Table of Contents
23.5.2 Failures of Pipe lines
Always again pipe lines stay causative in connection with aircraft accidents. Thereby the special danger of fretting strikes (Example 23.5.2-1). It develops at contact faces with other components. This situation worsens with the introduction of pipe lines from titanium alloys. They replace lines from the usual CrNi steels (example 23.5.2-3). In the fretting zone, especially at titanium alloys the fatigue strength drops heavily (Fig. "Risk of titanium pipe lines 2", volume 2, Ill. 6.1-8). Not to underestimate is the failire potential of the multiuse of P-clamps (Fig. "Resonance vibration caused by P-clamps" and Fig. "Welds in components under thermal fatigue"). Its function and safety depends crucial from the assembly. A further weak point seem V-bands. They serve as simple connection element of flanges (Fig. "Leakage of accessory devices").
Example 23.5.2-1 (Lit 23.5.2-2, Lit 23.5.2-9, Lit 23.5.2-10 and Lit 23.5.2-17):
The tiltrotor airplane (sketch above) practiced landing approaches with the autopilot. Thereby, in the last
phase at nearly 300 km/h approach speed, the tilting of the aeroengines begins. Suddenly, in seconds, the
rotor blade control and the power of the aeroengines
changed. After about 30 seconds the aircraft got
bumpy and gets into a firm flight attitude.13 km from the schduled landing place it crashed.
The following investigations showed, that a hydraulic line of the propeller actuation from the left aeroengine has ruptured. With this, the actuaton pressure dropped.
The wall thickness of this line from a titanium alloy for an inside pressure of about 345 bar is with 0,56 mm remarkable thin. It was touched by electric lines and worn through. Thereby an about 5 mm long crack developed. The touch of lines with each other is promoted by the constricted room of the aeroengine nacelles. The effect of the leak at the actuating of the propeller could, not sufficient fast corrected. This was due to problems with the backup systems.
Further dangers establish sharp edged damages of the pipes at contaminated contact/mounting faces of clamps and the tensioning during assembly (Lit. 23.5.2-25).
Comment: Here the failure cause is explit traced back at the friction of a titanium pipeline on other lines. This case has a similarity with a case which dates back about 30 years (eyample 23.5.2-3). It is interesting, that obviously electrical lines, in spite of plastic insulations are sufficient for dangerous wear.
Example 23.5.2-2 (Lit 23.5.2-1 and Lit 23.5.2-14):
During a transatlantic flight in minutes periods at about 13 000 m hight, both aeroengines of a large airliner (sketch above) extincted. Concerned was an `ETOPS-flight'. For this the drop out of both aeroebngines should be ruled out. The cause was fuel starvation. The airplane had to approach an alternate aerodrome without running aeroengines. For this an about 20-minutes gliding flight was necessary. During the hard landing 8 of the 12 landing gear tires burst.
The fuel starvation was caused by the leak of a large low pressure fuel line. A problematic fuel management lead to the fuel outage at both aeroengines. The leak developed because of friction contact of the fuel pipe line to the high pressure pump with a hydraulic line.
The lines have been modified, obviously after already former wear failures had occurred. In the case on hand, five days before the accident, simultaneously with a changed pipe version, the exchange of an aeroengine took place. Thereby it came, despite the warning of a technician, to the combination of not scheduled pipe versions. This lead to the contakt of the pipes with intense fretting.
Comment: As far it can be seen, the pipe material concerned is CrNi steel. The friction contact was such intense, that the cross section of the fuel line was weakened up to rupture. Probably the wall thickness of the low pressure fuel line was thinner than these of the hydraulic line. Remarkably is, that the wear process needed only 5 days.
Example 23.5.2-3 (Lit 23.5.2-3 and Lit 23.5.2-18):
1970 already during the 2nd testflight of a prototype from this fighter (Fig. "Dangerous situations by air duct leaks", sketch above) crashed. About 25 flight minutes after the start the accompanying aircraft observed a smoke plume behind the fighter. Apparently vaporised hydraulic fluid was concerned.
A following accident investigation showed, that two thin high pressure hydraulic lines from a titanium alloy had failed because of vibration fatigue. This was the consequence of a resonance vibration of the lines during idle of the aeroengine. Laboratory tests showed, that the fatigue fractures developed in less then 10 seconds.
