Aeroengine Safety

Figure "Surge causing blade fractures ": Compressor surges and limited compressor areas with stalls can cause a wide variety of damages throughout the engine. This diagram shows the locations of typical problems.

Overheating:

If the hot gases in the combustion chamber expand into the rear compressor area (Fig. "Stall problems with tight clearances" and Fig. "Operating loads on engine components by surges"), the compressor blades of the rear stages may overheat.

A special type of overheating of compressor rotor blades occurs during surges, when the lack of compressor airflow causes the power supplied by the turbine into the compressor to no longer be drawn continuously from the rotor. Agitation losses and friction heat up the blades (Ref. 11.2.1.2-3). The thin profiles and rear edges of modern compressor blades combined with materials with poor heat conductivity (titanium alloys, nickel alloys) promote overheating and permanent strength
losses. The part temperatures can be several hundred °C above the normal operating temperatures in localized areas. Titanium alloys can also become embrittled due to oxygen absorption.

This type of overheating is especially common in test compressors, since their operating conditions subject them to more frequent flow disturbances than normal. In addition, their power sources (electric motors or gas turbines) continue to supply full power even after the flow is disrupted (Ref. 11.2.1.2-3).

Hot parts are especially threatened by compressor surges. However, the overheating danger during a rotating stall is considerably lower than during surges. Due to the extremely small amount of air throughflow, the fuel/air ratio increases dramatically. In addition, the hot parts receive less cooling air. The overtemperatures in the combustion chamber and turbine are correspondingly high. In order to better understand the overheating effect, one should remember that for every mass unit of air involved in the combustion process, there are roughly three parts that are not involved. This extra air is required to guarantee acceptable temperature levels and an even temperature distribution in the hot gas flow ahead of the turbine, as well as cooling the hot parts. During a surge, there is only enough air for combustion.

Dynamic fatigue fractures in blading:

During a rotating stall (Fig. "Stall at compressor blades"), blades pass through the disrupted flow zones and are incited to high-frequency vibrations. If these resonate, they can result in dynamic fatigue fractures.
The physical forces during surges can be powerful enough to cause LCF-fractures in the blading after relatively few load changes.

Erosion of the blading:

During a surge, dust deposits, spalled abradable coating particles, and oxides from the expanding combustion chamber can all cause unusual erosion. This affects both the location (suction side and/or near blade root platform) and also the intensity of the material removal.

Contact between rotor blades and stators:

Pressure surges can axially deflect the rotor and blades and cause them to interfere with the neighboring stages (Example "Severity of compressor rotor rubbing "). Possible consequences include rubbing damage, nicks, blade failures, and titanium fires.

Flexural vibration of the rotor:

The asymmetrical pressure distribution around the rotor circumference during surges can lead to deflection and vibrations. These cause serious rubbing, during which stator vanes can damage the rotor drum (e.g. local overheating; Volume 2, Ill. 7.1.3-5 and Ill. 7.1.3-6) and rotor blades can wear unallowably large amounts of material out of the housing (increased clearances; Example "High pressure compressor blade fractures")

Bearing damage:

The extreme radial and axial forces acting on compressor rotors during surges, as well as pressure changes that prevent the necessary thrust compensation (insufficient cooling), can cause bearing overloads. This can result in violent fractures on the fastening lobes of the bearing rings or damage to the bearing track, which can later lead to material breaking out due to fatigue.

Labyrinth damage:

Rotor deflection can cause excessive material removal in the labyrinths, through the same process as at the blade tips. Typical consequences include deterioration (SFC increase; Volume 2, Chapter 7.0 ) and potentially damaging changes to the bearing loads in connection with changed pressures in the ring-shaped spaces that are responsible for the axial forces of the rotor.

Oil fires (Volume 2, Chapter 9.2):

Surges cause elastic deformation of rotors and housings, as well as compromising bearing chamber seals (labyrinth material removal), which can cause oil to escape and create oil fires. Pressure waves of the type expected during surges are suspected to cause oil fires in bearing chambers.

