Aeroengine Safety

Fig. "Appearance of bolt operation failures" (Lit. 23.3.1.2-3 and Lit. 23.3.1.2-7 up to Lit.23.3.1.2-9): Mechanical operation loads, which causative lead to bolt fractures, are primarily dynamic. Thereby fatigue fractures/cracks develop (HCF, LCF, Fig. "Where bolts faiL during operation", volume 3, chapter 12.6). These can as well develop through high frequency vibrations, as also low frequency loads like thermal fatigue or cyclic changes of centrifugal forces.
Fatigue fractures also show at ductile materials without plastic deformation. Anyway they are not summarized under the term brittle fractures (Fig. "Brittle failure modes of bolts and nuts"), because the material is not embrittled but appears only so by the crack propagation mechanism. If an interpretation in the SEM is still possible (Lit 23.3.1.2-10), they can be certain identified in most cases by the expert. With enough experience, the fracture surface enables also a macroscopic assessment (Fig. "What fracture features of bolts tell us", volume 3, Ill. 12.2.1-3). Thereby at least first hints about type and hight of the dynamic load can be expected (volume 3, Ill. 12.2.1-2 and Ill. 12.2.1-3).
At bolts caused by stress concentration at load transmissions, there are areas which are predestined for vibration fatigue fractures during overload (Fig. "Where bolts faiL during operation").
Fatigue fractures at design conform loads, can start from weak points, caused by production, assembly or operation deteriorations. to these belong (frame above):
`pittings' like from corrosion (at steels, volume 1. Ill. 5.4.1.2-6, Ill. 5.4.2.2-1; volume 2, Ill. 6.2-13) and sulfidation pittings (at Ni alloys, Fig. "Dangers to bolts at temperatures").
Fretting wear (volume 2, Ill. 6.2-9). During the assembly at the shaft `seizig grooves' can develop.

Static load causes at sufficient high operation temperature failures as creep fractures with plastic deformation. During long operation periods under low loads the creep strain can be wey small. Not necessarily it is noticed through markedly plastic deformation. Anyway, also in these cases, fracture causing material embrittlement can not supposed. A feature is intensified oxidation of the fracture surface, compared with the residual fracture. This complicates also the microscopic analysis, especially in the region of the crack origin. Metallografical, the formation of creep voids (volume 3, Ill. 12.5-7) can enable a certain confirmation.

Forced fractures because of mechanical overload are extremely seldom at bolts of aeroengines. They can develop as secondary failures, for example

• in the case of a containment (e.g., fragment impact, volume 2, Ill. 8.2-15) at the bolting or casing flanges.
• During extreme unbalances (blade fracture, bird impact) at rotor boltings.
• As shear fracture at centric tension bolts/rods after the fracture of the hub from a turbine wheel.
• At flange boltings in rotors by a swirling around fragment (Fig. "Bolt fracture at hot parts by silver") or a foreigen object (e.g., tool, Fig. "Foreign object remained by assembly" and Fig. "Design caused problematic assembly").

Depending from type and direction of the overload (tension, bending, shear, torsion), a typical fracture progress can form (frame below). From the macroscopic and microscopic fracture appearance, if necessary, with a macro etching (at a damaged fracture surface, volume 3, Ill. 12.2.1-5) forced fractures can be identified, as well as load direction and load type can be analysed.

Fig. "Brittle failure modes of bolts and nuts" (Lit. 23.3.1.2-3): By far the most fractures and cracks of bolts in gas turbines show, except forced fractures (Fig. "Appearance of bolt operation failures"), as normally secondary failures, at least macroscopic a brittle appearance. This can have different causes:

Causative embrittlement:

Stresscorrosion cracking (SCC, volume 1, Ill. 5.4.2.2-1) is a potential threat for bolts and nuts from high strength steels („A1”, „A2“). To cracks and fractures it comes only under design conform conditions, if the structure/material deviates from the specifications. Mostly this can be proved, if the hardness limits are exceeded (mostly 32 HRc). The fracture patterns (Fig. "Brittle failure modes of bolts and nuts" and volume 1, Ill. 5.4.2.1-6) seem often pronounced crystalline and show corrosion features (rust), especially at the origin. Microscopic, the specialist can identify this failure mode on evaluable fracture surfaces sure and without problems. Features show the relationship of the failure mechanism with hydrogen embrittlement (volume 1, Ill. 5.4.4.1-2 and Ill. 5.4.4.1-4).

