facility is in full operation at Vac-Aero for pre-braze cleaning of Ni-base superalloy components.

The repair of components made of 713C IN -738 and INCO 792 has been

, I I

Investigated. Figure 2 shows a service exposed nozzle supplied for evaluation of the ａｆｏｾＯｄｂｒ＠ process. This is an unserviceable part made of INCO 792, with excessive crack1ng of the outer shroud and vane airfoil surfaces. The nozzle has been sectioned for metallographic evaluation as shown in Figure 2. The surface condition before and ｾｦｴ･ｲ＠ AFOR/DBR processing is illustrated in Figure 3. The efficiency of the AFOR process

1n removing oxides from crack surfaces is clearly demonstrated in Figure 4. This

｣ｾｭｰ｡ｲ･ｳ＠ an oxidized crack from the component in the as-received condition , Figure 4a, With a similar crack repaired by AFOR/DBR, Figure 4b . Note that the service oxide has been entirely removed by AFOR cleaning and that the braze alloy has subsequently wetted the crack surfaces and filled the crack entirely.

Several other nickel-base (IN-738, IN-718) superalloy parts, shown in Figure 5, are currently being repaired using AFOR cleaning as part of the repair procedure. The AFOR process effectively removes oxides from these parts, allowing the braze alloy to wet the material effectively and to fill cracks and eroded areas.

ｾ＠ｾ＠ .

.

. .

.

.

.

.

.

.

.

ｾ＠ Ｎ Ｚﾷ Ｎ Ｚ ｾ Ｚ ﾷＺＺ ﾷ ＮＺ ﾷ ＺＺＺ＠ ': ｾ＠

ｾＭ ·- .

...

.

. h . . vestigatirn@ tne strength of

There is a concurrent program underway

wh1c IS In St ess rupture and low

AFOR/DBR processed

ｾｲ｡ｺ･＠

joints in nic.kel-base

ｳｾｰｾｲ｡ｬｬｯｹｾ

Ｎ＠ ｷｾｩ｣ｨ＠

closely simulates

cycle fatigue tests are be1ng conducted us1ng a braze JOint desJgn Preliminary data is

the geometry of a repaired crack in

ｾｮ＠

actual

｣ｯｭｰｯｮ･ｮｾ

Ｎ･＠

orted in previous test

encouragingf4_{) and compares favourably w1th the strength levels P}

programs(5

'6'7)

DURABILITY TESTING OF DIFFUSION BRAZE REPAIRED COBALT-BASE

3

.0 (X-40) NOZZLE GUIDE VANES UNDER SIMULATED SERVICE CONDITIONS IN A BURNER RIG

Source substantiation for new aero engine parts requires that they be subjected

to some form of accelerated endurance testing in an engine test ｾ･ｬｬＬ＠ for an assessment

of their durability and reliability , prior to being used in service . This would also

ｾｯｲｭ｡ｬｬｹ＠

be followed by field evaluation on wing prior to part release . ｂｯｴｾ＠ of these options are

expensive and time consuming for repair ｱｵ｡ｬｩｦｩ｣｡ｴｩｾｮ Ｎ＠ An ｡ｬｴ･ｲｮｾｴＱｶ･＠ ｡ｰｰｾ＿｡｣ｨ＠ could be

to test for durability in the laboratory under closely s1mulated serv1ce condJtJons. In order

to assess this possibility, a programme was undertaken to develop ｴ･ｳｾ＠ hardware and

procedures for accelerated component-level durability testing of ｾｇｾｳＬ＠ us1ng a laboratory

burner rig to simulate operating conditions through the appi.JcatJon of therma.l lo.ad

histories representative of component duty cycles in service eng1nes. A ｦｵｾｨ･ｲ＠ _obJect.Jve

was to compare the relative durability of new and DBR'd con:p?nents dunng ng test1ng under identical conditions with the view of optimizing and qual1fy1ng the DBR procedures. Hardware for burner rig testing has been commissioned for two of the components ｦｾｲ＠ which DBR procedures have been developed and the two components have been ng tested. The results of this work are presented below.

3.1 Test Components

The components subjected to rig testing were from Allison T56 Series II and Series Ill engines and included second stage Series II NGVs and first stage Series Ill NGVs. For Series II engines, two new and two DBR'd NGVs (with the same P/N: 6871352) were rig tested. For Series Ill engines , two new NGVs (P/N 23031745- coated with AEP-100) and two repaired NGVs (P /N 6896896 - recoated in-house with Codep B) were rig tested. Because the AEP coating has only been introduced recently, there were only components coated with Alpak available for repair evaluation at the onset of the programme. These components could not be compared against new Alpak-coated components (i .e. of the same part number) since they are no longer available.

3.2 Duty Cycle Analysis

The duty cycle of a gas turbine engine hot section component is defined as the thermal load history applied during service to that component. Burner rig testing of these components requires the simulation and application of thermal load conditions and thermally induced stresses representative of duty cycle in order to induce thermal fatigue damage which is representative of service.

Thermal stresses (loads) are developed in se · b

exposed to thermal transients associated with changes

ｾｾｊ｣･＠

. ecause

ｾｵｩ､･＠

vanes are

These stresses are cyclic in nature and are caused by th engJnl. e tohperatJnl g

ｴ･ｭｰｾｲ｡ｴｵｲ･Ｎ＠

t t. f e eye JC erma expans1on and

con rae Jon o components under mechanical restraints imposed by th ·

The resulting thermal stresses can cause local plastic deformat'

･Ｈｨｾｮｨｧｊｮ･＠

ｾｳｳ･ｮＺ｢ｬｹ

Ｎ＠

damage) ·n th t · ₁ h' h · Jon 1g stra1n fat1gue

. . J . e ma ena w JC 1n turn leads to thermal fatigue cracking. Successful

ｾｵｲ｡｢Ｑｬｩｴｾ＠

test1ng thu.s demands that,. in addition to simulating temperature histories, the

ng test f1xture restra1ns components 1n a manner simulating engine assembly restraints .

Analysis of the. mission profiles for Allison T56 Series 11 and Series Ill engines

reveale? that the cyclic exchange ratio (ratio of major thermal fatigue cycles to flight

ｨｯｾｲｾＩ＠

IS largely dependent .on the user's operating practices. According to Allison, th is

r.atJo

ＨｾＱ＠ ｡ｰｰｲｯｸｊｭＮ｡ｴｾｬｹ＠

0.5 (1.e. one thermal fatigue cycle for every two hours of flying

;1me), .

ｅｾ｣ｨ＠

miSSion has one primary thermal cycle: take-off brings the engine from

｣ｯｬｾ＠

(am?Jent)

ｴ･ｭｰ･ｾ｡ｴｵｲ･＠

to 'hot' (maximum operating temperature- 100% power) and

landing bnng.s the eng1ne back to 'cold'. The burner rig thermal cycle models the service

ｴｾｫ･Ｍｯｦｦｪｬ｡ｮ､Ｑｮｧ＠

thermal cycle (cold-hot-cold) by cycling the component between a cold a1r stream and the hot burner rig flame.

