21.2.3 Repair coatings
We find the processes of repair coating/plating (Fig. "Coatings for repair and overhaul") also in the production of new parts (volume 4, chapter 126.96.36.199). Problems like coating faults/failures or deteriorations (e.g., crack formation) however can develop as result of an operation caused change of the surface and/or the base material (Fig. "Typical proglems with repair coatings"). Dimensional changes can limit the number of repeatable removing cleanings (etching, abrasive blasting).
At some elder aeroengine types there are still used coating/plating processes like browning (Fig. "Stress corrosion cracking in burnishing baths") or cadmium plating (Fig. "Danger of LME and LMIE by galvanic platings"). Those have an especially endangering potential. Because the use of such processes gets more and more seldom, the danger exists, that those problems ar no more known or are no more sufficien aware by the personnel in overhaul and repair.
Frequently coatings/platings in the repair not only serve the replacement. They are also used for the restoration of worn component areas (fretting, rubbing, oxidation, erosion).
Figure "Coatings for repair and overhaul" (Lit. 21.2.3-1 and Lit. 21.2.3-2): The different coatings (volume 4, chapter188.8.131.52) are precondition for the necessary application specific operation behaviours. These parts are deteriorated during operation by fretting and rubbing, oxidation, aging, erosion and spalling.
In spite of the same coating processes for new parts and repair parts the requirements at the processes technology can quite differ. Typical repair coating processes and examples for its applications are:
- Thermal spray processes (A, B, D, E, F, I, K, N).
- Diffusion coatings (D,N, O).
- Galvanic processes (G, H).
- Vapor deposition: PVD (D).
- Hard facing by welding (E, N), brazing (C, I, K).
- Painting/finshing (M, O).
In the most cases repairs for the replacement of a coating may be concerned. Not seldom the lifetime of the coating is first limited by design. Such a coating must be periodically reconditioned with a repair during overhaul. For example this is the case for sealing surfaces with rub and rub in coatings (volume 2, Ill. 7.0-1) or sliding coatings/layers on blade roots.
But there are also cases where uncoated surfaces are repaired with a coating. Usually these are wear surfaces. An example is the crome plating of a sealing sliding surface or of a bearing seat at gear shafts (H). Therby the repair plating can be used as improvement of the operation behaviour (e.g., sliding properties) compared with the origuinal uncoated properties of the new part.
A special precondition for a successful re-coating will be often late aware. It concerns a, for the base material mild removing/stripping of the old deteriorated coating respectively its rests. The origin of the difficulties is often already the design phase. If not enough attention was payed at the repair, a problem can be created (Fig. "Welding repaired labyrinths"). An example is the attack of material specific carbides of forgings during stripping of ceramic spray coatings with glycolic acid. The repair process is hindered respectively prevented if a safe and effective remove of a coating is not possible without deteriorating other materials of the component. Problematic for example is a, compared with the base material, similar but chemically more stable coating like MCrAlY. This can exclude a removal by etching. A coating can change the structure during operation, by overheating, aging, oxidation or diffusion (e.g., depletion of the alloy). So a stripping process that has been proved at new parts can no more applied for repair without too high risks. This shows the necessity of a realistic testing of the repair process (chapter 21.3). Changes of the coating process parameters or the transfer of a process to an other part requires besides the test/proof also the approval of the OEM respectively of the responsible authorities.
Figure "Typical proglems with repair coatings": A recoating as it is typical for repair coatings, can make markedly more problems than a coating of a new part (volume 4, Ill. 184.108.40.206-1). Naturally the problems depend from influences like coating type, condition or the component and the coating process itself.
It is also a big difference if a recoating or a primary coating is concerned. Was a coating already at the new part, it's designed for this and adapted in its operation characteridstics. This is not true for a repair coating which should avoid a problem which was identified during operation.
An example is the later introduction of an erosion protection coating for a compressor blading. This was originally not planned by the designer. Therefore its influences on the component strength are not considered. Such a coating usually must be hard to have the necessary abrasion resistance. Sutch coatings are mostly brittle. This leads to a dangerous FOD behaviour (volume 3, Ill. 220.127.116.11-19.2). A further example is the oxidation protection of hot parts by a diffusion coating. Such coatings behave at least at lower temperatures brittle. So they can downgrade the thermal fatigue behaviour (LCF) and the HCF strength. The operation behaviour of vibration loaded parts can essential depend from the sliding conditions at the contact surfaces. This is especially true for the dovetail blade roots of compressor rotor blades. Even seemingly insignificant deviations of coatings and/or geometries, caused by the coating, can have an unexpected effect (volume 2, Ill. 6.1-15.1). Thereby very different effects can be concerned like damping or the load of the contact surfaces (combined tension and shear, volume 2, Ill. 6.1-11).
For wear/abrasion problems always the whole „tribological system“ must be considered. Is only one contact surface coated or the coating was changed, this can lead exactly to the contrary of the desired effect (Fig. "Use of new overhaul instructions during repair").