Spalling of coatings in housings:

Brittle coatings (such as ceramic coatings) and/or coatings that are weak or have low bonding strength (e.g. Ni-graphite abradable coatings) can spall due to the elastic deformations of the housings caused by pressure surges.

Figure "Stall problems with tight clearances" (Refs.11.2.1.2-1, 11.2.1.2-2): Compressor surges (also see page 11.2.1.1-5) are the most common type of compressor instability. A surge is a very high-energy, rotationally asymmetrical process (Ref. 11.2.1.2-2) that can be detected as a low-frequency vibration (530 Hz; Fig. "Preventing a compressor surge"). It is caused by a repeating flow disruption around the entire circumference. Detonation-like sounds with shock-like engine vibrations are typical of surges. This process occurs when the surge limit is reached, and its mechanism can be explained as follows:
Similar to rotating stalls, a local stall zone is created, but it spreads across the entire circumference and does not rotate. If the required compressor exit pressure can no longer be reached, it results in stall throughout the entire compressor. This creates a pressure wave that travels towards the engine inlet at the speed of sound (Fig. "Pressure shocks and vibrations during surges 2" and Fig. "Estimating engine loads during surges"), while the air throughput either ceases completely or even results in an airflow towards the inlet caused by the combustion chamber emptying itself “upstream” through the compressor (in extreme cases, with escaping flames). The decrease in pressure causes the airflow to resume its normal course so that the compressor again begins to move air. If an end pressure is reached that is above the surge limit, then the described process continues to repeat itself until the cause of the stall is removed.
However, there are also cases of unstoppable surges (lockin surge; Fig. "After burner triggered compressor surge"). These can be caused by, for example, serious compressor damage due to foreign objects or a fatigue fracture in a blade. Other possible causes include faulty regulators that do not react correctly to the pressure changes caused by a stall. In the case of a lockin surge during a rotating stall at low RPM, interrupting the fuel flow for at least 1-2 seconds (fuel blipping) can solve the problem. If this is not sufficient, then the engine must be shut down and restarted.
At higher RPM rates, the external surroundings must be altered. Typical measures include changing the thrust nozzle position or preventing inlet flow disturbances.
Understandably, lock-in surges lead to extreme dynamic loads with various consequences:

• Deflection of the blading that can cause interference of rotor and stator blades (1).

• Blade vibrations leading to blade failures (2).

• Rotor vibrations due to asymmetrical pressure distribution (Fig. "Pressure shocks and vibrations during surges 2"), resulting in dangerous rubbing.

• Powerful axial bearing forces created by the pulsating pressure build-ups and drops (Fig. "Pressure shocks and vibrations during surges 2").

• Extreme stress on the engine suspension due to shock-like forces and torsion moments (Fig. "Pressure shocks and vibrations during surges 2").

• Clearance changes at the blade tips and in seals. Pulsating forces on the rotors and housings cause dangerous rubbing. If large radial clearance gaps (3) are created by material removal, the surge limit will be permanently lowered.

• Overheating and erosion of the trailing edges of blades, due to the gas flow expanding forward from the combustion chamber, and/or an “agitation flow” (Fig. "Operating loads on engine components by surges") that transforms the thrust power of the turbine into heat in a very ineffective manner.

Figure "Pressure shocks and vibrations during surges 2" (Ref. 11.2.1.2-2): As has already been described in Fig. "Stall problems with tight clearances", compressor surges are accompanied by powerful pressure shocks. These act radially and axially on the rotor, stators, and housings.
The top diagram shows typical engine parts that are subject to high mechanical stress by surge shocks:

• Housings

• Compressor blading

• Engine suspension

• Main bearings

The shockwave from the development of the backflow towards the compressor inlet starts in one sector (Fig. "Estimating engine loads during surges"), which quickly expands to cover the entire circumference. This gives the surge process a one-dimensional character. The forces on the engine parts are caused by the initial asymmetrical pressure distribution relative to the rotor axis.