. The current.'time between overhaul' (TBO) for Canadian Forces (CF) T56 engines

1s 5400 hours wh1ch, based on the above, translates to approximately 2700 service

cyc!es. ａｾ＠ overhaul, over 60% of the Series II second stage NGVs and over 80% of the

Senes Ill ｦＱｲｾｴ＠ stage NGV.s are scrapped (i.e. damage exceeds serviceable limits)(9_{l. Also,}

CF and Un1ted States A1r Force (USAF) experience indicates that replacement of NGVs takes place in the field between overhauls, and that the cracking of NGVs at 2000-2500

hours (since previous overhaul) is considered normal(10l. However, these components

｡ｲｾ＠ n?t tracked in service and thus have unknown TSN's (Time Since New). Thus, the

reJection rate at overhaul for 'new' NGVs (with TSN's of 5400 hours) is not known . Based on the above, it may be assumed that burner rig testing, of new and DBR NGVs, to 2500 cycles should approximate one TBO (5400 hours).

The average maximum turbine inlet temperature (TIT) during service is approximately 971 o C (1780 oF) for Series II engines and 1076 o C (1970 oF) for Series Ill

engines(11l. Figure 6 shows airfoil temperature distribution and individual TIT readings on

Allison T56-A-15 (Series Ill) 1st stage NGVs and indicates that hot spots exist where local peak gas temperature may be 100 o C higher than the average maximum TIT.

Analytical mission profiles for Series II engines prepared by Allison suggests that

the inlet temperature to the 2nd stage NGV is 134 o C below the TIT(12) . It is assumed that,

locally, the peak TIT may be 100°C higher than the average maximum TIT of 971 oc which would correspond to a peak local NGV2 inlet temperature of 937 o C.

The IAR burner rig flame can simulate the flame temperature distribution in the engine (as explained later in Section 3.4). However, the rig cannot match the velocity, pressure, and mass flow experienced by the engines of interest. NGVs are exposed in service to a high pressure hot gas stream, of close to ten (1 0) atmospheres, wh ile test components in the rig are at atmospheric pressure. Gas velocity and mass flow in the

.··.

. · t 1 2 and 12 times greater respectively, than in the burner rig . Gas

engine are

｡､ｰｰｲｯｬｸｬｾｴ｡＠

eo{

mass flow are factors

ｾｨｩ｣ｨ＠

influence the rate of accumulation

pressure an ve oc1 Y · ·11 b t d · ·1

· mponents Thermal fatigue crack1ng WI e grea er, un er s1m1 ar

of damage In cod't'ons ·,n a. higher velocity hot gas stream on account of the greater

temperature con 1 1 , · kl · ·

transient thermal stresses generated (i .e. components heat up

ｭｾｲ･＠

qu1c y 1n

ｴｨｾ＠

ｨＱｾｨ･ｲ＠

I ·t stream) Similarly environmentally enhanced crack1ng (through ox1dat1on)

ve oc1 y gas . , . . .

may be expected to be more pronounced at a ｨｩｧｨｾｲ＠ ｰｲ･ｳｾｵｲ｢･＠ . . Howev.er, Itt 1s poss1ble

to compensate for the lower velocity and pressure 1n the ng y 1ncreas1ng emperature relative to the engine.

Based on the above, it was decided that accelerated burner rig ､ｵｲ｡｢ｾｬｩｴｹ＠ testing

of T56 series 11 NGV2 and Series Ill NGVI would be conducted at burner ng exhaust

temperatures of 1ooo·c and IIOO·C, respectively. Series Ill 1st stage NGVs are

internally cooled by 2.5% of total engine airflow which corresponds to. 0.0135 ｬ｢ｾｊｾ･｣＠ per

airfoil(1_{3l . The internal cooling air is at approximately 650 ·C. To s1mulate th1s Internal}

cooling effect during burner rig testing, room ｴ･ｭｰ･ｲｾｴｵｲ･＠ ｾＳＰﾰｃＩ＠ air ｶＮＺ｡ｾ＠ ｾｵｰｰｬｩ･､＠ at a

rate of 0.0022 lbnJsec per airfoil. This was done w1th a v1ew to ｭ｡ｾＱｭＱｺＱｮｧ＠ thermally

induced stresses and thereby accelerating fatigue damage accumulation .

3.3 Development of Test Hardware

The core of IAR's burner rig facility is a Becan LCS-4A laboratory combustor. The combustor burns aviation Kerosene (JP-4, wide cut gasoline type) to generate a high velocity, high temperature exhaust flame . A thermocouple placed in the flame (the exhaust control thermocouple) allows the flame temperature to be monitored and controlled (by means of a microprocessor-base control system). The combustor variables (e.g. fuel mass flow, air mass flow, primary and secondary combustor air pressures, etc.) are monitored and controlled on a console (the control console) in the burner rig control room.

Extensive development of test hardware (e.g . rig exhaust nozzles, component fixturing, cooling air piping , etc.) was required to allow for component-level durability testing to take place, as detailed below.

Burner rig exhaust nozzles were designed to provide a flame temperature distribution at the nozzle exit area which simulates the service temperature distributions acting on T56 Series II 2nd stage NGVs and Series Ill first stage NGVs. Figure 7 shows

the can-type burner exhaust nozzles, manufactured from Hastelloy

x

material, along with

an actual T56 combustor can nozzle which was the basis of the designs.

Engine assembly re.straints acting on Series 11 second stage and series Ill first

stage NGVs v.:ere determ.lned from the engine Training Manuals(11 ) and the engine

assembly draw1ngs . Test fixtures were designed, with critical dimensions and tolerances

take.n directly from the. drawings, ｴｾ＠ simulate a complete turbine stage assembly. A

sect1on through the Senes II NGV2 f1xture is shown in Figure 8.

, Each burner rig test (BRT) fixture holds a complete set of NGVs with two 'test

zones located 120 ° apart as shown in Figure 9. The BRT fixture rotates 120 o about its

｣･ｮｾｾ｡ｬ＠

shaft, by

ｭ･｡ｮｾ＠

ｾｦ＠

a pneumatic actuator, between two test positions. In test

pos1t1on I, ｾ･ｳｴ＠ ｾｯｮ･＠ I 1s 1n the burner exhaust flame and test zone 2 is cooled by one

of two cooling a1r

ｳｴｲ･｡ｭｾ＠

located at ± 120° relative to the burner exhaust nozzle, Figure

9. At th.e

ｾｮ､＠

.of the heat1ng cycle, the BRT fixture is rotated to test position 2 where test

zone 2 1s hot and test zone 1 is 'cold'.

