9.1.2 Damages from Metal Fires
Metal fires in engines are ignited primarily through the ignition of titanium structural parts (titanium fires) in the compressor. In a few cases, fires in magnesium alloys have also been reported (Example "Magnesium dust catching fire"). Titanium fires occur in the engines of both tactical and commercial aircraft. Their frequency of occurrence is about once every 106 flight hours. In line with experience and technical writings, the duration of titanium fires is between 5-20 seconds. Aside from indirect consequential damages (e.g. fuel fires), the consequential damages of the fire occurred during this short time.
Titanium fires can limit themselves to small regions of the compressor (Fig. "Damage symptoms of limited titanium fire"). These localized fires occur when, for example, a blade fragment strikes the housing wall. The blade fragment can ignite (Ref. 9.1.2-1) and melt onto or through the housing. Rotor blades burn primarily at the tips (Fig. "Titanium fire burn test"). This is understandable if ignition occurs due to rubbing, but even burning melt moving centrifugally out from the rotor tends to ignite the blades tips. This type of fire usually extinguishes after the melt is thrown off along with the energy it contains.
In many instances, however, the engine region behind the ignition point inthe flow direction suffers extreme damages. Experientially, titanium fires do not spread against the direction of flow. Only an overheating of the trailing edges of the preceding stage is observed. A situation that leads to this type of damage can be caused, for example, by a blade fragment jamming between the rotor blade tips and the housing. If the fragment rotates along with the blades, the heat of friction can cause ignition along a large part of the circumference.
The implementation of suitable materials increases the probability of titanium fires being contained in the housing (chapter 9.1.3). However, this also often means accepting an increase of damage inside the engine. This damage is causedby the burning melt being directed into the hot parts. The melt reacts with the nickel alloys, low-melting phases are created (Fig. "Melting temperatures of Ti-oxides") and the turbine blading is melted.
Recognizing a titanium fire from the damage symptoms:
One must be fundamentally aware, that titanium fires cannot ignite or be sustained through impacts or fires on the ground. If symptoms typical of titanium fires are found in parts of a damaged engine, it must be assumed that the damage occurred during flight. With the exception of multi-engine tactical aircraft, that fly at high speeds close to the ground, it must further be assumed that the fire occurred while the engine was running, since the conditions for igniting and sustaining titanium fires are only present at this time (Fig. "Titanium fires risk of flight mission").
There are characteristic damage symptoms of titanium fires that extinguished shortly after ignition, before reaching sustained burning. The beginning of a titanium fire, i.e. an ignition, is indicated by concentric tarnishing, surface oxidation (rough grey surface, embrittlement), and/or splatters of metal or heat cracks. If the burning process was already initiated, burn traces with melted grey oxidized edges surrounded by yellowish and grey oxides will form in the direction of the flow (Ill.9.1.2-5).
Housing breaches caused by titanium fires can be recognized by their undeformed, superficially fused, oxide-surrounded edges (Fig. "Characteristic traces of titanium fires").
Primary damages that cause titanium fires (see chapter 9.1.1):
Titanium fires do not spontaneously occur during normal engine operation. There must always be primary damage or unusual operating conditions in order to reach the ignition temperature (Fig. "Causes of titanium fires"). Therefore, when a titanium fire has been identified within the framework of an aircraft accident, it is just proof that a primary damage such as blade failure or foreign object damage caused serious rubbing. From this, the actual investigation as to the causes begins.
Consequential damages of titanium fires:
Even if it is extinguished after a short time (Fig. "Damage symptoms of limited titanium fire"), a self-sustaining titanium fire leads to a failure of the engine or to unacceptable operating trouble (e.g. compressor stall). Therefore, in single-engine aircraft a crash is unavoidable. Even if the titanium fire is contained in the engine by the housings, the aircraft can be indirectly endangered if important load-bearing engine parts (e.g. the suspension or supports of the main bearings) are weakened to such a degree that a catastrophic failure occurs.