Hydrogen embrittlement („B1”, „B2“, „B3”) is caused by hydrogen, which diffused(volume 1, Ill. 5.4.4.1-2) into the material during production or overhaul processes, when there existed a too long time interval than specified, till the anti embrittling process occurred. This embrittlement develops over a longer time during storage or in the operation. It is irreversible and can not be proven by an impact test (falling weight test, volume 1, Ill. 5.4.4.1-6). Typical processes that cause hydrogen embrittlement can be galvanic coating, etching and the stripping of coatings (volume 4, Ill. 16.2.1.7-14).

Embrittlement by diffusion of solid foreigen metals (SMIE, volume 3, Ill. 12.4-14, volume 4, Ill. 16.2.2.3-11). This danger exists at unforseen high operation temperatures. Cracks start preferential in the thread („C1“, Fig. "Bolt fracture at hot parts by silver").

Embrittlement by dipping of foreigen melts (LME, volume 4, Ill. 16.2.2.3-10.1, example 16.2.2.3-3, Ill. 16.2.2.3-11). Thereby, in a high speed process, wetting metal melt penetrates under enough high tension stresses the material. („D1”). The origin area of the fracture surface can show an unnormal discoloration (silvery), which can not be explained by oxidation.

Fig. "What fracture features of bolts tell us" (Lit. 23.3.1.2-3 and Lit. 23.3.1.2-10): Already the macroscopic fracture appearance enables the experienced expert inportant conclusions at the cause of the incipient crack (sketches above), as well as type hight and variation in time of the loads (volume 3, Ill. 12.2.1-2 and Ill. 12.2.1-3).
In the scanning electron microscope (SEM, volume 4, Ill. 17.3.2-7) undamaged fracture surfaces can be excellent evaluated (detail below). They enable conclusions at

• Crack origin.
• crack propagation, lapse of time,
• cause of the crack start /weak points/flaws,
• failure mechanism.

Metallographic investigations (volume 4, Ill. 17.3.2-6) evaluate usually the material from its structure deviances and strength (hardness). Especially for the assessment of the forging structure and a strengthening process (rolled head transition radius and/or thread, Fig. "SUP bolts with deviations" and Fig. "Identifying during assembly a too soft bogus nut") with the evaluation of the `'fibre flow' and of foreign material diffusion (LME, SMIE), these investigations are irreplacable.

Fig. "Where bolts faiL during operation" (Lit. 23.3.1.2-1, Lit. 23.3.1.2-2, Lit. 23.3.1.2-5 and Lit. 23.3.1.2-6): Force transmission and cross-sectional jumps are because of the notch effect (chart above right) highly loaded. This is especially noticeable at the fatigue strength of the connection. Therefore will be distinguished between the SCF α_k (stress concentration factor), which finally applied for a brittle material and the notch effect factor ß_k which considers a specific, good-natured behaviour of the real material. With appropriate measures like work-hardening or surface hardening; notches can be eased. This applies especially for rolled threads and a strain-hardened transition radius from the bolt head to the shaft. This applies also for potential material inhomogentities (weak points). They just can be beneficial influenced in the mentioned notch regions with an appropriate fibre flow. For this; for example; the bolt head is upset forged or the thread rolled (Fig. "SUP bolts with deviations").
The frame in the middle left, sHows the statistic likelihood of a fatigue fracture for exposed regions of a bolt. It can be observed, that the fatigue fractures concentrate with 65% at the first loaded thread turns. This reflects in the below shown stress distribution in the bolt. The high stress peak in the first loaded thread turns gets understandable, if the load distribution at the nut in the first three thread turns is considered (frame middle right).
The sketch below left shows schematic the stress field with lines of the same stress. Thereby the stress concentration in the head radius with the associated stress peak is good to identify. This can be met with an optimized stiffness distribution in the bolt head (sketch below right).