The s.eries II second stage BRT fixture holds twelve NGVs; one test NGV in test

zone 1, ?ne 1n ｴ･ｳｾ＠ zone 2, and 10 ｾｬ｡ｶ･＠ NGVs (Figure 8). The Series Ill first stage fixture

holds th1rty ｎｇｖｳｾ＠ two test NGVs 1n test zone 1, two in test zone 2, and twenty-six slave

vane segments (Figure 9). Both fixtures were manufactured from type 304 stainless steel. The test f!xture .is m.ounted on a stand in front of the burner rig exhaust nozzle as

ｳｨｯｾｮ＠ schematically 1n F1gure 10. Cyclic rotation of the fixture allows cyclic thermal

test1ng of two batches of components. In this way the durability of new and repaired parts can be compared under identical test conditions. The position of the fixture in relation to the rig exhaust nozzle and the two symmetrical cooling nozzles can be seen in Figure 11 .

Internal cooling air requirements for BRT of Series Ill first stage NGVs led to the development of 'cooling air units', made from Hastelloy X, to supply a regulated airflow through the component from the inner shroud cooling passage.

3.4 Development of Test Procedures

Extensive preliminary testing was performed on scrapped components to establish the surface temperature response to a burner rig cycle. This was done by spot welding fine-wire sheathed type thermocouples to the surface of components as indicated schematically in Figure 12. On the basis of these tests, it was decided that the BRT cycle would be of four minutes duration (2 minutes heating and 2 minutes cooling) since it was found that steady state maximum and minimum component surface temperatures were reached after two minutes heating and cooling, respectively (see Figure 12). It should be reiterated that thermal fatigue damage occurs primarily during the heating and cooling portions of a service thermal cycle. Since a service thermal cycle is of the order of 2 hours and the burner rig cycle of 4 minutes duration, testing in the burner rig allows damage to accumulate 30 times faster than in service .

It should be stated that component surface temperatures measured by the wire thermocouples are not 'true' surface temperatures. This is because the temperature readings are not corrected for the influence of radiation to the cold surroundings, conduction along the thermocouple wire, and disturbance caused by the thermocouples to the boundary layer that conditions convection from the gas stream to the component. True surface temperatures may be expected to differ by tens of degrees from the

measured temperatures(14_{l. However, no efforts were made to ｣｡ｬ｣ｵｬ｡ｴｾ Ｎ＠ these}

temperatures since it was more important in this programme to ensure repeatability and this was achieved by controlling flame temperature.

Flame temperature distributions at burner rig exhaust control th.ermocouple settings

of ₁₀₀₀ o

c

(1832 oF) and 11

oo

o

c

(2012 oF), respectively, were ｾｳｴ｡｢ｾＱｳｨ･､＠ f?r both Series

11 and Series Ill can-type nozzles by means of a thermocouple rake (see

ｆＱｧｾｲｾ＠

1.3) with

the thermocouples lying horizontally in the plane of the . burner flame to .minimize the

effects of gas flow disturbance around the bead (convection) and ｣＿ｮ､ｵ｣ｴＱｯｾ＠ along the

thermocouples. The variations in temperature across

ｴｾ･＠

ｾｯｺｾｬ･＠

ex1t.

｡ｲｾ｡＠

(Figures 14a

and 14b) are consistent with engine gas temperature dJ.stnbutJon which 1s known to be quite non-uniform across any one combustor can (see F1gure 6). The flame temperature distribution across the BRT nozzle was found to be stable and repeatable.

The airflow settings for internal cooling of Series Ill first stage NGVs during rig testing were determined through calibration with an airflow gauge. Internal cooling was provided during both the heating and cooling portions of the burner rig cycle. Component surface temperature response to a BRT heating cycle with internal cooling was investigated visually and by qualitative readings with spot-welded fine wire sheathed thermocouples (see Figure 12); there was a noticeable cooling effect which was later found to be insufficient as discussed in see Section 3.5.2.

The burner rig facility control systems were set up and programmed to

automatically monitor and control system variables during rig testing. Component

inspection (visual, stereo microscope , and liquid penetrant) and documentation procedures were developed prior to rig testing of components.

3.5 Durability Testing - Results and Discussion 3.5.1 T56-A-7B second stage NGV

The Diffusion Braze repair scheme for Series II second stage NGVs (see Table 1

and Figure 1) was applied to two unserviceable NGVs from CF inventories (P /N 6871352)

with unknown TSN 's. Each of the two repaired components was burner rig tested to

2500 BRT cycles along with one new Series II second stage NGV (same P /N) for baseline

durability. In all, four components were tested through two sets of Series II NGV2 burner rig tests, with one new and one DBR'd NGV per test (S/N CST 1805 with S/N HT 1552, and S/N HT 4324 with S/N 27875). A close up view of the components under test in the rig assembly is shown in Figure 15a.

. Visual , stereo mic:oscope, and liquid penetrant inspection were performed, at

Intervals throughout the ng test, during 'disassembly inspections' in order to document ｣ｾ｡｣ｫ＠ damage (by ｬ･ｮｧｴｾＬ＠ position, and crack type) . Disassembly inspection involved ､Ｑｳ｡ｳｳ･ｭ｢ｬｾ＠ of the BRT f1xture, removal , cleaning , and inspection of the test NGVs, crack documentation, and reassembly of NGVs in the BRT fixture. Disassembly inspections were performed at 0, 25, 50, 75, 100, 150, 200, 250, 275, 375, 500, 650, 850, 950, 1100, 1250, 1500, 2000 , and 2500 BRT cycles for the first test (of new NGV _ S/N CST 1805 and DBR NGV- S/N HT 1552) and at 0, 250, 500, 980, 1500, 2000, and 2500 BRT cycles

for the ｳ･｣ｾｮ､＠ test (of. new NGV - S/N HT 4324 and DBR NGV _ S/N 27875). The

nu:nber of d1sassem?ly Inspections for the second test was reduced based on experience ga1ned through the f1rst test, for which it was not clear when the cracks would initiate and

how fast they would propagate.

Thermal fatigue cracking induced by rig testin was . . . .

to service-induced cracking for new and repaired

ｎｇｾｳ＠

Inn sln:Jiar, In. type an? location ,

ｷ･ｲｾ＠

all noted

ｯｾ＠

new and repaired parts. This

ｩｮ､ｩ｣ｾｴ･ｳ＠

ｾ［［ｾｾｾﾷ＠

axJal, and fillet c:acks

applied by BRT

ｾｾ＠

representative of service conditions. As

ｲ･ｧ｡ｲｾ［ｾｾｾｭｲｾ＠

ｬｾｩｾｾ､＠

history

some of the repa1red cracks did reopen during testing altho h P .

ｾ｡ｲｴｳＬ＠

ft t

r

1 • ug many were st1ll 1ntact

a er. es 1ng. t . was also. found . that ｣ｲｾ｣ｫｩｮｧ＠ occurred in areas that had not been

repa1red , at locations consistent With cracking in new parts. For exampl th · ·

crack, of length 0.380", on the DBR'd vane (S/N HT 1552) wh· h h de,b e maJor axial

· · f h' 1c a een cause for

reJeCtion o t IS part from service was found to have remained intact after 2SOO BRT

cycles

｡ｬｴｨｯｾｧｨ＠

new

｡ｸ ｩ ｾｬ＠

cracks had developed adjacent to it during testing. A

metallograph1.c

ｾｲｯｳｾ

Ｎ＠

sect1on t.hrough this crack, which had originally extended right

ｴｨｲｯｾｧｨ＠ the a1rfoJI tra11ing edge, 1s shown in Figure 16. The good structural integrity of the

repa1red crack should be noted.