The diagrams show data taken from the high-pressure compressor of a large fan engine.
In diagram 1, the constant local pressure relative to the total pressure (constant + dynamic) at the inlet is given over each axial position (compressor stage). The solid curve represents the typical axial pressure levels. The dotted line shows the pressure levels in the shockwave that is spreading forward. The pressure changes after the stall are practically linear. For example, one can see that the pressure levels in the pressure wave in the fifth stage are roughly twice as high as the compressor pressure was before the stall. Depending on the size and distribution of the pressure wave, this pressure difference creates a powerful radial force at the circumference, putting flexural stress on the rotor. At the compressor inlet, the relative pressure increase (not the absolute pressure) is greatest (also compare with bottom diagram).
Diagram 2 shows the pressure changes in their axial positions during the blow down phase (Fig. "Stall problems with tight clearances"). After 5 milliseconds, the pressure increase due to the shockwave is at its greatest, and then rapidly decreases. At 50 milliseconds, the pressure in the shockwave along the entire axial length of the compressor is already below the pressure during normal operation. This shows that the actual dangerous powerful pressure shocks cover a time span of 10 milliseconds.
Diagram 3 shows that the pressure increase relative to normal operating pressures is especially great at the inlet. These values can also roughly indicate the mechanical surge loads on affected components.

Figure "Estimating engine loads during surges": Some guidelines for estimating the stress on components during surge shocks are given in accordance with Ref. 11.2.1.2-2. If the reader is interested in more specific stress calculations for a concrete case, he or she should refer to the indicated reference works.

Housing stress:

The pressure increase in the shockwave can put dangerously large amounts of stress on the compressor housings. The loads can be roughly estimated with use of the middle diagrams. However, the force distribution (on the rotor, as well) is very complex, corresponding to the makeup of the pressure wave. The pressure distribution depends on factors such as the surface condition of the blades (erosion), tip clearance size and distribution, and the pattern of flow disturbances at the compressor inlet. In general, it can be assumed that a dangerously large asymmetrical force typically lasts about two rotor rotations in compressors of large fan engines. This means that it clearly exceeds the time required for the shockwave to expand forward to the compressor inlet (Fig. "Stall problems with tight clearances").

The stress the shockwave puts on the housing depends on the shape of the housing and the temporal increase of the shock impulse. The impulse is influenced by the speed at which the shockwave spreads and the length of the compressor. Although the surge does not usually originate at the last stage, the force is calculated for the entire compressor length. Experience has shown that the expected resulting margin of error in the estimation of the impulse time is about 10%. The path of the shockwave relevant to the effects of the force is, at most, about 20-30% of the axial length of the compressor. This path is related to the different stagger angles of the rotor blades and stator vanes in each stage.
Estimation of the housing stress also demands knowledge of the advancing speed of the shockwave. The mach number of the shockwave (relative to the mach number of the normal air flow) is a function of the shock strength. This is represented by the local overpressure (diagram 4). Diagram 5 depicts an impression of the speed at which a shockwave spreads towards the compressor inlet. The relative mach number is 1 at the compressor exit and reaches a maximum at the inlet.
Diagram 5 shows how the mach number (M) of the shockwave speed changes over the axial length of the compressor. The speed at which the shockwave advances, v_shock can be estimated for every axial position by using the formula

v_shock=( M_{shock Rel}-M_flow ). c_local

Because the shock mach number increases about equally with the speed of sound (c_local = local speed of sound) towards the rear due to the temperature increase, the value of v_shock at the inlet is almost the same as it is across the entire compressor length. Based on this, the time of the pressure surge caused by the shockwave at all axial compressor locations is

D t_stress= Vcompressor length/v_{shock inlet}.

It is recommended to determine the overpressure ratio at the splitter without the fan (bypass ratio 0). The advancing speed in the fan can be calculated with the bypass ratio. The spreading speed of the shockwave determined in this way is slightly low, since the shockwave requires time to completely expand out of the compressor of the core engine into the fan. The short length of the fan allows assumptions of a constant expansion speed towards the inlet inside the fan.

