21. Overhaul and repair of used parts.

Figure "Determination of overhaul intervals" (Lit. 21-2): Both diagrams show by means of the bathtub curve (volume1, Ill. 5.2.1.3-5) two philosophies. They differ in the determination of the overhaul intervals corresponding the fixed interval maintenance (Hard Times), and the „on(line) condition maintenance” (Fig. "Concepts of overhaul fixed or on condition").

The upper diagram shows the philosophy of the former `Eastern Block' . Here in the foreground stood the military application. It can be characterised as „conflict related philosophy“. This based on the consideration, that in a at every time possible conflict the aeroengine must be available in the best as possible condition. This demands a replacement considerably before the deterioration caused operation life time is reached. Naturally this approach leads to relatively short overhaul intervals and high costs. They even rise further if the design of the parts is oriented at the short intervals. Especially concerned are expensive rotor components with LCF-limited lifetime. Because a repair is no more possible, even after a relatively short lifetime the scrapping is needed. Those disadvantages obviously can not be compensated by more beneficial repair costs for the reason of relatively short operation times of less deteriorated parts. Obviously today, everywhere in the civil and military aviation, the „peace related philosophy” prevailed for economic reasons (diagram below). Here with fixed overhaul intervals the really achievable safe lifetime of the of the appropriate designed component is utilised.

The trend seems to progress even a step further. Thereby will be changed over from fixed overhaul intervals to „on-condition maintenance“ (Fig. "Concepts of overhaul fixed or on condition").

Figure "Concepts of overhaul fixed or on condition" (Lit. 21-1): For the time intervals after which an aeroengine must be overhauled, two concepts prevailed:

Fixed interval maintenance (Fig. "Determination of overhaul intervals") and on-condition maintenance. The „philosophy” is not necessarily linked to a aeroengine type, but can depend from the airplane type. Thereby the approach is operator specific. Also different versions of an aeroengine type can be treated different.

Both approaches have its advantages and disadvantages.

Fixed interval maintenance is the traditional approach of the aeroengine overhaul. It considers especially the design lifetime/guaranteed lifetime of the hot parts. Should the overhaul interval be extended, certificates/proofs are necessary. For these, the condition of the operation typical aeroengine/-es is very thoroughly investigated during the disassembly and the overhaul and the condition of the parts determined („sampling“). Those activities can be very extensive, time consuming and expensive. For example if the approved cycle number of rotor disks should be increased. To this security providing cyclic spintests can get necessary. They are carried out with disks which had operated under representative service conditions. For this the OEM gives support and recommendations. The approval/certification gives the responsible authority. This, than by the OEM in the manual scheduled overhaul interval, refers usually to operation hours and/or start-shut down cycles (Fig. "Result of unsufficient lifetime tracking").

During the on-condition maintenance the overhaul takes place as a targeted exchange of a component which reached its lifetime limit. Naturally there will be tried to combine consumed lifetimes respectively remaining service life of the components of an aeroengine as favorable as possible. This is true as well for new parts as also for used parts. This concerns not at least a demanding logistic task.

So it depends on weather to identify the condition („health”) of the part with sufficient certainty. For this different approaches are possible, which also can be combined (chapter 25.1). Sensors/probes monitor and record the lifetime determining operation parameters. From those the lifetime usage respectively remaining lifetime is calculated with the algorithms, established by the OEM. This is supported by the modern „trend monitoring“ and electronic control units (full authority digital electronic control = FADEC). Typical example is the lifetime (cyclic, LCF) of rotor disks which is determined by starts/shut downs. A damage of those parts can not be detected in time, whatever visual or with non destructive tests (volume 3, Ill. 12.6.1-8). In contrast borescope inspections (chapter 25.2.2.1) can be used at components with visual identifiable damage criteria in the mounted/assembled condition. They can assure the evaluations of the trend monitoring based on data of sensors/probes and/or used alone as lifetime criterion. An example are failures and crack development by thermal fatigue or oxidation on vanes/nozzles of the high pressure turbine (volume 3, chapter 12.6.2). To exchange the concerned parts a module design is helpful. It enables the exchange of the parts with a limited effort.

Wants an operator to proceed according to on-condition, an detailed scheduling with the OEM is necessary. The approval finally is carried out from the responsible aviation authority (e.g., FAA).

Figure "Cost optimisation by overhaul interval" (Lit. 21-1): Naturally the operation costs with compliance of the necessary safety, stand in the foreground when an optimal overhaul interval respectively the „overhaul philosophy” is determined.