Fig. "Production caused bolt head failing" (Lit. 23.3.1.2-3, Lit. 23.3.1.2-5 and Lit. 23.3.1.2-7): Although seldom, bolt failures caused by production can not be fully ruled out. Some attract the attention with an atypical location and progress. To these belong fractures inside the bolt head. During upset forging of the head, an inner crack formation can occur. It triggers during operation the fracture. A critical region is the transition radius (sketch right). Even fractures in the bolthead, at the bottom of the centric sinking or below it (sketch left), have been observed.
Besides these forging flaws there are also further failure causes triggered by the production (Fig. "Brittle failure modes of bolts and nuts"):

• Stress corrosion cracking at steel bolts of too high hardness (volume 1, Ill. 5.4.4.1-1, Ill. 5.4.4.1-2).
• Hydrogen embrittlement (volume 4, Ill. 16.2.1.8.3-7).
• Rolling flaws in the thread (Fig. "Identifying during assembly a too soft bogus nut"),
• threaded (cutted) thread instead of a rolled thread (Fig. "SUP bolts with deviations" ).

Fig. "Bolt fracture by etching residues": This picture represents an exception: It is a combination of faults from production/repair and assembly in connection with blind holes.
Auxilary materials of the production, like cooling lubricants (volume 4, Ill. 16.2.1.1-13) of the chipping (drilling) or fluids from etching, cleaning and solvents from repair processes, can get into the thread of blind holes. There the possibility exists, that they dry or thicken. So rests will stay (sketch left). Will now a high strength i.e. susceptible bolt (Fig. "What fracture features of bolts tell us") screwed in, contact with the mecdium arises. The dried up residues develop a dangerous electrolyte with the air humidity respectively condensate if the flages and bolting are untight. Thereby especially chlorine containing residues must be named. These can trigger stress corrosion cracking in and the fracture of the bolt.

Fig. "Zipper effect at bolting interconnections" (Lit. 23.3.1.2-11): At connections which depend of the bearing function of a multitude similar elements (flange boltings, sketch above right; fixing of turbine guide vanes, sketch below), a typical failure mechanism can be observed:
Fails/breaks a single element (bolt), especially the neighboured elements must bear additionally the `free' load. Are these bolts already loadet in the limit range, possibly already deteriorated, they will be overloaded and fail also (sketch above right, Fig. "Evaluation of the risk of a broken bolt"). So it comes to a further increase of the load at the remaining bolts. To this self-energising, accelerated failure process, fits the term „zipper effect“.

Fig. "Evaluation of the risk of a broken bolt" (Lit. 23.3.1.2-11): The left aeroengine of the airplane (sketch above) suffered a failure during rolling at the start. After that the start was cancelled.
The following investigations showed, that fixings/lockings of the vane segments (sketch and detail below) of the 2nd stage stator from the low pressure turbine had failed (sketch middle right). They prevented a rotation of the stator, caused by the forces of the gas deflection.
Since the overhaul, the aeroengine had about 12 000 operation hours with about 1800 start/shut down cycles.
The concerned fixings have been 8 „lockings ” for the 16 vane segments. They consist of a sort U-shaped `bridge' from Waspalloy. Its lugs latch into the outer shrouds of respectively two neighboured segments. The `bridge' itself is fixed with a threaded pin, radial through the casing to the outside (sketch below). The nut, tightened from the outside has a silver plating. The threaded pins (stud bolts) have been broken, as in former cases, at the transition to the bridge. Unfortunately there are no mor failed parts in the bolted condition. So the break loose torque could no more determined. The lockings of the 3rd and 4th turbine stage correlate these of the 2nd and do still function. At two „bridges“, the break loose torque was markedly lower than at others.