After this, the testing with the 2nd prototype was so long delayed, until the titanium lines had been replaced by lines ot CrNi steel.
Comment: The short date of origin of the fractures identifies these as high frequency LCF (!; Volume 3, Ill. 12.6.1-6). This correlates an extremely intense vibration load in the region uf cyclic plastic deformation.
The modification of the fighter aircrafts to lines from steel, permits the conclusion rather at a basic problem than merely a resonance. This should have been corrected with relatively simple measures. Thinkable would be further mountings (P-clamps) and/or an additional damping.
Unfortunately the available papers contain no hints at contributory effects like fretting or sharp edged damages (Fig. "Risk of titanium pipe lines 1"). However, it can be supposed, that the risks became aware with the accident. They stay in connection with the fretting problem of titanium alloys (Fig. "Leakages on pipe lines" and Fig. "Risk of titanium pipe lines 2"). This would be a plausible reason for such a modification. Obviously Offenbar also later series versions had pipe lines from CrNi steel. This reveals the dimension of the problem. It affected longtime the design philosophy (configuration, stresses) and handling (maintenance, assembly) of pipe lines. Up to the turn of the millenium, no applications of titanium pipes in series use in aeroengines emerged. Not until in the newest fighter airplanes we find aeroengines with pipe lines of titanium alloys.
Fig. "Dangerous situations by air duct leaks" (Lit. 23.5.2-6, Lit. 23.5.2-9, Lit. and Lit.23.5.2-16): Although `only' air was
concerned, failures with leakages at bleed air ducts/lines can have dangerous consequences. The
bleed air of the rear compressor stages is needed for the cabin supply, cooling of the turbine blading and the
turbine casing. At modern aeroengines it possesses pressures above
40 bar (newest aeroengines in development up to 70 bar) and temperatures up to
600 °C (example 23.5.2-5). Under these conditions at a
critical crack length, an explosion like failing of the
line must be expected. Even smaller leaks can
overheat neghbouring parts with air temperatures of several hundred °C.
Even if no fires develops, during exit of hot gas a fire warning should be triggered. This result of a hot gas exit at least an inflight shutdown, must be expected (example 23.5.2-4 and example 23.5.2-5).
Obviously certain lines are especially susceptible.
- Electronic and elekctrical lines (example 23.5.2-4).
- Near positioned tanks for fuel or oil and hoses can during overheating fail and leak. The hot air can easily ignite these flammable liquids (example 23.5.2-6).
Example 23.5.2-4 (Lit 23.5.2-16): Shortly after lift off, during the start, there was fire alarm at the
aeroegine No. 2. The airplane returned and landed safe.
The investigation arose, that the two-piece V-band at the casing to the 14th compressor stage had partly loosened. So the hot air with a temperature of more than 600 °C could exit. It damaged the outer insulation of electrical cables.
At the circumferential fracture surface from the ruptured semicircle of the clamp, a fatigue fracture is concerned. Ihis pointed at a cyclic axial load. It showed, that the fatigue overload of the clamp is a secondary failure. Causative was the failing of relief clamps. These limit the load of the clamp. Obviously the safety wires failed from wear. So the nuts of the clamp boltings could twist loose.
The OEM published an instruction to all operators of the concerned aeroengine type. It dealt with the securing wires of the clamps.
Comment: Seemingly the safety wires did touch in a manner, that they where worn through by vibrations. This obviously was no single case. Probably therefore the instruction of the OEM refered to an unproblematic arrangement of the security wires.
Example 23.5.2-5 (Lit 23.5.2-6): Shortly after the start the pilot reported a fire at the right
aeroengine. After this an emergency landing took place. An investigation showed, that the
V-clamps (band clamps) at the high pressure bleed valve to the aeroengine had been loosened. They had
not been sufficient tightened during exchange of the valve before the flight.
Comment: This case obviously didn't deal with a fire. The alarm was triggered by the hot air.
Example 23.5.2-6 (Lit 23.5.2-9): A reconnessance airplane of the U2 type crashed, because
of a fire. Cause was the bleed air of the aeroengine which hit a fuel tank.
Comment: The aeroengines of the concerned aircraft type are integrated into the fuselage. This promoted the situation.