If the housings are breached by a titanium fire (uncontained), the aircraft is in immediate danger (especially in engines inside the fuselage). A outwardly-directed housing breach is more likely, if the regulator increases fuel flow after a loss of pressure in order to maintain compressor rpm. It is especially dangerous, if consequential damages include fuel or oil lines developing leaks (Fig. "Primary and secondary damages of titanium fires") and causing a more extensive fire of longer duration. Damage to adjustment systems (e.g actuators), electrical wiring and/or mechanical adjustment cables (feedback cables in older regulator systems) can also have disastrous results.
Contained titanium fires, especially, can result in extreme overheating in the hot part region due to the burning melt itself and as a result of the additional heating-up of the air flow.
Figure "Characteristic traces of titanium fires": Titanium fires leave characteristic traces that differ from those of fires on the ground (e.g. after a crash). While plastic deformations (usually inwardly directed from outside) can be expected from an impact force, a hole created by a titanium fire through burning, flame cutting, and melting (Ill. 9.1.2-8), will not be plastically deformed (bottom left diagram, Fig. "Pipelines penetrated by titanium fire"). It is possible that the “ballooning-effect” (Ill. 9.1.2-8) caused by the internal pressure tore open the housing wall in an outward direction. In this case, the inner wall of the housing has to show traces of a titanium fire, such as splatter of metal and traces of heating-up. The detail view shows typical symptoms at the band edge. These include melted structures, defined oxide formation, and splattered, hardened titanium droplets.
If traces of a titanium fire are found at an accident scene or in a damaged engine, it proves that a primary damage occurred in the compressor region during flight.
Figure "Primary and secondary damages of titanium fires" (Ref. 9.1.2-1): A titanium fire can directly and indirectly cause consequential damages.
Small melt splatters (clearly smaller than a pin head, top left diagram) can severely damage the affected surface in the same manner as welding splatters: localized overheating (oxygen absorption, loss of hardness, residual stresses), fusing to the surface (loss of hardness, embrittlement), embrittlement due to oxygen absorption, crack initiation, and stress concentration. Titanium alloys are especially sensitive to these types of damages. Because the compressor blading is constantly under fatigue stress, this type of damage increases the risk of fatigue fractures in the blade leaves. In this way, splattered particles from a starting ignition can lead to extensive consequential damage and engine failure.
Even limited titanium fires roughen up the compressor blading (detail at top left), change profiles, and enlarge cracks that impair operating behaviour, all of which result in an unallowable drop in performance and/or a stall. The resulting lack of cooling air and/or increased fuel flow increase the risk of the hot parts overheating. Rotating blades throw the burning melt off in such a way, that the damages are concentrated in the tip region. Because the gap at the tip has a strong influence on the compressor`s degree of efficiency, even small damages are clearly noticeable.
In guide vanes, centrifugally moving melt causes undercuts primarily near the foot. This promotes fatigue failures in the leaf region of guide vanes.
If pronounced titanium fires are contained in the engine, the melt and heated-up gas flow cause particularly extensive overheating damage in the hot part region (melting down, burning up). Further damages include increases of roughness by adhering splatters, diffusion, melting-on, changes in the flow cross-sections (e.g. plugging), and/or changes in the profiles.
If a titanium fire breaches the housing, a normal firewall designed to contain oil or fuel fires provides insufficient protection. It cannot prevent the damaging of important components on the engine or a parallel engine (mounted components such as regulators, fuel and oil systems, wires and cables, hydraulics) or damage to the airframe (load bearing structure, tanks, steering)(Example "Most common cause of titanium fires"). Therefore, even in multi-engine aircraft, titanium fires can result in crashes. However, it is evidently possible to use mats (even retrofitted) to provide acceptable protection from titanium fires that breach housings (Example "Most common cause of titanium fires").
The bottom diagram schematically shows the risks inherent in an escaping titanium fire. The risk is especially great in tactical aircraft where both engines are aligned next to one another in the fuselage (bottom left diagram). In engines mounted directly under the wing (second from left), such as in larger supersonic aircraft (bombers, commercial aircraft), the danger of a fuel tank being damaged is considerably higher.