The fracture surface of a threaded pin of the 2nd stage and a crack showed no signs of corrosion. One crack of the 3rd stage proceeded, intercrystalline and was oxidized. In the fracture origin silver of the nut thread was smeared. The fracture structure let suggest at LCF (thermal fatigue? bending gas forces?) and creep as causative load. It showed, that both load types acted in combination. With this a time depending failing could be explained.
With the failing of all fixings, the stator segments could rotate in the casing. These `milled' through the casing wall and ripped the outer casing (sketch middle left). As result, hot fragments of the turbine stator segments exited. Obviously they have beek catapulted up to 6o meters by the expanding hot gases. Fortunately these did not penetrate the wing tank. So essentially it kept at the aeroengine failure.

During an overhaul the `bridges' undergo a penetrant inspection. Without signs and with acceptable wear the reassembly can occur. There is no life period for the `bridges' up to a failing or the rejection during overhaul (on condition parts).

History: Before the current failure already further 4 have been occurred. Also at these the fixing failed. In three cases it also came to the exit of fragments.
About 7 yers ago, the stud bolt and the `bridge' have been strengthened because of cracks and fractures. Additionally it was changed from the material Rene 41 to Waspalloy. So the strength (LCF and creep) at operation temperature should be increased.

Remedies:

• A favorable located plug of a borescope bore was as additional anti rotation locking strengthened and elongated.
• Ultrasonic testing of the stud bolts from the outside at the mounted aeroengine. In case a suspect part was found and the operatorhad no possibility, timed tight limited interim solutions are offered.
• During the overhaul, used fixings are replaced by new. A reassembly does not take place, even with a good indication.
• Also a single broken fixing per stage 3 and 4 is no more allowed.
• Introduction of a further, by design, improved fixing. To this belonged an about 15 % enlarged diameter of the stud bolt. Additionally the number of fixings was doubled. This version has proven in a related aeroengine type.

Fig. "Bolt fracture at hot parts by silver" (Lit. 23.3.1.2-8 and Lit. 23.3.1.2-13): At the operator this turboshaft aeroengine (sketch above) showed several abnormalities:

• Start problems at which the aeroengine did not accelerate above a low rotation speed („hung starts”). Thereby it came to an aborded take-offstart because of the danger of overheating (volume 3, Ill. 11.1-15.2 and Ill. 11.1-16).
• Vibrations at rotation speeds below (volume 3, Ill. 11.2.1.1-6).
• Unusual noises, which could be assigned the high pressure turbine rotor (gas producer).

For the clarification of the symptoms the aeroengine was disassembled. It showed, that the connection between the disk of the high pressure turbine (frame below left) and the stubshaft failed. Two of the five tightening bolts had been broken in the thread (sketch middle right), as consequence the remaining bolts are bended in different intensity and the threads are damaged. The bolts of a Ni alloy had about 3000 operation hours with about the same number of start cycles.

Investigation result: Both bolts have been fractured in the thread at the contact plane of the nut. The not fractured three bolts showed crack development in the root radius of the thread (detail middle). The fracture took place by crack propagation as cyclic fatigue.
A metallographic investigation of the incipient cracks showed, that these are filled with oxide and silver. On the thread sulfur was found. At all threads below the nuts, lubricant like residues with a `bead like structure' could be observed. These tiny beads have been analyzed in the SEM. They contain sulfur, silver, silicon and carbon.
The silver obviously exists in colloidal state.
The residues could be removed in water.
There are different explanations for the sulfur and silver containing residues on the threads:

• The silver derives from the silver plating of the nut. This would at the opinion of the investigators only expect silver in zones, which have been in contact with the thread of the nut. To this in the concluding comment a view will be given.
• The silver originates from a silver containing lubricant of the thread.Indeed in the manual of the OEM a lubrication of the thread is demanded. However synthetic aeroengine lubrication oil, which contains tricresylphosphate should be used. But the microanalysis showed no phosphor, which should be expected in this case. This speaks against a lubrication, according to the instructions.
• The sulfur in the residues at the thread could origin from the sulfur containing compounds of the bleed air from the compressor. In rests of the dust at the radius of the bolthead to the shaft, sulfur, natrium,and chlorine have been found. This is typical for sea atmosphere. At other locations of the bolt, like the shaft, there was no dust.
• The sulfur derives from a lubricant which was during the assembly applied at the bolt thread.