Deflection of the rotor blades:

The force that acts to deflect the rotor blades is created by a combination of the shockwave and the backflow at the blading. Evidently, the complex conditions prevent exact calculation of these forces, so one is usually limited to qualitative estimations. The shockwave is partially reflected by the trailing rear edges of the rotor blades. The rest of the shockwave passes through the blading. The rotation speed of the blades increases the reflection of the shockwave. This reflection depends less on the shock itself than it does on the speed triangles at the rear blade edges. Typical circumferential speeds in modern compressors are close to the speed of sound. This means that increasing factors of 2.5 to 3.0 are plausible. In modern engines, the circumferential speeds are sometimes clearly above the speed of sound (450-480 m/s). The older literature from the early 80`s does not mention these conditions.

Engine suspension:

A consequence of the asymmetry of the flow is the stress axial forces and torsion moments place on the engine suspension. The axial force of the surge process F_surge (see diagram 6) is derived from the mass flows m_postS (post-surge), m_preS (pre-surge), and the average sonic speed using the following formula

F_Surge = m _preS . c (1- m_{post S} / m_preS )

F_Surge is directed rearward against the reaction force of the thrust. A conservative approach is sufficient for estimating the moment that the asymmetry of the surge puts on the engine suspension.
For this, it is assumed that the shockwave has spread across half of the circumference. The resulting moment is derived from half of the force of F_surge and attacks at the central point of the core engine.

Main bearings:

The main bearings (thrust bearings) of the engine are primarily stressed by the surge shocks. The bearing forces are made up of the the piston forces based on the pressure differences at the disk surfaces and blades (Volume 2, Chapter 7.2.1). Surges primarily mean rapid pressure differences at the compressor blades. The turbine is less affected due to the damping effect of the combustion chamber`s volume. The shockwave is too fast to greatly change the pressures in the ring-shaped spaces through the seal gaps (labyrinths). The resulting force is made up of the effect of the shockwave on the ring surfaces, rotor blades, and stator vanes (projection cross-sections). The estimated, rearward-acting, shock-like force on the rotor roughly corresponds to half of the total axial force from the surge. This total force is considerably greater than the single force from the reflection of the shockwave at a single blading.

Illustration 11.2.1.2-5(Ref. 11.2.1.2-2): During a compressor surge (Fig. "Stall problems with tight clearances"), hot gases from the combustion chamber can flow backwards into the compressor. This will probably cause the compressor blading to overheat. The thin edges and cross-sections of the blades in modern compressors, combined with the poor heat conductivity of titanium alloys and relatively poor heat conductivity of nickel alloys (in the rear stages), promote dangerous localized temperature increases. At especially high temperatures, structural changes that decrease strength are to be expected. Titanium alloys can additionally become embrittled through oxygen absorption, which causes a dangerous decrease in dynamic strength (especially resistance to FOD).

Figure "Damages by rotating stall" (Ref. 11.2.1.2-1): As a consequence of the reduced efficiency in areas with rotating stalls, the thermodynamic cycle process in gas turbines can deteriorate dangerously. In this case, the engine cannot be accelerated after it is started (also see Fig. "Typical problems during start-up"), i.e. the RPM level is “hung up” (stall stagnation, stagnation stall). If, after reaching the stall limit, the compressor enters the range of tertiary characteristics (Fig. "Axial compressors operating characteristics"), the cycle process of the engine deteriorates extremely. As a result, as the RPM rate decreases, the turbine inlet temperature increases rapidly, causing corresponding overheating problems. Modern, quick-reacting regulators (electronic/digital) can prevent or limit overheating damage by reducing the fuel flow sufficiently quickly. At this point, it is possible that the compressor returns to its primary characteristic after deceleration and the engine can be accelerated normally. However, if the compressor remains in the tertiary characteristic, i.e. the RPM rate decreases until a new, extremely inefficient balance is reached, then the engine must be shut down.
Every rotating stall subjects the blades to a slight periodic fluctuation in the static and especially the dynamic pressure as they pass through the stall cells. The frequency of these periodic pressure shocks depends on the rotating frequencies of the rotor and the stall cells, as well as the number of cells (top diagrams). The pressure shocks can be square or TRIANGULAR (bottom left diagram), which means that they contain many harmonic portions in their fundamental frequency. This increases the probability that the blades will resonate with one of their flexural or torsional natural frequencies. Expected consequences include dynamic fractures in the blade or blade root (bottom right diagram).