Fig. "Resonance vibration caused by P-clamps" (Lit. 23.5.2-11): The investigation showed, that the
fuel leak occurred at a broken dumping
line. A similar fracture of the line already developed at the same aeroengine during a flight
three days befor! The position was the same and had the same failure
features. Then the airplane returned to the depart airport. There the line was exchanged. Also at an other operator, already two
parallel cases existed. However, these concerned in each case different aeroengines.
The pipe line consists of stainless CrNi steel and is screwed with a ring line (detail sketch middle left). This distributes the fuel to the injection nozzles. In the heat affected zone at the transition of the weld to the end screwing (Fig. "Pipe line crack by tensioning" and Fig. "Fatigue cracks starting from inside"), a HCF fatigue crack (volume 3, Ill. 12.6.1-6) has developed. The picture of the fracture is typical for a one sided dynamic bending load.
Three P-clamps serve as support of the line (frame below). Its position was optimised in tests by the OEM, to minimize the vibration load of the line. As interlayer/bandage wrappings with teflon tape have been used. These act damping at vibrations and should prevent fretting at the line (Ill. 23.5.2-10). All three teflon bandages showed unusual heavy wear. This partially penetrated up to the pipe surface. The ends of the tape did not lay firm at the line. They have been frazzled and took off. Obviously the wrapping was very loose. There were besides the bandages at two spots markedly wear patterns. These indicated, that also other components came in contact with the line. This argued for deviations during the laying of the line.
A check of the form of the line showed slight deviations from the nominal geometry. The cause for this remained unclear. Also the position of the clamps was slightly offset, compared with the sheduled. A check of other aeroengines showed, that such deviations frequently exist without the development of a failure.
For the understanding of the failure causing vibration load the OEM carried out tests. Especially concerned was the resonance exitations of the line. These correlated the typical low frequency, however not usual combustion chamber fluctuations (rumble, volume 3, Ill. 11.2.4-11). Such vibrations can be transferred at the pipe line from the combustor casing, as well through the pipe ends as also the P-clamps. With these tests the positions of the P-clamps have been optimized for a low vibration load. Because these improvements obviously have been estimated as low, also the pipe line geometry was changed.
Conclusion: Obviously the dynamic load of the pipe line is aeroengine individual. It depends from scatters of the exitation frequency and the exitation intensity, as well as the natural frequency of the pipe line.
Comment: Most of the attention of the OEM at the pipe line support let suggest a weak point since long time. The unusual heavy combustor vibrations obviously could not be avoided.
In this connection it should be remembered at the influence of the fuel properties at the combustor vibrations in modern aeroengines (volume 3, Ill. 220.127.116.11-4.1). Even if these are inside the specifications, they can act failure causative. An unfafourable combination of component tolerances can promote the susceptibility of an aeroengine for such vibrations. Thereby e.g., the fuel injection or the flow through cross section of the turbine nozzel behind the combustion chamber may be concerned.
Fig. "Welds in components under thermal fatigue": A weld always means also a
notch (volume 3, Ill. 13-18), even if it is specification
conform. It acts as shape notch. The
fatigue strength of the weld material (casting structure, weak points)
and/or of the heat affected zone, is also lower as of the base material. With this the deteriorating effect of
a vibration load intensifies (Fig. "Fatigue cracks at pipe line weldings"). For pipe lines additional the weldseams frequently are
positioned at the pipe ends. They are the connection to the screwings. These zones are especially loaded from
the usual bending vibrations of the pipe lines (Fig. "Causes of pipe line vibration").
In the displayed case, it came to the fracture of a scavenge pipe (detail). So hotgas could get into the oil system and ignite there an oil fire. This continued inside the pipe over a longer distance.
The pipe is guided through a strut of the turbine exit casing to the bearing chamber. Differnces in the thermal expansion between the strut and the scavenge line, as well assembly caused tensionings can, especially to the pipe ends, trigger a dangerously high prestressing. With the vibrations of the combustion chamber, the fatigue crack can be explained.
Note: During the laying of the securing wires attention must be payed, that there is no loosened, vibrating contact with other components.
Fig. "Danger by a rubbing safety wire" (Lit. 23.5.2-2 11 and Lit. 23.5.2-2 12): During the flight at a military aircraft it came
to the extinction of the aeroengine. Cause was the leak at the distribution line to the fuel
injection nozzles (sketch right). With this the fuel supply failed.