The overhaul intervals for fixed interval maintenance but also the point in time for on-condition maintenance of the overhaul can be extended with increasing experience (e.g., fleetleader principle) from the safety point of view. Then the question for the most cost-efficient optmised overhaul interval arises. The costs of the aeroengine alone and its total costs, that means under consideration of the indirect costs, follow a bath tub curve (Fig. "Determination of overhaul intervals"). First the costs drop with longer overhaul intervals, however will increase any time again. This may depend thereon, that expensive components will be damaged above a feasibility of a repair and so must be replaced with new expensive parts (Fig. "Identifying deterioration of components by aging"). As the curves show just the indirect costs markedly rise after long operation times. This may be due to an increased fuel consumption. It develops as a result of the rub in abrasion of the labyrinth seals, the increase of of the tip gaps in compressor and turbine, erosion wear at the compressor blading as well as distortion and change of the profiles in the turbine (volume 2, Ill. 7.0-3). Additionally the compensation of the performance drop (e.g., take off power) requires increased hot gas temperatures. This leads to a accelerated deterioration of expensive hotparts and so degrades the on-condition-benefits.

Figure "Identifying deterioration of components by aging": Which failures/deteriorations can be repaired? How long can be waited till a repair must be carried out? What characterises failures which can be repaired, correspondent to the overhaul manual or specifications?

Basically counts, same failure modes at different or equal parts of different aeroengines don' t mean, that the same repair limits are valid. This gets understandable with the example of a turbine stator blade (vane/nozzle) compared with a turbine rotor. A role plays the component specific and also application specific operating loading. Finally decides the approving/certificating authority after recommendations of the OEM. In the picture on hand with the help of rotation turbine components the problem should be explained.

Blade edges („A“): The orange peel effect (volume 3, Ill. I11. 2.3.1-10) by oxidation and thermal fatigue is well identifiable for an expert. This deterioration is limited and can be reworked by removal if there is sufficient remaining blade geometry/profile. A build up by brazing or welding as possible at guide vanes can be ruled out for rotor blades because of the high operating loads. At least the same should be valid for the erosion of leading edges (carbon particles/carbon erosion). This deterioration can be easy distinguished from thermal caused damages by burr formation (Fig. "Identification of a failure").

Blade regions with suspicion of overheating („D”) are usually not sufficient sure to identify. For an experienced specialist the surface structure or deformations can be hints (volume 3, Ill. 11.2.3.2-8). In less critical applications a repair will be abandoned (local heat treatment) respectively this is not sheduled. That counts if it can act on the assumption that during normal operation temperatures an age hardening effect occurres and the strength is again increasing.

Longtime deterioration by creep at the blade(„H“) is hardly visible if it does not show with markedly deformations (volume 3, Ill. 11.2.3.2-8 and Ill. 12.5-15). Does the deterioration occur with a formation of creep voids (normally only at blades from wrought/forged alloys of elder aeroengine types, volume 3, Ill. 12.5-7) and isn't yet in an too extensive advanced stage (standard necessary), a regeneration with HIP can be performed (Fig. "HIP process limits" and Fig. "Hot isostatic press device").

Longtime change of structure („E”) like an orientation („raftening“) and/or coarsening of the `gamma prime' phase or the formation of brittle pases (e.g., laves phase, sigma phase) can mean drop of strength and embrittlement. Those changes are not visible from the outside. With a suitable heat treating a regneration is thinkable.

Damage of the blade by vibration fatigue (HCF) („F”)is neither identifiable in time nor controllable. This is caused by the fast crack propagation during high frequency vibrations (volume 3, Ill. 12.2-21). A repair, even of a deterioration without crack formation is not possible, a success not verifiable.

Fretting wear at the outer shroud („G“) can be identified in time and repairable by welding on (hardfacing).

Drop of strain hardening at the blade root („J”) is not visible, however considered in the design and will be regenerated by shot peening usually during every overhaul.

Labyrinth tips („B“): The repairability (Fig. "Welding repaired labyrinths") depends from the operation loads (volume 2, Ill. 7.2.3-12). If high dynamic loads must be expected (LCF, HCF) this can markedly limit a rework. Hot tears, caused by rubbing (volume 2, Ill. 7.2.2-9.2) usually, can be detected with a penetrant inspection. Also a visual recognisability of lager cracks is possible.