Figure "Pipelines penetrated by titanium fire" (Ref. 9.1.2-1): The top diagram shows a CrNi steel fuel line that was perforated when struck by Titanium splatter. The formation of a low-melting phase should have promoted the perforation. This example shows the concentration of energy in a burning drop of titanium.
The bottom diagram is of a compressor housing after an uncontained titanium fire and shows several fuel lines that have been burned through and other typical symptoms of a housing breach (the edges of the hole are undeformed and surrounded by light oxides (Fig. "Characteristic traces of titanium fires").
Example "Probable titanium fire " (Ref. 9.1.2-3):
Excerpt: “…(the aircraft) suffered an in-flight engine failure…shortly after departure from Heathrow Airport….According to airline sources, engine gas temperature went `sky high' and the level of vibration prompted the crew to shut down the engine immediately and return to Heathrow. The engine was found to have considerable damage in the high-pressure compressor (HPC) `around stages 5 and 6' and was transported to (the OEM)…for closer examination. It is unclear whether the engine encountered mechanical problems or suffered foreign object damage….Meanwhile,…(an other Airline) found cracking in the sixth-stage HPC blades of its…engines…“
Comment: Even though a titanium fire is not explicitly reported, all factors indicate that one occured (extremely highy gas temperatures, short duration of the damage sequence): the failure of a rotor blade in the compressor region, where the fire is suspected to have begun, would be a typical cause of a titanium fire and thus a plausible explanation (see Fig. "Causes of titanium fires").
This type of blade damage becomes probable if, as in the other case mentioned, cracks form in the blades.
Figure "Titanium fire ignited by blade fracture": The primary damage that most frequently results in a titanium fire is the failure of a compressor blade (see Fig. "Causes of titanium fires"). Depending on the weak points specific to the engine, primarily rotor- or stator blades can be involved in the ignition process. However, not many blade failures in the compressor result in titanium fires. It seems that, in order for a titanium fire to ignite, several special conditions must be met. In larger titanium fires, the ignition point is usually completely destroyed and no longer analyzable. Therefore, the damage sequences depicted here are based on observations of the damage symptoms in limited titanium fires (Fig. "Damage symptoms of limited titanium fire") and/or experience with weak points of certain engine types and should be taken hypothetically. If one assumes that ignition was initiated by an abrasion cloud ( ), the following typical scenarios may apply:
Guide vane failure: Phase “1”: guide vane failures due to fatigue fractures tend to happen near the “clamping point”, i.e. above the foot platform mounted on the housing wall. If there is no inner shroud, the failed blade leaf jams between the rotor (spacer) and stator, leading to dangerous rubbing. The resulting “abrasion cloud”(“2”) ignites the neighboring blade in the direction of rotation as in Fig. "Ignition of titanium fires by rubbing" and then the entire blading in the direction of the gas flow (“3”).
Rotor blade failure: In this case, fatigue fractures usually occur above the foot platform near the rotor (Phase “1”). The leaf lies flat against the housing and is run over by the other rotor blades and/or drags along the housing. In this case, the guide vanes of the following stage exhibit melt zones typical of self-extinguishing titanium fires (Phase “3”, Fig. "Damage symptoms of limited titanium fire").
Figure "Damage symptoms of limited titanium fire" (Ref. 9.1.2-1): The top diagram shows the typical damage symptoms of a limited titanium fire. The impact area of the rotor blade fragment can be distinguished clearly. A region where rubbing occurred originates here, and was probably created when the leaf fragment was dragged by the rotor blade tips. There is a section of the following stator assembly behind the impact area where several neighboring blades have burned down to the foot platform and the surrounding blades have typical undercuts that start at the leading edge (see bottom diagram). The ignition of the blades was most likely caused by the rubbing of the blade fragment that was jammed against the housing (see ).