Conclusion of the investigation report:
The connection bolt fractured by a fatigue crack, which started from the root radius of the loaded thread turns to the nut.
The deposits in the thread originate from a silver containing lubricant, which does not correlate the demanded aeroengine oil. The sulfur could as well originate from the lubricant, as also from the cooling air.

Fig. "Locks and securings of bolts" (Lit. 23.3.1.2-14 and Lit. 23.3.1.2-17): During an overflight failed the (planet-) gear of the helicopter at the aeroengine side. This interrupted the power transfer to the main rotor. The pilot tried an emergency landing with autorotation. He touched the tree tips. The helicopter was catastrophically destroyed. A first investigtaion showed at the chip detector on the aeroengine side considerable metallic deposits. The concerned gear exhibited heavy damages.
Since the last overhaul the operation time for the aeroengine was about 400 hours. About 2 months before the accident, the gear had been disassembled because there where copper coloured chips in the oil filter of the fuselage side. There was no explanation for these filter residues. It was supposed, the chips can be assigned to the the pressure oilpump. It was exvhanged. A check found brass chips, caused by an assembly failure.
In the gear, the following failures/damages where found:

• The oilsystem was contaminated with copper respectively brass chips.
• The center gear (sun gear) was overheated. The engaget planet gear wheels had milled off all gear teeth. Two teeth obviously suffered (according to the results of the laboratory investigations) before an oilpump failure, caused by fatigue fractures. This could be concluded from separated tooth fragments and its flank condition. This pointed at an oil supply up to the fracture. The fatigue cracks could be identified as secondary failures of a derogated tooth engagement. The reason was an offset/misalignment of the shafts in connection with damaged shaft bearings.
• The gear wheel on the shaft of the low pressure rev counter had got loose. Groove nut and cup washer (sketch below right) missed. The groove nut was found in the accessory gear. The cup washer lay in the scavange sieve of the accessory gear. It showed a broken securing lug. The fracture obviously is in causative connection with a forced overload during the assembly. The wear track in the deformed zone let conclude at the gear assembly during the last overhaul. Such damages at the cup washers are known and are mentioned by the OEM in the assembly instruction.
  

Obviously the misalignment of the gear wheel had lead to the destruction of the driving counter wheel. This sits at the driveshaft of the aeroengine. The other damages are secondary failures. During the gear failure metal chips and abrasion developed. This triggered a jamming (seizing) of the scavenge pump. Therefore the drive shaft of the pump got overloaded. The oil pressure suddenly dropped and a power loss with overtemperatures occurred.

Two scenes for the failure sequence offer itself: 1. Scenario:

• During the last assembly, the securing lug fractured from overload.
• The groove nut at the radial shaft got loose, slipped back and separated with the securing lug.
• The loose gear wheel wore the multi spline coupling of the shaft and damaged the gear casing.
• Chips and abrasive got into the planet gear and lead to its destruction.

2. Scenario: A documented rework of the bearing seats in the gear during the last overhaul lead to misalignments of the shaft from the center gear in the planet gear. From the so caused vibrations, the safty lug of the cup washer fractured. The results correlate the 1st scene.

Remedies of the OEM:

• Design improved cup washer..
• Actualization of the assembly instruction for the fixing of the safety locking. This measure can be traced back at a former failure alread one year ago. At this, the safety lug fractured from a fatigue crack.
• Position marking of the grovve nut and gear shaft.
• Lubrication of the groove nut during the assemnly.
• Doubling of the torque of the groove nut.