In order to compare crack growth in new and repaired parts, crack lengths versus

ｮｾｭ｢･ｲＮ＠ of BRT cycles was plotted, for cracks having the same initiation sites and

ｯｮ･ｮｴ｡ｴｾｯｮＬ＠ fo\ all tes.ted components. It was found that cracking generally initiated

s?on.er 1n ｄｾｒ＠ d than 1n new NGVs. Meanwhile, crack growth rates generally decreased

w1th Ｎ Ｑｮ｣ｲ･｡ｳＱｾｧ＠ BRT cycles for all components. Finally, all four (4) tested NGVs were still

ｳ･ｲｾＱ｣･｡｢ｬ･＠ (1.e. no dar:nage ｾｸ｣･･､ｩｮｧ＠ serviceable limits) after 2500 BRT cycles (or the

eqUivalent ｾｦ＠ one TBO 1n serv1ce at the assumed cyclic usage rate). An example of crack

growth for. fillet cracks at a trailing edge

I

outer ring location is shown in Figure 17. From

an analys1s of all the crack growth data, it was estimated that cracks initiated at

ｾｰｰｲｯｸｩ Ｎｭ｡ｴ･ｬｹ＠ 700 and 300 cycles in new and repaired parts, respectively . It would be

Ｑｾｴ･ｲ･ｳｴＱｮｾ＠ to compare these crack initiation data with field experience but unfortunately f1eld data 1s not available for the components under test. However, it is interesting to note that a large fraction of first stage NGVs in a Series Ill engine were found recently, in a programme undertaken by Standard Aero Ltd. on behalf of NRC and DND, to contain

many cracks after only about 1200 hours of service since new (i.e. TSN

=

i 200 hours)

which, at the assumed cyclic usage rate, corresponds to about 600 rig cycles.

While the amount of rig testing of repaired components is limited, the results bear out the viability of the Vac-Aero Diffusion Braze Repair technique and suggest that repaired T56-A-7B second stage NGVs could be returned to service with confidence to endure at least one TBO. Based on extrapolation from the rig data, it would appear that repaired cracks will not grow beyond rejectable limits until approximately 3000 rig cycles or about 6000 hours of service. Vac-Aero International now offers this repair on a commercial basis and anyone interested in this service is welcome to contact the

company. It is worth noting that the repair will be field evaluated shortly in Canadian

Forces aero engines in collaboration with Standard Aero Ltd., the overhaul facility of CF T56 engines.

3.5.2 T56-A-14LFE/-15 first stage NGV

The Diffusion Braze Repair scheme for Series Ill first stage NGVs (see Table i and

' .. : ..• :.: :; •·: i :::

::[,:::?/'

. . · ··· : • . , , . . ,. '._.,.).1 : .•. ,: ! :,!' •I I' ' \ :',;::!:: ｾ＠: . I . I I .. ... . . . "'., .. ,, .... - . ... "1"" ''· :: ｾＺ ＺＺ＠ = :· ﾷ ｾ ｲＺＺＺｴ ＺＭ［Ｚ ＮＡＺﾷ＠

.:::m:\:kl,

.. :? \\

｜｜Ｎｾ＠

:.

ｾ｜Ｚｊ ［Ｋ ＺＺＺＺＧ｜｜ｩ ｜Ｚ｜Ｇ［ ＺＱＨ＠

, • • • : • • • t I ,, .. , • ＮＬ ＬＬｾ＠ '•' •P1'1'TJ,,, ,I j : · · .

ｮＺﾷＺ｜ｾＺＺ｜Ｚ ｜｜｜［ＡＺＺ ｷ ［Ｉ［｜ＺＺＺ［Ｑ＠

;. , . .• " !,' ... ,, ,1·\··. ,\,t

i :

I :

:

!

·: .· );::'

ｩ｜｜ｬｾ｜ Ｚ｜Ｚ

Ｑ

ｴ＠｜ｾﾷｾ＠ ﾷｾＢ［ ＡＺＩ［Ｚ＠

.' .

. 1· d to two unserviceable NGVs (P /N 6896896; S/N 39637 and S/N

F1gure 1) was app 1e DBR'd t

. h k TSN's (T'_{1me Since New) These two} componen s and two

195600) w1t un nown ·

new Series ₁₁₁ first stage NGVs (P /N 23031745; S/N 5066M and ?/N 381 K) were

｢ｵｲｾ･ｲ＠

· _{d t 250}

o

BRT cycles concurrently through one burner ng test. A close up v1ew

ng teste o • . · F. 15b

of the components under test in the rig assembly IS shown 1n 1gure .

Visual , stereo microscope, and liquid ｰ･ｮ･ｴｲ｡ｾｴ＠ ｩｮｳｰｾ｣ｴｩｾｾ＠ were performed, at

· t Is throughout the rig test during 'disassembly 1nspect1ons 1n order to document

tn erva ' . bl · t' 'd ·

crack damage (by length, position, and crack type). ｄｴｾ｡ｳｳ･ｮＺ＠ Y tnsp_ec ton ｷ｡ｾ＠ 1 entt.cal

in scope and extent to that used for Series II NGV2 ng test1ng as dtscussed 1n sectton 3.5.1 above. Disassembly inspections were performed at 0, 50, 100, 150, 200, 300, 400, 500, 750, 1000, 1500, 2000, and 2500 BRT cycles.

At about 1000 cycles, it was found that the internal cooling air distribution tubes had become loose, thereby compromising the efficiency of internal cooling . However, based on the apparent soundness of the external surfaces of the components, testing was continued to 2500 cycles as originally planned .