As rework the local removing of damaged rubbing zones or a welding on of abraded labyrinth tips is possible.

Hub bore and other highly loaded disk zones („C”) permit at the end of the lifetime no repair of a possible LCF deterioration. Crack formation is neither allowed nor repairable. This is also true for the absence of cracks, if the lifetime is exhausted according to the specifications (Fig. "Retirement for cause" and chapter 24). This is caused by the fact, that till now no for series applicable test/procedure exists,which can prove sufficient sure a LCF deterioration (Ill. 25.2.2.2-2). The removing of small damages like scratches with the necessary attention is only possible according to the repair specifications (Fig. "Failure by improper blending of a turbine disk"). Usually after this a shot peening is required.

Figure "Overhaul intervals and military mission" (Lit. 21-7): This type of helicopter gets in desert environments in military operation. Thereby ingested sand caused serious problems. Primarily it's erosion of the blading and the inside of the casing from the compressor (frame in the middle right, volume 1, Ill. 5.3.1-10). This could also not be sufficiently avoided with an upsteam sand separator (sketch down left and frame down right). As consequence the overhaul intervals had to be markedly shortened.

Figure "Replacement by a repaired or new part" (Lit. 21-3): The more expensive a part, the greater are the efforts for a cost-saving repair. The usual limit from which the repair of a part is no more worth is reached if the repair costs are more than 60% the new part. This border region is marked by the angular dark stripe.

The frequency of a repair to be expected (chapter 21.3.2) is an important criteria which decides about the development of a repair process. This demands proofs/tests and approvals/certifications and can be very expensive. A repair which is seldom needed may be therefore not worthwile, even if it does not exceed the 60% limit.

There are definitely cases in which the repair is prevented because of limited technical means (e.g., machinery) of a shop or a technology till now not available.

As it can be seen in the diagram, primarily for the repair parts with damage mechanisms like wear, oxidation, thermal fatigue and erosion come into consideration. Those are rather longtime effects. In this time period a spontaneuos catastrophic failure is not to be expected. They than can be considered as controllable (blade tips, labyrinth tips). For foreign object damages only limited sizes are concerned (specifications in the overhaul manual, Fig. "Rework instruction for fan blades"). For the particular application with sufficient safety they must not expect a crack development.

Figure "Retirement for cause" (Lit. 21-4, Lit 21-5 and Lit. 21-6): The „retirement for cause-principle“ was already discussed in the early 80s. It is caused by the LCF lifetime limitation (volume 3, chapter 12.6.1) of rotor components like disks and rings of at this time modern aeroengine types („damage tolerant design”, volume 3, Ill. 13-14). The high safety demands also consider statistical failure probabilities. Because these are extremely low (10^-9 for disc fractures per flight hour), a further use over the lifetime, verified by calculations and tests (e.g., with a sampling program) is forbidden. For the operation primarily the so called incubation is used. In it no distinct crack propagation occures. Even if in single cases such one should emerge it is ensured, that it is far from endangering the part (upper diagram).

Only in a negligibly small percentage of those expensive parts a deterioration can be expected which influences the lifetime in kind of a dangerous crack growth. This probability may lie clearly below 10 ^-3. So the attempt is suggested to use the high potential for life time extension and cost reduction of the components.

However the precondition was the sufficient sure finding of deteriorations (Ill. 25.2.2.2-2 and volume 4, Ill. 17.3.1-2 and Ill. 17.3.1-3.1) which guarantees in the overhaul interval a sufficient safety against spontanepous failing (fracture, diagram below).

During investigations emerged, that till now for civil application with its typical long operation times no sufficient safe non destructive testings for practical series use are offered. The application meanwhile emerged for military aeroengines in only few rather experimental applications.

Figure "Repair limits of used parts with material change": A special problematic of the repair is the change respectively deterioration of the parts/components under influences of the operation (Fig. "Changes of material properties by repair"). These changings can trigger serious problems up to non applicability for prosesses like welding, brazing, etching or coating and stripping which make no earnest problems during the production of new parts.

Increase of the sensibility for external loads:Problematic is a drop in strength, especially in combination with a dangerous embrittlement. Cause are deteriorations of grain boundaries or the depletion of alloy components by diffusion processes which starting at the surface can also reach deeper into the base material. This promotes failures like crack development under tensioning, machining forces or straightening during repair.