Figure "Damage mechanisms of titanium fire ignition": Several cases of titanium fires igniting in the labyrinth region have been reported (Fig. "Titanium fire by labyrinth rubbing"). The rotating rings of these labyrinths were made of titanium alloys and the pressure differences and/or the leak flow speeds reached critical values for ignition (Fig. "Titanium fire ignition in labyrinths"). Unarmored labyrinth tips are especially prone to fire. In this case, several factors promote ignition (see chapter 7.2.2):
- Because there is no rough armor to enable cutting and the titanium is in direct contact, a great deal of frictional heat is created.
- Heat transfer directly into the labyrinth tips
- unsuitable combination of materials. E.g. a titanium tip against a honeycomb seal made of nickel alloy (top diagram) or against a rub coating containing Ni (bottom diagram, Fig. "Weak points at compressor housings for titanium fires").
Depending on the labyrinth, the rubbing can be accelerated by heat strain and loss of hardness in the rotating labyrinth ring (Fig. "Labyrinth rings bulging by local heating").
Figure "Titanium fire by labyrinth rubbing" and Figure "Titanium fires damage at compress rotor": In the development phase of this older engine type, an extensive titanium fire occurred that was evidently caused by an intermediate stage labyrinth made of a titanium alloy. The damage sequence was reconstructed as follows (center diagram):
The rub coating on the inner shroud of the compressor stator assembly was made of a thick “fill layer” and a thin rub coating (bottom detail). The fill layer was too thick (3 mm) for the sprayed-on material that was used and tended to separate due to the high thermal stress.
The radial clearance was only 0.5 mm and was easily bridged by a localized separation of the layer.
When rubbing occurred, both the unarmored labyrinth tips as well as the thin ring cross-section quickly heated up (Fig. "Damage mechanisms of titanium fire ignition").
The rings located farther back in the rotor showed that the filigreed T-shaped labyrinth rings tend to expand when heated due to heat strain and mechanical overstress due to centrifugal forces (see diagram of the damaged rotor).
In this way, self-perpetuating rubbing led to the ignition of the stator assembly above it (Fig. "Damage mechanisms of titanium fire ignition"). This can be concluded, because titanium fires do not spread against the flow. Consequentially, the compressor blading including the Ni-alloy blades in the rear compressor region were burned up, i.e. melted off (Ill. 9.1.2-8).
Figure "How a titanium fire burns through a housing wall": The greatest danger lies in titanium fires that escape the housing. There are several damage sequence models relevant to the perforation of housings:
“Ballooning” (top left diagrams): this damage sequence requires overpressure typical for compressor housings on the fire side. In Phase “1” the housing wall is heated up by the fire. Local thermal expansion and loss of hardness cause it to bulge outwards (Phase “2”). The bulge (balloon) breaks open in Phase “3” and the fire escapes.
Burning through (top right diagrams): In this case no overpressure is necessary on the fire side. The stream of fire strikes the housing wall (“1”) and heats it up until it softens (“2”), in some cases supported by the formation of a lower melting-point alloy with the titanium melt (Fig. "Melting temperatures of Ti-oxides"), and breaks through the wall that has been melted open (“3”).
“Cutting”(bottom diagram): some damage symptoms indicate this damage sequence. First, a housing under internal pressure is perforated (burned through or “ballooned”). The walls of this hole continue to burn in a radially outward direction. At the same time, the air flow carries burning drops in a circular pattern corresponding to the edge of the hole (bottom left diagram). These drops strike the next wall (e.g. a partition wall between two parallel engines or a tank wall), causing it to soften and possibly fuse to it, so that a flat section of wall comes loose (bottom right diagram).
Note: for effects of a hot gas stream on a housing wall see chapter 9.3
Figure "Titanium fire damage at a compressor housing" (Ref. 9.1.2-1): Titanium fires almost always occur in a region specific to the engine type. In the above case of a military triple shaft engine, ignition occurred in the intermediate-pressure compressor. The fire then escaped directly behind the swan neck of the wind tunnel to the high pressure compressor (see Fig. "Weak points at compressor housings for titanium fires"), although its location along the circumference seemed to be coincidental. The fire perforated several housing walls (bottom diagram).