Effects of a rotating stall:

Limited stalls and, in extreme cases, stalls of the entire airflow can cause extensive damages directly and indirectly not only in the compressor itself, but also in many other areas of the engine (e.g. hot parts, bearings, etc.; Fig. "Surge causing blade fractures ").
If only small areas of the compressor are affected, a rotating stall may merely cause a small decrease in the air throughflow. In this case, there is no rapid temperature increase in the hot parts (mild stall, cold stall). If a rotating stall acts on the compressor blading for a sufficiently long time, then it can cause unnoticed damage which, during later operation, can result in spontaneous, catastrophic failure due to a dynamic fatigue fracture of a compressor blade.
During a deep stall, the stall cell is sufficiently large that the gas turbine cannot be accelerated after start-up (hang-up, Fig. "Turbine rotor overheating by 'hang up' during start"). It is no longer possible to increase RPM, and the engine stagnates. The performance balance of the compressor and turbine worsens to the point that the turbine can no longer supply sufficient acceleration power. In extreme cases, the regulator will supply overly large amounts of fuel until unallowable temperatures are reached at the turbine inlet, and there is immediate danger of extensive overheating damage.
Rotating stalls can cause flexural and torsional vibrations of the blading (Fig. "Compressor surge influenced by cannons"). Because smaller stall zones are not yet detectable from the outside, they can create unnoticed resonant vibrations, causing fatigue damage (blade failure) to engine parts.

Figure "Turbine rotor overheating by 'hang up' during start": Older engine types are outfitted with mechanical/hydraulic fuel regulators. If a compressor stall occurs during start-up or due to an operational anomaly such as a poor inlet flow or FOD, the hot parts, and especially the turbine blading, are at risk of overheating.
In the depicted case, the personnel at the testing rig were distracted during a final acceptance test, and overly fast acceleration was initiated. This resulted in compressor surges (see Fig. "Damages by rotating stall") in the lower RPM range (“hang up”, middle left diagram; also see Fig. "Axial compressors operating characteristics"). The regulator was unable to reduce the fuel flow quickly enough to correspond to the small airflow. Within seconds, the turbine blades overheated and sustained the damage shown in the bottom diagram. The resulting damage cost was almost equal to that of the entire engine.

Therefore, the following rule should be enforced:

Testing rig personnel must not be distracted (by visitors, etc.) during tests under any circumstances!

Further typical damage symptoms of overheated turbines and signs of overheating on hot parts suspected of overheating are depicted in Chapters 11.2.3.2 and 12.4.

The identification and analysis of overheating damage is extremely important, because in many cases, maintenance handbooks include specifications that make the continued use of certain hot parts dependant on the extent of possible damage. Misinterpreting apparent damage can be expensive and/or dangerous, as it can result in the scrapping of usable parts or continued use of unallowably damaged ones.
In modern engine types with electronic regulators (digital regulator) the danger of overheating during compressor stalls has decreased considerably. These systems evidently react sufficiently quickly and appropriately.

Figure "Overheating of blades by agitation losses during stalls": Through a process of heat input and heat removal, the normal airflow in a compressor leads to blade temperatures that stabilise near the local compression temperature. The temporarily interrupted airflow during a surge (Fig. "Surge causing blade fractures " and following) heats up both itself and the blading through friction and turbulence (agitation losses), because it does not remove any heat. The consequence may be very high local part temperatures.
Depending on the material involved, these temperatures can cause unallowable damage to the compressor blades in the shape of structural changes, strength losses, and embrittlement (Ref. 11.2.1.2-3).