Etching processes (chapter 21.2.4) can act markedly more damaging as at new parts (volume 4, chapter 16.2.1.7). This may be caused by chemically noticeable more stable oxides than the base material. The consequence can be attacks of the grain boundaries or locally very differing removal of the material.

Sensitizing the material structure by operation temperatures (volume 1, Ill. 5.4.1.1-9) also intergranular attack/corrosion (IGA, volume 1, Ill. 5.4.2.1-2).

Stripping (chapter 21.2.3) of coatings which have been altered during operation (oxidation, change of structure, forming of carbides, depletion of the alloy) can show markedly more difficult as in the new condition. Coatings like MCrAlY or rub coatings, especially in the rear of the compressor and on hot parts, can be chemically stable in an extend, that an increased danger of a damaging for the base material exists.

Brazing repairs of hot parts are especially hindred by oxides on the braze surfaces, esceptionally in cracks which must be closed. To remove the oxides, treatments with aggressive media (process baths, heat treatment in hot gas atmosphere) are necessary.

Welding repairs on Ni-alloys, mostly because of its problematic weldability caused by the development of hot tears (= heat cracking, Fig. "Replacement by a repaired or new part" and Fig. "Welding of creep resistant alloys") are an especial challenge. Changes of material structure caused by the operation (e.g., coarsening of the age hardening phase = `gamma prime' - phase) can promote the formation of hot tears. Not sufficient removed surface contaminations (e.g., sulfidation) in the welding zone affect the strength and ductility of the welding itself.

Figure "Risk of overhaul aberrations" (Lit. 21-8): Already one year before the 1st incident during the overhaul of the aeroengine corrosion (Detail Mitte) was identified beneath the root platform of the turbine rotor blades from the 2nd stage (left frame). This lead at that time to the exchange of the whole set by refurbished blades out of two other aeroengines. Those blades got a protection coating (platin aluminide = PtAl diffusion coating). Actually such coatings are rather an oxidation protection.

The aeroengine of the here concerned 2nd incident which occurred only 10 days after the 1st incidnet was also overhauled one year before the failure. Also here all blades of the concerned stage had to be exchanged because of corrosion below the root platform. However in this case new blades with protection coating have been introduced.

After the incidents both aeroengines had been disassembled at the OEM. It arose, that in both cases turbine rotorblades of the 2nd stage were broken due to stress corrosion cracking.

As cause sulfurous and salt containing deposits below the root platform had been identified. Those triggered corrosion pittings whose high notch effect released the cracks.

Such failures are known since about 5 years from diverse different aeroengine types. These obviously occurred depending from the operator. Since the blades got a protection coating. One year before the here discussed incidences there was a change to a less corrosion susceptible blade material. Obviously the coating was not sufficient to meet the problem. So it was once more replaced by a new version. But obviously also this did not prove in case of deposits.

Especially the aeroengine of the 2nd incident had unexpected with 1020 start/stop cycles an even lower lifetime as this of the 1st incident (1306 cycles). The corrosion of the associated new turbine blades was markedly stronger (about 0,3 mm deep pittings) than this at the somewhat longer operated (already reworked) of the 1st incident (about 0,15 mm deep pittings).

In both cases obviously the corrosion damage increased after the overhaul dramatically in spite of the protection coatings. In contrast the uncoated blades survived with 4000 cycles longer without cracking („).

This suggests the conclusion, that a increasing deposition of corrosive dust (dolomite + sulfur) can be seen as main failure influence after the overhaul.
As cause operation induced gaps between the rear side plate and the blade roots have been identified (lower sketch right). Obviously those have not been shut in spite ot the rework (increased?). In the aeroengine of the 2nd incident those have been especially pronounced.

Note: Not before the failure mechanism is sufficiently understood and the causative influences are identified a sufficient remediy can be expected!

Figure "Personnel for high risky repair processes" (Lit. 21-9): Some repair processes of elder aeroenging types have an especial high danger potential. For this reason those processes will be replaced as far as possible by not so hazardous. But this can also again trigger problems (e.g. changed sliding features).

Plating with cadmium: Besides the toxicity in connection with cadmium (Cd) two extremely nasty effects which are especially pronounced emerge:

• Hydrogen embrittlement (volume1, Ill. 5.4.4.2-1; and volume 4, Ill. 16.2.1.8.3-7): During the galvanic cadmium plating process from experience an especially high danger of embrittlenet exists.
  