For a recoating the described effects should be spared. However, also then it must be reckoned with problems which did not occur respectively have not been expected during the new parts production. Concerned are operation caused changes of the base material, like:
- Dimensional caused limitations for the stripping process.
- Remained residues and contaminations (Fig. "Problems by thr removal of coatings").
These can complicate the recoating and deteriorate the operation behaviour compared with the new part coating.
At such problems increased coating faults/flaws can be traced back. With these the lifetime of aluminium diffusion coatings for the oxidation protection of turbine blades(vanes) shortens. Was the old coating removed by alumina blasting (vapour blasting), as a consequence of the „charging effect” (volume 1, Ill. 5.3.1-7) too much fractured blasting particles can stick in the surface. This hinders afterwards the formation of an optimal diffusion coating.
Figure "Problems of cooling holes by repair coating": The coating of cooled turbine blades/vanes is already during the new part production a demanding task. A repair takes place under complicated conditions. Especially critical are the cooling air holes (cooling veil) to the airfoil surface (right sketch). The problem is, to guarantee a sufficient protecting layer also in the operation influenced/changed holes/channels. Does this not succeed, the danger exists, that in the holes increased oxidation occurs. This can restrict the cross section (detail). The bore geometry can also change and the cooling deteriorates, if the oxides breake out. A so increased blade temperature can markedly shorten the lifetime.
Figure "Danger of LME and LMIE by galvanic platings" (Lit. 21.2.3-3): „Liquid Metal Embrittlement“ (LME, volume 4, Ill. 18.104.22.168-11) can be triggered by many metal melts. It's a brittle crack formation. Thereby a lower melting metal „dips” into the base material which underlies tensile stresses. This danger may also be the reason why successful used rub in coatings based on silver brazing metal (e.g., „Easy Flow“) in elder aeroengine types can not be used just like that in modern aeroengine types. The explanation could be a susceptibility for this failure mode of nickel alloys and tanium alloys with silver alloys. The typical steels of elder aeroengine types seem less prone for this failure mechanism.
Generally also for steels cadmium may count as especially dangerous (Fig. "Casing behaviour like apressure vessel"). Its melting point and/or temperature for a dangerous diffusion is often below the usual operation temperatures in aeroengines. In such cases, for the cadmium corrosion protection platings of steels (e.g. casings), attention must be payed at an instruction conform coverage/masking (sketch right) of the part zones marked in the manuals (sketch left). To this also belong the dense, sufficient thick nickel barrier coating to the base material.
Figure "Stress corrosion cracking in burnishing baths": The at least former used burnishing (blacking) process used a hot brine. Mostly it is applied at case hardened steel parts, usually gears and/or shafts. Its technically justified sense was always, at least in cycles of experts, questionable. Obviously the attractive appearance of these parts was in the foreground. A markedly corrosion protection could not be expected. Also the sticking of a protecting oil film, in contrast to porous layers like phosphate layers, was not given.
If the parts for burnishing have internal tensile stresses a special type of stress corrosion cracking (SCC), the so called „caustic cracking/embrittlement” can occur (volume 1, Ill. 22.214.171.124-4 and volume 3, Ill. 126.96.36.199.3-11).
Shot peening serves as a remedy (steel shot, glass beads). With this the protecting effect of the induced internal compression stresses at the surface is used. However this protection can be only used for sufficient accessible and geometric suitable component regions. For example the often inside sleeves/gears positioned spline toothings can be hardly protected in this manner (volume1, Ill. 188.8.131.52-4).
Cracks which run along the toothing are especially difficult to find with magnetic crack detection, usually used for these parts.
Because of the potential danger of cracking, some OEMs now forbid burnishing.
Figure "Engine failure after faulty repair" and 21.2.3-6.2 (Lit. 21.2.3-4): The airplane was damaged at several positions. A first investigation of the failed aeroengine showed:
- In the plane between 5th and 6th rotor stage an ripped open casing of the low pressure turbine (LPT, sketch in the frame).
- The rim of the 5th stage rotor disk was separated (sketch below left).
- The rotorblades of the 5th stage had fractured in the middle of the airfoil.
- In the „6-12 o'clock “ region of the circumference heavy damaged stator blades and rotor blades of the 6th LPT stage.
- Heavy damages at the turbine exit casing and the exit guide vanes.
History primarily failed part: About 4 years before the incident, the LPT module was disassembled from an other aeroengine and overhauled. Thereby the inner airseal (interstage airseal ring) was repaired at a sub contractor. About 1 year later the module was assembled into the failed engine. The mounting into the concerned airplane took place about half a year before the incident.
About 6 weeks before the incident, during a borescope inspection a crack in the nozzle guide vane (NGV) of the high pressure turbine (HPT) was identified. After this the inspection period was shortened from 500 operation hours to 250.