Fig. "Avoiding fatigue cracks at bolts": In this case of a larger turboprop engine (sketch below), fractures at the transition radius to the bolt head at a flange connection in the planet gear (frame above) occurred. Concerned was vibration fatigue. At the bolts no failure causative weak points have been found. An examination showed, that obviously the pretension was not enough and it came to the lift-off of the flange. As remedy the pretension of the bolts was increased, what the static strength allowed. After this measure no more bolt fractures occurred.

Fig. "Problems with additional bolt locking" (Lit. 23.3.1.2-15): The aeroengine caught fire. The investigation showed, that the blaze started at the accessory gear. The gear box and the aeroengine cowl showed burnings, soot and through burned regions. The magnesium casing to the generator had burned and was molten. Thereby the gear wheels inside could be seen. The „Constant Speed Drive“ (CSD) and the generator are opposite positioned at the gear. The connection took a hollow shaft (frame middle right). The shaft was separated at the entrance into the generator.
About 1,5 years before this incident, the same CSD was repaired by an external shop.
An important question was, why the electric disconnection command, with the switch for the generator-CSD-connection, did not work. However, the examination of the electric actuation showed no malfunction.
In the CSD, the four attatchment bolts of the bearing support at the drive side had loosened. At the casing three of four bolts have been totally pulled out. The fourth bolt was hindered from dropping out by neighboured parts. Some of the bolts have been shorter than the CSD manual of the OEM prescribes.
A second CSD, which was overhauled by the same shop, was compared. It arose, that from the five attatchment bolts of the closing-off cover, also four had been shorter than in the manual demanded. Bolts of the bearing support and at other locations, although in the manuals not mentioned, had a securing wire. According to the OEM of the CSD there should no securing wires used inside the CSD, because the danger of FOD/OOD and constricted handling, should basically not be used. Bolts without securing wire had a curing fluid securing (detail below left), although this was not additionally to the self locking thread inserts demanded in the manual.

History: Within the last six years at relevant CSD types of airliners, 51 failures occurred. 37 lead to a unexpected landing and 10 to abortion of the start. In 9 cases the connection to the CSD could not be disconnected as specified. In spite of these seemingly similarities with the current case, its characteristics and those of the comparing part are apperarently not existing. However these could be assigned the repair shop of the CSDs.

Conclusion about the failure cause: The use of too short bolts and a bolt locking device which deviates from the manual, lead to the loosening and pull-out of the bolts at the bearing support, inside the speed control unit (CSD). As consequence the disconnection of the connection to the generator, which can be triggered in case of a failure, did not function.

Fig. "Failing of nosecone bolting", 23.3.1.2-13.2 (Lit 23.3.1.2-16): The failed engine was dismounted after the incident and sent for repair. Since the last overhaul, about 5400 hours with about 6400 start-shut down cycles have passed. The nose cone and the connection attachment went to the flight safety authority for a laboratory investigation. These found:
The four stud bolts/studs from a steel with a hardness of about 30 HRC, are corresponding the drawing, screwed without thread inserts into the magnesium casting of the bearing casing (sketch above right).
In the failure case they have been pulled out and stayed with the nut in the nose cone. The threads of three bolts are filled with sheared material of the connected casing. One stud („1”) looked different. Its thread showed heavy damages and smearings of the tips.
In the associated threaded bore, at the whole screwed in depth, the thread missed. This surface looked as if polished. The bores of the other bolts looked similar. Indeed the thread tips have been disappeared to a large fraction. However, rests of the thread turns still existed.
Metallographic axial sections through the bores of the casing showed, that in each case of nine thread turns, the thread was sheared off. The removal of the thread turns was especially pronounced in the bore of stud „1“. There were traces, which pointed at an axial movement, acting over a loger time. The other threads showed features of an additional bending.

Conclusion: The thread of stud „1” must have been markedly damged befor the screwing in. The magnesium casting is too soft, to damage the thread of the steel stud during screwing, although the thread of the casing was damaged. Obviously it could no more bear the operation load and pulled out. The vibrationmovements of the stud thread then produced the polished bore surface. As result the neighboured threaded bores have been overloaded and loosened. In the end phase additionally a bending load developed. In the bores of the casing, obviously already oftener damaged threads have been observed. They have been repaired with threaded inserts from stainless (Cr-?) steel.