• • • • t' ot• t •lffiJ'loiH' ' ' ' '. '

''t•

,,J,r. I II

·:i

i

ＺＨＺＩＺ［ｩ［ＡｩｬｾＡＺ

Ｑ

ｬＡ ｾｬ ｾｬ［ ｕ＠ ＺＺ ｲ ｴｩＡｾｽｽ＾＠

Upon completion of the test, the components were inspected for cracks and

)

: i

ｾ＠ ｾＮ］ＺＡｩＺｅ ｐｩ Ａ ｩ ｾＡ ｾ ｆｕ ＺＡＧ Ｑ_{Ｇｈ Ｚ＠}

_::::;.:-·;

1_{: • subsequently sectioned for metallography and also to find out why the cooling tubes had}

ｪ ［Ｉ＼［［ＺＺＡ［

ｾＺ ｬＱ［ ［［ｾ ｪ ＺｪﾷＡＺＡＧｱＺ［

Ｚ ＺＺＺＺＮＺＺ＾Ｚﾷ＠

..

become loose. It was found that the tube material (347 stainless steel) had been entirely

Ｚｾ Ｚ Ｚ［ＮＺﾷＺＺＺ［ＺＺ［ Ａ ＺｕＺＺＮ ＬＺ ｮＺｲ［ｕＬ＠ !:·:.::::.::_:.:··;·_. oxidized to the point that the tubes were crumbling, particularly near the outer shroud of

ﾷｩＧｩＭＧ ｇＺ ［［ ［Ｚ ｾ ｩｻ ｈ［ ｗ ｾ［ ｾ［ Ｚ ｦｩ ｬＺ ｈｷ ［Ｚ［ＬＮＺＺＺＧＮ Ｚ＠

:;·;'\

the components. This indicated that the tubes were overheated during testing,

ﾷｪＺｪＡ ｾＡ ｕ ｬｩｾ ｬ ｊ ｪ［ＡＺ Ａ ［［ ﾷｮ ｾ［ ｪ ［ Ａｩｊ ｨ ＮＯ ＺｨﾷｩＺＺＯｃ＼＠

wesumably owing to insufficient internal cooling. In hindsight, a rate of supply of internal

Ｚｻ｜ｾｾＺｾ Ｚ［ ［ ［ ﾷ ｾ Ｚｾｩ ｾ ］ｲ ｰ ｦ ｆ Ｇｐ Ｏｻ ｩ ＨＺ ｊ Ａ Ａ ｽ ｊ Ｈ ｊＩｩＮＬ＠

·.ooling air

･ｱｵｾｬ＠

to

ｴｨ｡ｾ＠

provided

ｩｾ＠

engines rather than one apportioned to the hot gas

· . r . · ass flow rate 1n the ng (see sect1on 3.2), should have been used. In all future tests,

internal cooling air will be supplied at a rate of 0.0135 lb₁₁jsec per airfoil which is

approximately six times more than was used in the rig test.

As a result of insufficient internal cooling, the thermal fatigue cracking in rig tested

components differed somewhat from service induced damage. In particular, airfoil

cracking of test NGVs was confined to axial cracking occurring between mid airfoil locations and the outer shroud . However, the rig-induced damage for new parts was similar to that for repaired parts. Completely new cracks, on new and DBR'd test NGVs, were initiated and propagated during BRT. This indicates that the thermal load history ｡ｰｰｬｩｾ､＠ by BRT led to thermal fatigue cracking . However, the differences in cracking ｬｯ｣｡ｴＱ ｟ｾｮｳ＠ ｳｵｧｧｾｳｴ＠ _that the _thern:al history was not entirely representative of service

cond1t1ons . Th1s ､Ｑｦｦ･ｲ･ｾ｣･＠ 1s matnly due to the insufficient internal cooling of rig tested

components. However, 1t also appears that the outer shroud needs to be better shielded from the ?urner rig exhaust flame, since it was found that the outer shroud regions were

systematically_ more ｾｲ｡｣ｫ･､＠ than the inner shroud regions. Nonetheless, cracking was

､ｯ｣ｵｭ･ｮｾ･ｾ＠ Ｑｾ＠ deta1l to evalu_ate repair integrity and to compare the crack growth

｣ｨ｡ｲ｡｣ｴ･ｮｳｴＱｾｳ＠ 1n. new and repa1red ｾ｡ｲｴｳＮ＠ Some repaired cracks did reopen but several

others rema1ne_d 1ntact after. ?omplet1on ?f 2500 BRT cycles . Figure 18 is a cross section throug_h a repatred area_ (tra1l1ng e?ge ax1al crack - original length of 0.120") showing good

braze 1ntegnty and coat1ng adhes1on (to both the base metal

x-

₄

o

and the braze material)

even after 2500 BRT cycles. Based on metallographic evaluation of repaired and tested

components, ｴｨ･ｲｾ＠ appeared to be no adverse reactions between the Codep B coating

employed on repa1red parts and the repaired (brazed) areas.

Cracking generally initiated sooner in the DBR'd NGVs than in the new components

as shown in Figure 19. for. lead ｩｮｾ＠ edge

1

outer shroud cracks. Crack growth rates

ｧ･ｮｾｲ｡ｬｬｹ＠ ､･｣ｲ･ｾｳ･､＠ w1th 1ncreas1ng BRT cycles and were comparable in new and

repa1red parts, Ftgure 1 ｾ ﾷ＠ In ｣ｾｮｴｲ｡ｳｴ＠ to results from rig testing of T56-A-7B second stage

NGVs, all four (4) Senes Ill f1rst stage NGVs incurred damage during testing which exceeded serviceable limits. This is not entirely surprising because the parts were in all

evidenc_e overheated during testing. S/N 195600 (DBR'd NGV) and S/N 381 K (new NGV)

were reJectable at 300 and 750 BRT cycles, respectively , owing to multiple trailing edge axial cracks which were not separated by a distance equal to the length of the longer of

two cracks(15l. S/N 39637 (DBR'd NGV) and S/N 5066M (new NGV) were rejectable at

750 and 1500 cycles , respectively, owing to leading edge axial cracks exceeding 0.125"

in length(1s)_ Although the NGVs were all tested beyond their serviceable limit, to 2500

BRT cycles, no metal breakout occurred.

Although the test components incurred severe distress through rig testing, many of the repaired areas looked structurally sound after testing. This attests to the viability of the repair, which is not inconsistent with earlier results for Series II T56-A-7B second stage NGVs, which are made from the same base material and which were repaired following virtually identical procedures . For the Series Ill first stage NGVs, the rig testing conditions were obviously too severe and therefore it is difficult to draw any parallel with service experience . If the repair is to be qualified for use in aero engines partly through rig testing of the type described in this paper, test hardware modifications will need to be made such that operating conditions are better reproduced in the rig.

Notwithstanding the above testing shortcomings, the repair is an attractive

alternative to component replacement and is offered comr:nercially ?Y ｖ｡｣Ｍｾ･ｲｯ＠

International along with the repair for Series II engines (discussed 1n the prev1ous sect1on).

4.0 CONCLUSIONS

It was the intent of this paper to report on recent ､･ｶ･ｬｯｾｭ･ｮｴｳ＠ _at Vac-Aero

International concerning diffusion braze repair techniques for gas turbt_ne eng1ne h?t ｰ｡ｾｳ Ｎ＠

Extensive coupon and component-level testing has been performed 1n collaboration w1th

N RC-IAR to optimize and evaluate repair durability to: several components.

ｂ｡ｾ･､＠

on the

results achieved, repairs are now offered commercially for both cobalt and n1ckel-base superalloy components.