After the disassembly of the aeroengine could be seen:
- From the separated disk rim of the 5th LPT stage only 2 fragments with about 40% of the total disk slots had been remained. The laboratory investigation unfolded from oxidation respectively thermal discolourations the front side of the rim must have been markedly hotter than the rear side. The fracture surface showed a crystalline rupture (90° to the surface, typical for a tension forced fracture), bordered by two surface near typical smooth shear surfaces each in 45 ° angle (volume 3, Ill. 12.3-3).A metallographic investigation showed at the frontside of the disk, besides the main crack, also additional cracks. Those were filled with oxide. This hints at an locally, unusual high temperature of the part.
The hardness (31-33 HRC) corresponds with the requirements of the specification.
- The ring of the interstage seal has 60 bores for the needed cooling air. During the assembly, attention was attracted from 56 bores which are clogged by a plasma spray coating. The remaining 4 holes had been only partly blocked because coating material had chipped off.The plasma coating was more careful analysed. It corresponds with the certified coating from 95% nickel and 5% aluminium.
Consequence of clogged cooling air holes/bores in the interstage sealring shows Fig. "Warning hints in process sheets"
Examination of the work sheet for the repair of the ring from the interstage seal. The work sheet contains as well the process steps as also the approval comments.
After the plasma coating process the cooling air holes had to be deburred and then approved by an inspection. However there was in the manual no instructuion how the inspection of the cooling air holes must be carried our.
At the following warning in the overhaul manual (Fig. "Warning hints in process sheets")
„No plasma coat is permitted in airseal holes or on hole edge breaks”
was not pointed in the accompanying documents of them part. Because the part was repaired at a subcontractor the ordering repair shop had not compiled a process sheet.
The responsible inspector declared when questioned, that when he inspected the part, the cooling holes had been open (!?).
An inspection of the storage by the responsible authority found a further part on which merely 2 holes have been partly clogged.
The following preventive measures have been introduced by the responsible shop:
- Preliminary cancellation of the plasma spray process and withdrawal of the internal approval.
- Amendment of the process sheets (work sheets). Now they require, that the inspector signs especially the inspection of the cooling air holes.
- The personnel and the inspector of the plasma coating got an intense training about the process specifications, the masking procedure and the test technologies.
- Examination/inspection of the whole procedure related to the accompanying documents to insure, that they contain the warning notes in the manuals.
- Elctronic information (telephone, internet/E-mail, fax) of all concerned customers about the necessity of an inspection for clogging of the cooling air holes from the interstage seal rings. However further concerned parts have not been found.
- Search of the causes of the process failure. Therefore an examination of all processes at the concerned shop was carried out.
- All testing personnel had to attend trainings about „Human Factors“.
- From the FAA all operators of the concerned aeroengine type got an accordant Airworthiness Directive (AD-note).
The failure investigation showed as an important factor of the failure development, that obviously the consequences of clogged cooling air holes was not known respectively was not aware by the personnel. A warning note as „sheet anchor” lacked in the accompanying documents.
Understood and aware perceptions are a requirement to avoid malbehaviour and failures.
Figure "Disk overheating by lacking cooling air" (see Fig. "Engine failure after faulty repair", Lit. 21.2.3-4): The cooling air holes supply the disk rim with air jets. Correct functioning (detail right) they lower the temperature from the rim to the hub sufficiently (diagram).
The thermal gradient is during idle about 100 °C, at the end of the start 230°C and during cruise about 150°C.
Are the coolimg air holes clogged (detail left) hotgas can impinge the disk rim and heat it up almost to the gas temperature. This is especially true for the start during which also the failure accurred (Fig. "Engine failure after faulty repair"). In this operational state the temperature difference between rim and hub bore is about 450 °C, about the double as with open holes. This means a rim temperature of nearly 685 ° C (gas temperature) in contrast to the normal component temperature of about 470°C. Within this the creep strength and with it the creep life drops extremely (several orders of magnitude! volume 3, Ill. 12.5-4). Simultaneously the thermal stresses rise markedly. This caused the fracture of the disk.
Literature of chapter 21.2.3
21.2.3-1 P.Adam, „Fertigungsverfahren von Turboflugtriebwerken“, Birkhäuser Verlag, 1998, ISBN 3-7643-5971-4, page 112-148, 153-160, .
21.2.3-2 „ Engine casing repair: present and future”, Zeitschrift „Aircraft Technology Engineering Maintenance“, June/July 1997, page 68-74.
21.2.3-3 National Transportation Safety Board, Aircraft Accident Report No. AAR-96/03 „Uncontained Engine Failure/Fire Valujet airlines Flight 597, Douglas DC-9-32, N908VJ, Atlanta, Georgia, June 8,1995” page 1-117.
21.2.3-3 National Transportation Safety Committee, Department of Communications, Republic of Indonesia, Aircraft Accident Report 2003, „Garuda Indonesia Flight GA880, Boeing B747-200 PK-GSD, In flight (21 minutes after takeoff from Denpasar, Bali), 23. November 2001“, page 1-16 + addendum,.