Fig. "Loosening of bolting from hot parts" (Lit 23.3.1.2-12): A disassembly of the failed aeroengine with the investigation of the aviation authority showed:

• The flange connection between the disks of the stages 1 and 2 of the low pressure turbine had failed. The rotor stages had separated.
• The cavity between Turbine Midframe (Lit. 23.3.1.2-18, = TMF, see aeroenging sketch) and Low pressure turbine rotor was laced with notches from FODs/OODs.

In this region, before about 700 start-shut down cycles, assembly work took place.
In the cavity a smashed bolt (details left) of the turbine midframe (frame below right) and a multitude of fragments from the flange bolting have been found.

Technical description: In the turbine midframe, eight four cornered bolts, have been screwed in. In the history there were different vesions of bolts and anti-twist devices (details below). In all cases a wire locking is used. Does the locking wire break, it comes to loosening and untwisting up to separation of the bolt. From the closed cavity neither the bolt, nor smaller fragments can excape.

History: Alltogether several similar faulures emerged. These have been distributed at two types from the same aeroengine family. In all six cases, it came to the failing of the flange bolting between 1st and 2nd low pressure turbine stage. Always the separation of the rotor stood in causative connection with damages by foreigen objects/own objects. Mostly primary fragments of single flange bolts were concerned. However, three cases got known, where a leaven tool was found (Fig. "Foreign object remained by assembly" and Ill. 20.2-7.2). Obviously the danger of the locking wire fracture was early identified (25 years?) due to correspondent failures. It is the only locking device of the screw joint at the turbine midframe.
Therefore already years ago, a version with two locking wires was introduced (detail below). In the acute failure case such a double locking existed.
At an other aeroengine version, the bolting consisted (upper detail) of a persistent stud. Its head was secured by a tack weld. At the threaded side, a self securing nut sat.
The experience showed, that no locking version was sufficient safe against untwisting. For example cracks developed in the tack weld.

Usually, as consequence of a separation of the flange connection, `merely' fragments of the rotorblades exited from the aeroengine. However in such a case it came to the fracture of a LPT disk. It penetrated a tank and triggered a fire.

Measures: The responsible aviation authority demanded a supplement of the aeroengine manual. For every overhaul was requested:

• Inspection of the bolts/studs as well as thebolted parts.
• Tightening of the bolting.
• Renewal of the wire locking.

Comment: Obviously there are extremely high forces, which lead to the loosening of the bolts. They can be explained with an especially high operation load duribng a special service. This concerns:

• High cyclic thermal expansions and stresses (cracks in the tack weld).
• Large thickness of the bolted cross sections enables under themal gradients an time depending overload (creep strain) and a loosening.
• Heavy vibrations from the rotor and pressure oscillations in the shut cavity.

This is an impressive example, how difficult it can be to `sanify' a seemingly simple bolting.

Fig. "Identifying fractured bolts at rotors" (Lit 23.3.1.2-12): Foreign objects, especially bolt fragments, can trigger inside rotors extremely dangerous secondary failures (Fig. "Assembly influences at weak points" and Fig. "Loosening of bolting from hot parts"). Not only, that the damage of the bolts from inside located flange connections by the „zip effect“ (Fig. "Zipper effect at bolting interconnections") can lead to the spontaneous separation of the rotor assembly. Even from little notches of fragment impacts in highly loaded, life determining, disc zones fatigue cracks and fractures can start.
Therefore attention must be payed at noises, which occur during the maintenence and the rotating (starter, by hand) of aeroengine rotors at the ground and come from small loose fragments in a rotor drum (Ill. 20.1-7.1 and Ill. 20.1-7.2).
Unfortunately the experience shows, that in many cases fragments obviously get jammed/clamped and so no noises will be heared.