It was shown that it is possible to closely simulate NGV

ｯｰｾｲ｡ｴｩｮｧ＠

conditions in the

laboratory using the IAR burner rig facility . Using this rig and dedicated

｣ｯｭｾｯｮ･ｮｴＭｬ･ｶ･ｬ＠

testing hardware thermal fatigue damage typical of service can be

ｰｲｯ､ｵｾ･､＠

In both ne_w

and repaired

｣ｯｾｰｯｮ･ｮｴｳ

Ｎ＠

By comparing the response of

ｾ･ｷ＠

and repaired parts

ｴｯＮｾｬｧ＠

testing in terms of crack initiation and growth, it is then

ｰｯｦｳｳｬ｢ｬｾ＠

ｾ

Ｐ＠

｡ｴｳｳｾｾｾ＠

ｲｾｩｾｾｾｾｩｾｾｲｾｾｾｾｾ＠

expediently and at minimal cost. Based on the results 0 _{sue es s, . d}

ｾＧ＠ H;. '1 • • : ｾ＠_... . ' •

,. ｾ＠ .. ,

. . .

viable alternative to component replacement for service-damaged parts.

. f th repair for nickel-base superalloy components is currently Bench test1ng o e · · t b tt t.

I bt · d to date also demonstrate th1s repa1r o e a rae 1ve to underway. The resu ts o a1ne

turbine operators .

5.0 ACKNOWLEDGEMENTS

The authors acknowledge early contributions to this w?rk

ｦｲｯｾ＠

their former Vac-Aero colleague Pierre Azar, as well as cooperation and technical ass1stance from Verna Taylor and Bruce Orbanski of

ｳｴ｡ｾ､｡ｲ､＠

A.ero Limited. The

ｷｾｲｫ＠

was performed under IAR project JHFOO and under financial assistance from the Directorate of T:ansport and Helicopter Engineering and Maintenance (DTHEM) of the

ｏ･ｰ｡ｲｴｾ･ｮｴ＠

of Nat1onal Defence (DND) (FE 847788FACCG) and from the Industrial Research Assistance Program of NRC (CA103-7-1131). 6.0 REFERENCES 1. 2. 3. 4. 5. 6 . 7. 8. 9 .

s. Stoykewich, R. Armstrong, "Development and Application of Diffusion Braze Repair Techniques", 1987 NRCC Symposium on Industrial Applications of Gas Turbines.

D. Manente, "Repair Techniques for Nickel Base Turbine Components", Vac-Aero Progress Report for Period 1 July - 30 September 1987 under 1 RAP Contribution Arrangement No . CA103-7-1131, 18 November 1987.

P. Stoute , J-P. lmmarigeon , and R. Dainty, "Diffusion Braze Repair Developments for Nickel-base Superalloy Gas Turbine Components", NRCC, IAR LTR-ST-1661,

8 April 1988.

D. Manente, Milestone No.5 , IRAP Contribution Arrangement No . CA103-7-1131, 21 March 1991 .

W. Young , "Materials Synergisms", (Conference Proceedings), Kiamesha Lake, N.Y., U.S .A., 1978, pp. 924-938.

M. Haafkens, "Experiences in Repair of Hot Section Gas Turbine Components", Aerospace Cong ress and Exposition , Anaheim, California, 1982, SAE.

K.

Schneider et al, "Materials Science and Technology" Volume 1 August 1985,

pp . 613-619. ' '

ａｬｬｩｾｯｮ＠ T56 CIP ｾｯ｣ｵｭ･ｮｴｳＬ＠ Pre-CIP .Meeting and Problem Solving conference, lnd1anapol1s, lnd1ana , June 1991 , on f1le at DND-DTHEM 4-2 Ottawa ontario.

J I

T56 Repair and Protective coatings - Proposal for foil _ k St d d Aero L d Wt ., . . 1nn1peg, M . b an1to a,March1991. ow on war ' an ar

12

10. T56 Engine IEMP Review Meeting , Kelly AFB, Texas, October 1989, Item 89-27. 11. Allison Gas Turbine, "T56-A-14LFE / -15 & -78 Training Manuals", Indianapolis,

Indiana, 1971 .

12. Detroit Diesel Allison , Engineering Program Notice (EPN) 5624.2.23R1-4-4, 2 January 1982.

13. T56 Problems and SRD Review Conference, April25-27, 1989, Indianapolis Hilton Airport, Allison presentation .

14. C . Korn, N. Marchand, and S. Gendron, "On-Line Damage Assessment of Coated and Uncoated Blades in Simulated Service Conditions", 5111 Interim Report, COT Project P1380, NRC Contract No. 31946-9-0001, August 1990.

15. Detroit Diesel Allison , T56-A-14LFE/ -15 Overhaul Manual, Indianapolis, Indiana, 1978.

. . .. :. : · .. . .. .. -· ·

ｾＨｾＭＮＺＮ ｾ＠

.•.•.

ｌ ＮｾＮ Ｍ

.:c•• ;.

<; .•.•

L

•> . ··· ••-•-•. · :· •-• •• • .··

- - - :. - ··- -:":.: . ｾＭ ... :.: .:· ·

-TABLE 1 Diffusion Braze Repair schemes for Allison T56 series II 2nd stage and

series Ill 1st stage nozzle guide vanes

SERIES II SERIES Ill

PROCESS STEPS T56-A-7B T56-A-14LFE/ -15

2nd STAGE NGV 1st STAGE NGV

Perform preliminary cleaning operations ./ ./

Remove inner and outer cool ing cavity - ./

caps and air distribution tube

Strip coating - ./

L.P. process and document component ./ ./

condition

Perform dimensional inspection ./ ./

Clean by hydrogen reduction ./ ./

Apply diffusion braze alloy ./ ./

Vacuum furnace braze ./ ./

Inspect: visually, L.P., dimensional, and ./ ./

Xray

Rebraze if necessary ./ ./

Perform diffusion cycle ./ ./

Inspect: visual ly ./ ./

Bench dress braze alloy ./ ./

Inspect: visually, L.P. , and dimensional ./ ./

Install and vacuum furnace braze cool ing - ./

cavity caps and tube

Apply Codep B coating - ./

.. ::· ... ... . : .. ; .. _::. : . · : · · · ·:-.. . . :···-···· .. --- -- -. . ..

ｾＭｾｧ［｟

Ｎ＠ＭｾＮ＠ ＭｾＭＭｾＭｾ Ｎ＠ＭｾＭ ﾷｬｾ＠ ｾＭ ﾷｧＮ］Ｍ ｾＭｾ ＺｾｾｾＺＺＺＺｾ＠ ｾ］Ｇﾷ ｾＺｾｾＭﾷＺＬ｜ＨｾｩＲ

ｾｾｬｴ｟ｾ｜［＠

;::':;::· __

ＡＺＺＮＭＮＺＺ｟ｾ ］ｾｾ＠

b) FIGURE I _a) b)

Allison T56 Series II second stage NGV; solid (no internal cooling) and uncoated.

Allison T56 Series Ill first stage NGV; hollow, internally cooled and coated (as received) .

FIGURE 3

ｾﾷＭＭＭＭＭＭＭＭＭＭＭＭＭＭＭＭＭＭＭＭＭＭＭＭＭＭ

...

\NCO 792 C.T. nozzle as received - sectioned for metallographic evaluation.