Figure "Weak points at compressor housings for titanium fires" (Ref. 9.1.2-1): similar to bird strikes, “swan neck tunnels” seem to promote the escape of titanium fires (top right diagram). The effect which radial gaps in the housing (top detail) can have in these cases must not be underestimated. Burning titanium melt gliding along the outer contour can be caught at these points and ignite the exposed edge of the slit (bottom diagram). Further weak points that can promote this phenomenon:
Ni-graphite abradables that form a low melting-point phase with the melt and support the burning process.
A flammable titanium alloy edge or an abradable coating that fails to hinder, or perhaps even encourages, fires (for remedies see chapter 9.1.3).
Figure "Titanium fire by blade failures" (Ref. 9.1.2-3): Evidently, this twin-jet tactical aircraft type was prone to titanium fires that could be traced to fatigue fractures in the high-pressure compressor blades (bottom diagram, arrow; Example "Most common cause of titanium fires"). In order to protect the parallel engine, the firewall between the engines was equipped with additional insulating mats.
Excerpt from Ref. 9.1.2-4: “The Navy has grounded roughly 25% of its …(aircraft of the affected type) following a series of uncontained fires in the twin-engine aircrafts….engines. A Marine Corps…(aircraft) crashed in the Pacific Ocean…Preliminary indications are that the aircraft's engine caught on fire after suffering a compressor blade failure….All… engines with more than 800 hr. flight time will be removed and replaced with spares and engines with fewer than 800 hr. The engines' outer ducts will be fitted with a new fire retardant rubber mat, a process that could take up to nine months….
Compressor blade failures can result in fires, and fire retardant material will be provided…The company's (OEM) technicians also will carry out associated modifications to the engine's high-pressure compressor…
The company will send three technician crews to …an operating basis to apply an additional retardant coating to the firewalls …“
Excerpt from Ref. 9.1.2-5: ”…The Service grounded aircraft with engines over 8000 hr. when several crashes were traced to uncontained engine fires following titanium compressor-blade fatigue failures…
All high-time compressors should be fixed by…Modifications include redesign of the first and third compressor stage blades to avoid fatigue, and redesign of the lever arms controlling the variable stators.
In addition, a Viton fire retardant mat will be bonded to the inside of the …(engines) outer bypass duct. This will contain any titanium fire long enough for the blades to melt and pass through the engine, rather than through the duct and into aircraft as before.”
Comment: This spectacular case reconfirms blade failures as the most common cause for the ignition of titanium fires. The affixing of mats on the inner side of the bypass duct and on the fire wall between the engines (see also Fig. "Firewall safety") is worth noting. Evidently, these mats are not made of normal rubber, but of polyfluoroethylene (PTFE). It is astounding, that the toxic and extremely aggressive flouric vapours expected to result from a fire were acceptable.
9.1.2-1 T. Uihlein, H. Schlegel, “Titanium Fire in Jet Engines”, Proceedings AGARD-CP-587 of the AGARD conference “Aircraft Fire Safety”, 14-17 October 1996, chapter 25-1 to 25-12.
9.1.2-2 T.R. Strobridge, J.C. Moulder, A.F. Clark, “Titanium Combustion in Turbine Engines”,
Report Nos. FAA-RD-79-51 and NDSIR 79-1010 , July 1979, about 100 pages.
9.1.2-3 I. Sheppard, “British Midland V2500 fails after only 150h”, magazine “Flight International”, 27 May-2 June, 1998, page 19.
9.1.2-4 “Navy Grounds F/A-18 Aircraft Following Engine Fire Incidents”, magazine “Aviation Week & Space Technology”, November 23, 1987, page 31.
9.1.2-5 “GE pays to fix F4o4”, magazine “Flight International”, 23 January, 1988, page 62.