• LME, SMIE (volume 4, Ill. 16.2.2.3-11): The sketches in the upper frame show a compressor disk (volume 4, Ill. 16.2.1.8.3-8) with a Ni-Cd plating. The Ni-barrier layer has to avoid the metallic contact of the cadmium with the steel disk. This did not work in the shown case. It came to a crack formation and the fracture of the disk.
  

Stress corrosion cracking during burnishing (volume 1, Ill. 5.4.2.2-4): Gears get their typical brown colour with a process in hot brine. Have the gears sufficient high internal tension stresses at the surface, this can trigger crack formation (lower sketch; volume 4, Ill. 16.2.1.8.3-11). Because of this danger this process was banned in the last years, at least by some OEMs.

Figure "One disaster occurs seldom alone": It strikes, that frequently later investigations at many heavy incidents come upon similar cases which can be assessed as parallel cases. They merely had not comparable serious consequences.

Here also a connection exists with the problem of the „single case“ (volume 1, Ill. 2-3.3).
Not seldom little differences of the cases are taken as reason to classify every as single case. Thereby the connection with other cases with quite similar failure mode is not identified or will be denied. Seemingly the classification as a single event/case has an additional, however very risky benefit. It is believed, because a follow-up is ruled out, measures, remedies and reports can be avoided.

Therefore at incidents with similar failure modes it should be insistently looked for the connection. This is the key and the chance to find the real failure mechanism and to identify the causative influences. This is the requirement for targeted, promising remedies and measures. This counts not at least for a repair.

Our particular attention should apply for effects which strike while comparing the cases. For example a seemingly discrepance like more severe failures at components with less operation hours. Also agreements like a concentration of the failures at distinctive operators give hints. (Fig. "Risk of overhaul aberrations").

Note: Parallel cases are a big chance to identify the real causes of failures. So they are a precondition for successful repairs. Important are differences which influence the failure mode unexpected by comparing the cases.

Figure "Trend analysis for problems indication" (Lit 21-9, Lit. 21-10 and Lit 21-11): If certain parameters of the aeroengine change during the operation time without conscious assistance from outside, a damage/deterioration can be suggested at. Thereby the chance exists to border/locate and identify possible failure locations and failure types. This picture contains typical examples from an older civil two shaft aeroengine type (JT8D).
In the diagram typical operation parameters are upright plotted, the operation time is at the abscissa.

Leaks in the high pressure compressor: Examples are failures on air removal tubes and bleed valves (volume 1, Ill. 5.3.2-16). In steady state operation such a leakage leads to the drop of the compressor end pressure. More fuel must be supplied for performance maintenance, in order to raise the turbine inlet temperature and the corresponding performance. This happens on condition that the speed of the compressor is increased and the demanded end pressure is built up again.

A compressor leak has a similar effect as the deterioration of compressor efficiency (Ill. 3.1.1-2). All the described influences lead to a remarkable alteration of the monitored parameter.

Combustor failure (fractures of the combustor walls volume 3, Ill. 11.2.2.2-1, failures on the injection nozzles): These failures are difficult to identify from the monitoring data. Often, secondary damages in the turbine make a recognition already possible. These are typical parameter changes that are also to be expected in turbine failures: Drop of speed, rise of turbine inlet temperature, as a consequence of increased fuel addition, in order to guarantee the demanded performance.

Failure on HPT guide vanes (nozzles, volume e, Ill. 2.3.2-7): If a larger share of the vanes is affected, the speed falls significantly. Turbine inlet temperature and fuel consumption increases, similar to the combustor failures. To be noticed is the clearly higher vibration measurements not explained in the literature. To be considered is, however, a vibration excitation, through flow irregularity at the periphery in the turbine area.

Failure on the HPC bearing: The HPC speed recedes as expected. Entirely surprising is, however, the clear reduction of the vibration level, observable in the rear engine area. This clearly shows how much expertise is imperative for the correct evaluation of the monitoring parameter.

Figure "Result of unsufficient lifetime tracking" (Lit 21-12): The investigation of the promary failed aeroengine was carried out at the OEM with attendance of the flight safety authorities. It could be seen, that three neighbouring firtrees of the high pressure turbine wheel were broken and the four associated blades lacked (sketch in the middle). The turbine disk underwent a laboratory investigation.

The investigation of the fracture surfaces and the material structure suggested that the lifetime was exceeded. For this, the combination of the operation loads, typical for this component have been responsible: Cyclic load (LCF), creep and hot gas corrosion. Further, not yet broken firtrees of the disk already showed damage features (micro cracking).