INCO 792 C.T. nozzle test pieces·

AFOR / DBR processing (on right) . ' as-received (on left) and atter

a)

b)

FIGURE 4 INCO 792 C.T. nozzle.

nickel plating protective coating

a) Typical oxidized crack in as-received condition.

Kalling's reagent. Nickel plated tor metallography.

b) Similar crack after repair by AFOR/ DBR process.

f . . . ｾ＠

a)

b)

FIGURE 5 _{Current Vac-Aero braze repairs}

_{un .}

a) IN-738 nozzles; AFOR

｡ｮ､ｾｾｾｾｾｾ＠

ｾｆｏｾ＠

cleaning .

b) IN-718 component· AFOR epatred.

, and braze repaired.

FIGURE 6 Circumferential temperature distribution on first stage NGV from

T56-A-15 engine (Rolls Royce data supplied by Standard Aero Ltd.)

CAN #6

Thermocouple Positions 1-18

1059

CAN # 1 _{VIEW LOOKING FROM REAR}

1062

1126

DIRECTION OF ROTATION 1129 CAN #2

1112 1083 1088 CAN #5 METAL TEMPERATURE

0

. .

OVER 520 · C BELOW 780·C DOVER 7BO·C BELOW 940•C

. E2j

OVER 940·C BEL.OW 1030·C

ｾｾ＠ OVER 1030·C BELOW 10BO·C

CAN #4

1088

1112

1156

CAN #3

PEAK GAS STREAM TEMPERATURES IN ·CARE SHOWN AGAINST

FIGURE 7

b)

a)

c)

Can-type burner rig exhaust nozzles for rig testing of Series II 2nd

stage and Series 1111 st stage NGVs; along with an actual combustor can nozzle from a T56 Series II engine.

...

.. . ·.··

ﾷ ﾷ ｾＮ＠

. .··

Ｎ Ｍ ＾ｾＱＺｾｾｽ＠

ＭｉｾｾｾｾｾｩＩ ＭｾＬＭＭｾ ＭＬｬｾｾ ｾｾ［ ｾ ｾ ｴｾｩｾｾｩｬ｟ｾｾｾｾｾｬｦ＠

ｾ＠ 0 Ill '

I

'

I

'

I

'

I

I

I

ＭＭＭｾＭＭＭｾｾＭＭ

zz

I

ｾｷ＠

I

(J)Q} ｾＭｾ＠

I

ww t-(J)

I

'

I

<.!)

z

a:

a: w

z

z

z

0 i= () w (f)

t?

w 1-<.!)

z

<.!)

z

a:

a: w

5

0 <.!)

z

a:

<.!)

z

z

ｾ＠

a: >-:0

E

Q) ({) ({) co Q) Ol co ..., ({) C\J

>

(9

z

... 0

-

,., ' ''":! ' ; • ... ; •' '' 'P •, ' " ,:· ':

TEST ZONE #2 (»W..Sf &I..Cf TEST ZONE #1 TEST POSITION 1 TEST ZONE #2 TEST POSITION 2

FIGURE 9 Schematic of test zones and ·test fixture positions for Series Ill NGV1 BAT

FIGURE 10

tO..ttTD«i P'LAl[

linG I

Schematic of Series II NG\V2 BRT assembly .

IOSlDH PILLOY ll..IXX SRPJ6-I <ZJ atff[llt LK cr II.AHU ···E·-·--·-·-·-·----·-·----·-·-·-·- "'""' L>< or

=• .,. ..._,

JASC ｛ＱＶＬＶｾ＠ OUTCR: RtHC ["952 1.-.!C ｾｉｄｓＺ＠ ,,," ..

BURH[R SUPPORT TAB1..C:

TEST ZONE #1

Ｚｾ＠ ｾＺＭ｟［ＺＺ ﾷ＠

""'-q fi. :·

'> :·.

l'' ' . •' .' 11 0 ·• , I ｉ ｾ＠ i:, j I, 1

.i.i;

_{:j. l I I : : ::.}

:ii·i ..

:;{: Iii!..

_{1' 1 I I , I l l} f i 1! ::::: I I . I · : ; .. ·j; : i } ::If:

1-i:!j:: !: :

l! it:

·i: :·

il!h!:i!:::· ·.

'!IiI :,:, i .. .

!·:l:rr! : : : : :.

!·:.1 ''.: '··:· :: 1_;1;;:: i _{i: : : .. :} 1_{1 'i · I; : ...} \:i'l. i-l ·l· .. : . . , ! . ｾ＠:.I : ｾ＠; : . . i ! '! ; ｾＺＺ＠ .': . :. : ' :: . l ':J,jJ :·::: ; : : :: 0 ｉｉ ｾ Ｚ＠I :. : : . 1 t : : • I'! i l i o • ｾ＠I I : , , I ll ,· ! .. . \; ... . '\· ｔｾＮ［＠ ｾ＠ :" !:: : ' Ｇ｜ｾｬｻ ｜Ｂ ＮＱ＠ •I ' .

!:rl:!:;:.t,! \ .. :.

' ' Ｍ ｾ＠ i•t'' I ' +!:1 \, t!!·\ ｾ＠ ｾＺ＠

r·l:n\ \:::: ::.

_{l. :,\, .'} I : : .. .. ｜ ｾ｜｜ ｜ ＠ ｾ＠"\.'\ :. ; . . " . ｜ﾷＭｾｾ ｜ ［＠ ﾷｸ Ｎｾ Ｎ Ｌ＠ Ｚ Ｎ ｾ＠

.: ·. _

0 ' 01 \•\<t•\: 0 I. : ' ' • ' ' I ( "l ·\' ｜ｾ＠ <\' . I , ; I ﾷﾷ ｾﾷ｜｜ｾ＠ .. ｜ｾ｜｜ＧＮｾ｜ｾ｜ＺｾＭｾ＠ ; ,', : . :-::::. ', :.•',. >:· ｾＺﾷｾﾷｾＺＮ＠ '.\::\: '. :

....-...

u

0

..._

1200

1100

1000

900

Tl GJ c :::0 m __, __, OJ c --. :J CD --. --.

ca·

..-+ CD (J) ..-+ ---h OJ Q . rl-' '<

FIGURE 12 Typical surface temperature response of Series Ill NGV1 to BRT cycle (of 1200 o C).

... . o .... -.. -· o ... .... o ···· .... ·o ... ...

.... .

ＰｾＶＭＶＭＰＬ＠

ＭＶｾＰ＠

0/6

ｾ＠

6

6 internal cool 1ng

01

r

connected

ｾ＠

800

o internal cool 1ng o1r

not connected

ｾ＠

ｾ＠

₇₀₀

E--4

ｾ＠ ｾ＠ ｾ＠ Pressure Side

P-4

ｾ＠

ｾ＠ ｾ＠ ₀ L...::.