The leak developed through a fatigue fracture (sketch below left), which started at a deep wear mark (fretting, sketch above left) at the fuel line. Obviously the wear mark was caused from a vibrating safety/securing wire. How dangerous wear during contact with safety wires can be, shows also the example 23.5.2-4. In this case, safety wires from nuts have been worn through. The nuts loosened and it came to a dangerous secondary failure.
Fig. "Assembly caused tension at pipe lines" (Lit. 23.5.2-13): During the aeroengine testing it came to an emergency landing.
Cause was a long circumferential crack at an aluminium elbow of the fuel supply to the high pressure
pump of the aeroengine. An investigation showed:
A production failure in the region of the seal at the connection to the fuel pump lead during the assembly to the pretension of the line. The so increased medium stress lowered the usable fatigue strength (Fig. "Pipe line crack by tensioning"). This pipe line underlied pressure shocks in the fuel during the actuation of the afterburner. Such pressure shocks act because of the deflection of the fuel at the elbow of the lines (Fig. "Fatigue cracks starting from inside"). Already after fifty of such shock loads the failure occurred. With this a LCF failure was concerned (volume 3, Ill. 12.6.1-6). As reaction to this incident, a design change was carried out:
- The aluminium elbow with the, for this material typocal low fatigue strength, was replaced with a pipe from CrNi steel.
- Changing of the contour from the line. Probably a `disarming' of the deflection.
- The pump was changed in a manner, so that dangerous pressure shocks are prevented.
23.5.2-1 D.Learmount, „A330 lands safely after gliding for 20 minutes“, Zeitschrift „Flight
International”, 4-10 Septermber 2001, Page 32.
23.5.2-2 “Wachsende Probleme für die Osprey”, Zeitschrift, Flug Revue“ Juni 2001, Page 68.
23.5.2-3 „Grumman F-14 Tomcat”, Zeitschrift „Air International“, January 1982, Page 28.
23.5.2-4 J.Neff, “Bad tubing grounds Osprey - Problems with Titanium hydraulic lines have plagued the innovative aircraft for years”, www.Newsobserver.com, November 29, 2003, Page 1 -3.
23.5.2-5 M.R.Brendt, “Briefing on V-22 Accident by Maj.Gen. Berndt”, www.defenselink.mil., April 5, 2001, Page 1 -33.
23.5.2-6 NTSB Identification MIA89IA016, „Airbus Industrie A-300-600R, Incident OCT-24-88”, Page 1.
23.5.2-7 „Turbine Engine Hazard - Incompatibility of Chloride Based Materials and Stainless Steel or NK Titanium Based Components“, www.casa.gov.au, AAC/Part-1/1-013.HTM, Page 1.
23.5.2-8 Zeitschrift “Aviator Aviation”, Volume 4 Issue 4, April 2001, Page 1-5.
23.5.2-9 D.M.North, “Pilot Selection Process Arduous”, Zeitschrift „Aviation Week & Space Technology”, April 12, 1999, Page 65.
23.5.2-10 “PWA PT6T3B Fuel controlling system pipe cracked”, Ref. 510003464, Flight Safety Australia November-December 2006,“Selected Service Difficulty Reports ” , Page 53.
23.5.2-11 Australian Transport Safety Bureau (ATSB). Investigation Report No. 200505952 from November 2006, „ In-flight engine fuel leak…Boeing Company 747-438…18November 2005“, Page 1-7.
23.5.2-12 W.D.Pridemore, „Introduction to Gas Turbine Engine Failure Analysis”, Ohio State University Class Lecture 4-26-2006, Page 9.
23.5.2-13 “F-16 / 101 Fuel System Modified”, Zeitschrift „Aviation Week & Space Technology“, February 23, 1981, Page 24.
23.5.2-14 P.Koring, “Air Transat executive confirms warning from mechanic”, www.iasa.com.au, September 2005, Page 1-6.
23.5.2-15 R.K.Smyth, „Flugerprobung Grumman F-14 Tomcat”, Auszug in Zeitschrift „Interavia“, 1971.
23.5.2-16 AAIU Synoptic Report No: 2006-006, AAIU File No: 2004/0029, Published: 24/4/96. “Airbus A330-301, Incident 4 June 2004”, 5 Pages.
23.5.2-17 J.Neff, “Bad tubing grounds Osprey - Problems with Titanium hydraulic lines have plagued the innovative aircraft for years”, www.Newsobserver.com, November 29, 2003, Page 1 -3.