History: During a 150 hours inspection chips were found in the oil filter. The OEM got an oil sample. From this he diagnosed a bearing failure in the accessory gear. After further 18 flight hours, 1 day before the flight accident, the aeroengine was repaired.

Influence of the maintenance at the failed aeroengine (frame below): It became apparent, that the lifetime, limited by a service bulletin from the OEM at 5200 start-stop-cycles, was exceeded from the high pressure turbine disk. Before the disk was assembled in two other aeroengines („A“ and „B”). In „B“ deviating from an OEM instruction about 4000 cycles have not been registered. With this the disk had already at the assembly about 6600 cycles. So the permitted lifetime cycles have been markedly exceeded.

Comment: The question arises, if there is a connection between the bearing failure in the accessory gear and the turbine failure. Thinkable is, that an unblance of the turbine wheel announced the failure development. Can it have been transferred from the high pressure shaft through the radial shaft into the gear and damged here the bearing?Although there was obviously no plausible explanation for the bearing failure, these symptoms have been cured. However probably this was not the real cause which lead later to the failure of the turbine wheel.

Against this consideration should speak the experience of the OEM with bearing failures in the gear and its assessment.

Figure "Problem estimating used lifetime" (Lit 21-13 and Lit. 21-14): The aeroengine had at the point in time of the accident about 8000 operation hours with about 6800 start-shut down-cycles.

The investigation of the failed aeroengine (sketch in the middle) showed:
The firtree root of a single blade from the 2nd stage gas generator turbine wheel was broken (sketch below left). The fracture proceeds in the very short shaft of the blade, close below the platform (detail below right). The incipient crack is a fatigue crack (LCF), which propagated from an edge of the root shaft on the exit side (detail below right).

History: About 7 years before (Lit. 21-14) a similar case is reported. Also here a fatigue fracture is concerned (obviously LCF) which was also positioned above the firtree root. It has well propagated from the rear side. Further 27 blades showed at the same position incipient cracks. There was no connection with faults. Therefore already then an excess of the cyclic lifetime (LCF) was suggested.

Counting method of cycles: At this aeroengine type the procedure is as following:
In the new condition and after an overhaul, in a certificate the usuable lifetime is specified in cycles and operation hours for every component. If one of those values is reached the aeroengine must be overhauled.

Distinguished are full cycles/main cycles (start-shut down-cycles) and part cycles/mini cycles (for every shaft speed below 85 % ) during operation.
With this operation characteristics are considered. From these data the lifetime relevant cycles are calculated. Two methods are used of which one produces a higher comsumption of lifetime.

The gas genreator consists of five lifetime limited components, all are rotordisks from the compressor and the turbine. The blades of the 2nd HPT rotorstage are not lifetime limited. However tests suggest a safe lifetime in the size of two certified disk lifes.

In the former failure case the aeroengine was mounted at an ovrehaul engine. Before about 4 600 cycles have been accumulated. Also the blades have been further used. Since the aeroengine had about 2 300 start cycles during an almost identical number of operation hours. l. Tis correlates with the mini cycles of about 4 400 - 5 500 main cycles (according the calculation method).

The consumed lifetime at the former operator of the blades from the gas generator module also was recalculated. It lay between 9 000 and 11000 cycles. This agreed with the lifetime evaluation of a material investigation. The design life obviously is about 12 000 hours. That so the lifetime of the blades was almost reached was not realized, because the operation data have been not sufficient exact. So olso thic chance to identify the deteriorated turbine wheel was lost.

Comment: Unfortunately it can not be seen from the literature of the former failure to what extent there have been further problems with the registration of the lifetime. However if there is a parallel case after many years, it can be supposed that the problems are still not definitely solved. Possibly also the lifetime determining operation load is not sufficient identified.

Figure "Seemingly harmeless maintenanve procedures" (Lit 21-17, see also example 9.2-12): After the start during climb an aeroengine failure with the exit of fragments occurred. The following investigation showed, that an oilfire in the region of a main bearing of the turbine arose. The disk of the last stage from low pressure turbine had plastically expanded about 25 mm by overheating. All blades of this stage had separated.

The leakage of the pressure oil was causexcd by a vibration fatigue fracture about 90° circumferential in an oil line. This started from the inside of the tube at a spiral scratch. This tubes of the oil line must be cleaned periodically because of oil coke formation. For this the use of tools like wire brushs, drills or steel needles is not allowed. Those could injure the inner side of the tube.