\

0

200

100

ｾ ﾷ＠

ＢＭＭＧｾＰ＠

o

ｾｄＭｯ＠

o

ｯｾｾｾＭＭｾｾＭＭｾｾＭＭｾｾＭＭｾｾＭＭｾｾＭＭｾｾＭＭｾｾ＠

0

15 JO 45 60 75 90 105 120 135 150 165 180 195 2 10 225 240

TIME <SECONDS)

ＬＬＬｔｾｾｾｾｾｾｾＧｾＧ

Ｌｻ｛ｉ｡ｾｲＮｬｬｩｾｴｾｾｾｭ［［ｩｲＧｾ＠

Ｌｲｲｬｾ ｾ［ｾ［ＬｩｾＬｾ［ｾ ｾ ｦｷｬｾｾｊｾｭｾ

ＺｾＺ［ＺＺ＠

-.

""'

I

I

ｾｾ＠

..J u ｾ＠ " " " 0 ... ..J 0 ;;: z M "' 0 t..n (\J u 0 0 0

-w ...J a. :J 0 u 0 l: 0: w J: 1-...J 0 0: 1-ｾ＠ 0 z 0 u w _J _J N LL N _(J) 0 z <t: 1-(!) tl) f-::> ₀ <! I I X w LL w 0 0..

z

>- ₀ 1-I f-z ｾ＠ 0 u

w

a: 0

.

-

.

. _... , _: . . .. _{··· ·· .} .

..

... •• 0 .. ..

ｾ｜ｾｩｾｾｾｾｾｩｾ

ｦｾｾｾｾｬｾ］ ｾｾｾｉｉＺ＠

c

0 ... ro ... .0 co (_) Q) ｾ＠ 0 c ... (/) ::J co J: X Q) OJ

·-=

_... Q) c ... ::J .:0 ... 0 '+-Q) .::r:. co 1-Q) Q. ::J 0 (_) 0

E

,_ Q) J:

f-NOZZLE EXIT AREA

§39 §34 .gl 18 §12 J74

ｾＵＰ＠ ｾＳＵ＠

ｾＲＸ＠

ｾＴＲ＠

f63

ｾＷＳ＠

Jl80 Jl89 ,$6 - ｾＸ ｟Ｓ＠ _ _ - ---=.Q.. ｾＶＹ＠ ｾＸＵ＠ 1000 .1 011 1012 01003 1009 d' v - - - - _ o_ - 0----<> - "---::1-G-1-9- _ 1012 1043 cl 0 1025 o;..; _ cl J002 ｾＹＹ＠ J006 77 J 030 1030 0 J025 o 1030 o1029 J 039 R38 JlOS 26 J378 ° 01029 §32 /:;" J336 J392 o1 031 Jl84 a) o569

NOZZLE EXIT AREA

b)

FIGURE 14 a) Series II NGV2 BRT nozzle temperature distribution at 1" from nozzle ti p (exhaust control temperature 1000 ' C).

b) Series Ill NGV1 BRT nozzle temperature distribution at 1" from

nozzle tip (exh aust control temperature 1100 ' C) .

a)

FIGURE 15 _a)

b) Close-up view of S . _{Close-up view of Se .} enes II 2nd t _{sage NGV under ng test.} . nes 111 1st stage NGV under rig test.

FIGUF={E 16 Series II 2nd stage NGV; repaired and rig tested to 2500 cycles . Cross-section of diffusion braze repaired crack after testing. Kalling's reagent.

.

-

·--. _..

:.::::·::::

. ':' ·· ... -: ·· ·· ·· . : · . ..

... Q) 0 ..c ｾ＠₀ '-O'l ｾ＠ 0 (lj ' -0 ... (f) Q) ... O'l ·;:: '-Q)

E

c:n

::J c ..0 ·;:: '> '-0.!!)

z

6

Q)..._ O'lQ)

i90'l

(J)""O -o Q) c O'l N _s;

=

·c.u '--(/)+-' Q) ... ·c ro Q) (f) ＨｦＩｾ＠ coo U") (lj f-

0

w 0: ::l 0 u. 0 LD 0 Ｍ ｾ Ｍ·- . .. ·- .: ｾ＠ tn tn ｾ＠ ｾ＠

0::

z

-rf).

>

0

z

ｾ＠ ｾ＠

A

<I <I tn 0 co ｾ＠ ｾ＠ rf).

u

z

-

rf).

>

0

z

ｾ＠

ｾ＠

z

0 LD 0 <I 0 0 0 ＼ｓＳＺｈｾｎ ｄ＠ 0

ｾＰ＠

<J

· ... <l '·. ·. ·. ·. '• .

<J

LD _{I'-- 0 LO} 0

ｾ＠

(\J 0 ₀ 0 _.

ｈＮｌｾｎＺｉＱ＠

ｬｩＸｖＱｉｾ＠

o

0 0 0 ( ) 0 LD ["--.. C\J 0 0 LD C\J 0 LD C\J C\J or.t;, Ｐ ｾ＠ Ｐ ｾ＠ C\J U

Ｐｾ＠

m U ["--.. ｾ＠ or.t;, Ｐｾ＠ 1-(-J ｾ＠ 0 0 LO 0 LO (\j 0 0 0 0 0 0

p

a) b) ｆｉｇｕ ｾｅ＠

IS

nickel plating Codep B coating

Series Ill 1st stage NGV; repaired and rig tested to 2500 cycles. a) Cross-section of diffusion braze repai red crack after testing .

As-polished . Nickel plated for metallography.

b) SEM micrograph of surface area away from braze . Backscattered electron image, as-polished . Nickel plated for metallography.

: :':. : ! • ,: ,.· . .. . '•. ··· Ｇｾｾ Ｚ＠ ' . I '

...

0.6

I

0.5 :

ｾ＠ ＰＮｾ＠ ｾ＠

u

z

... ｾ＠ ｾ＠

O.J

0

z

ｾ＠ H ｾ＠ u ｾ＠ p:j

u

0 . 2

0. 1

FIGURE 19 T56 Series Ill 1st stage NGJ burner rig test; crack growth of axial

cracks at leading edge

I

otter ring .

, -6 DllR NG V SIN 39-637 0 DllR NG V SIN 195600 0 NEW NG V SIN 5066M 0 NEW NG V SIN 381K ＬＮＮＭ ﾷＧ ＧＡＧｾＮ＠ A j ... /

oo

ｯ ｾ ｯＺＺＺｯﾷﾷﾷﾷｯ＠

u ... ···

o···

_D,. 0"""""""""

I

6 0

L

c

6 - - - - 6 0···"''''' ''"'"'''''"'''" 0 / / / / / /

_o

/ / / / / / / 0 ...

... ..· o ---· ---o

...

ｯＮｯｾｾＭＭｾＭＭｾＭＭｾＭＭｾＭＭ ｾＭＭｾＭＭｌ｟ｾｌ｟｟ｊ＠

__

_ i _ _ ｾ＠

750

1 000 1 250 1 5DO 1 750 2000 2250 2500 2750 JOOO

0

250

500

UURNER RIG 'IEST CYCLES

ｾＮ＠ ｾＭ