Experience has shown that the thin (rear-) edges of rotor blades (top left diagram) are especially threatened. Discoloration and/or changes in surface shine (matt) are characteristics of overheated zones.
This type of damage is observed most often in externally-powered compressors in industrial applications and externally-powered test compressors (bottom diagram). The external power source inputs high amounts of energy independent of the supplied airflow.
Agitation losses also occur in ring-shaped ducts which are enclosed by rotating surfaces such as disks, spacers, and labyrinths (Ref. 11.2.1.2-18). This can heat up these parts considerably. In labyrinth seals with insufficient leakage flow, especially, dangerous heat levels may be reached. A temperature increase of over 100°C is entirely possible under poor conditions.
A similar effect can occur in vacuum testing rigs if air is let in during the shutdown process while RPM levels are still too high (top right diagram).
A similar overheating effect of the rear blade edges, which should not however be confused with agitation losses, is a backflow of hot gases from the combustion chamber during a surge.

Figure "Flutter problems at a fan" (Ref. 11.2.1.2-16): Fluttering vibrations (Fig. "Overloading blades by flutter" and Fig. "Types of flutter excitement") can cause blade fractures within seconds. This case concerns the unusual phenomenon of LCF-failures following high-frequency vibration (Ill. 12.6.1-17). These loads occur spontaneously, are extremely powerful, and, unlike resonance, the process by which they occur makes it impossible to control them sufficiently quickly with RPM adjustments. Experience shows that this can result in dynamic overloading of several blades.
Multiple blade fractures with comparable dynamic fatigue fracture locations, sizes, and high constant crack growth speeds (LCF-range) indicate that flutter was the damage mechanism.
Long, slender blades with thin profiles, as are found in the fan of fighter aircraft engines and front HPC stages, are especially at risk for this damage mechanism.
The diagram shows such a case. In a two-shaft engine (top diagrams), a clapper at 65% blade height was unable to prevent dynamic fatigue cracks in the fan blades of the first rotor stage (bottom left diagram). Cracks above the clapper were observed following testing rig runs at simulated high mach flight speeds. The blade profile was constructed for high tip circumference speeds. Instead, during flight operation in the flight envelope, the damage occurred at subsonic (flight-!) mach numbers. Unlike testing rig runs, in this state of operation, the flight altitude and inflow conditions at the compressor inlet evidently resulted in high mach speeds at the blade tips (bottom right diagram). Also, the angles at which the flow struck the blade profiles were in the stall range (Fig. "Overloading blades by flutter"). Contrary to expectations, the flutter was also influenced by the inlet pressure at the compressor. In line with the described considerations, the operating conditions and the damage symptoms indicated that the damage mechanism was stall flutter. This is primarily a torsional vibration with large amplitudes. Interestingly, calculations and earlier test runs had failed to indicate an aeroelastic problem. The flutter was prevented by reducing the flow angle and lowering the reduced inflow speed. In addition, the normal frequency of the blade was raised by optimizing the profile and changing the clapper position so that the causal resonance no longer occurred (Ill. 12.2.1.3-11).

Figure "Compressor surge influenced by cannons" (Example "Severity of compressor rotor rubbing ", Refs. 11.2.1.2-9 and 11.2.1.2-10): On-board weapons and rockets can temporarily change the temperature and pressure of the inlet flow of engines to a degree that it results in compressor surges. In addition, the danger of a surge increases considerably when remnants in the exhaust gases are deposited on the compressor blades. These deposits can unallowably change the roughness and profiles, worsening the aerodynamics of the blading (see Ill. 11.2.1.1-9). Of course, stalls are also promoted by other influences. These include increased surrounding temperatures and lower inlet pressures (higher altitudes). This should be taken into account in engine design by ensuring that the surge limit is sufficiently high. An important factor is proper positioning of the cannon relative to the air intake, since unimpeded ingestion of the hot gas cloud will certainly lead to surges.

Figure "Factors increasing intake temperature": The recirculation of hot engine gases is especially significant for aircraft that takeoff vertically (top diagram, Ref. 11.2.1.2-7). The design of the aircraft must address this problem. According to the cited literature, the inlet temperature into the compressor can increase to roughly 15% of the exhaust gas temperature, i.e. for a typical single-shaft in-hull engine (turbojet) this is above 140 °C, but only about 30 °C in a in-hull turbofan engine. The significance of these values becomes clear when one considers that increasing the inlet temperature by about 7°C results in a thrust loss of 3 - 4 %.
However, even commercial aircraft and helicopters can ingest their own hot gases under certain circumstances, such as during thrust reverser operation or poor wind conditions (see Fig. "Situations for intake of hot gas").
During catapult takeoff from aircraft carriers, fighters can ingest hot steam. This causes unstable compressor behavior and surges (Ref. 11.2.1.2-6). In this case, three effects are especially important:

• disturbing the distribution of the inlet temperature at the compressor (middle diagram).

• uneven distribution of the physical gas properties at the compressor inlet (gas constant R, specific heat)

• ingestion of liquid condensation

A stall-inducing pressure distribution was not found while investigating catapult takeoffs with steam ingestion. Stalls may be promoted by a change in the compressor clearance gaps caused by the high g-forces during catapult takeoff (see Volume 2 Chapter 7).
Another influence which potentially promotes surges seems to be related to the creation of a ground vortex (see Volume 1, Chapter 5.2 and Fig. "Ground vortex caused by thrust reverser").
One solution was using the bleed valve in the high-pressure compressor, which is usually used in the investigated engine type to prevent stalls during start-up and rapid acceleration.

Figure "Situations for intake of hot gas": The bottom diagram shows a four-engine commercial aircraft with large first-generation fan engines. In this case, activation of the thrust reversers caused compressor surges in the outer engines (also see Fig. "Compressor problems by ground vortex"). These ingested hot gases from the thrust reversers of the inner engines (Ref. 11.2.1.2-8).
The top diagram shows a large helicopter. While the engines were powered up in overly strong tailwind on the ground, their hot gases and those of the APU were ingested in intolerably large amounts.

Figure "Effects of flow disturbances over long distances": In the development phase of the large first-generation fan engines, the pylon of all three competing engine types interfered with the fan airflow, causing the fan blades to experience dangerous vibrations (Ref. 11.2.1.2-9). In one case, the pressure increase ahead of the pylon reached far enough forward to affect the core engine (i.e. high-pressure compressor) via the splitter. This promoted compressor stalls.
In other engine types, the fan blades were incited to dangerous vibrations.
The problems were solved in different ways:
In order to improve the operating stability of the compressor, the flow disturbance was minimized with the aid of varying exit angles of the fan exit stator vanes near the pylon (bottom right diagram, Ref. 11.2.1.2-9). The left diagram shows another solution. In this case, the axial spacing of the interfering pylon section and the fan rotor and exit guide vanes was increased.
In addition, improved sculpting of the pylon cross-section may have been helpful.

Figure "After burner triggered compressor surge" (Ref. 11.2.1.2-11): Due to insufficient air inflow, the afterburner of the first series of a fighter engine in the 1980`s (developed in 70`s) had problems igniting and/or maintaining stable operation at high altitudes (about 13,000 m) at speeds of about 830 km/h (Example "Unfavourable afterburner conditions at high altitude"). The critical range is depicted in the flight envelope (bottom diagram).
During a delayed or repeated ignition of the afterburner, the fuel cloud that had formed in the afterburner ignited. This resulted in a low-speed detonation, the pressure wave of which was especially intense when the thrust jet was still closed. The pressure wave spread out forward into the fan (middle diagram), causing stalls in the fan and in the high-pressure compressor, which could not be resolved sufficiently quickly through the usual method of suitably reducing the power (stall stagnation, hung stall, Fig. "Axial compressors operating characteristics"). The insufficient air supply during a surge causes the afterburner to extinguish again. This allows a cloud of fuel to form and ignite as soons as the compressor resumes operation and enough air flows into the afterburner. This process can repeat itself several times. In this case, it caused serious fire damage in the hot part area. Evidently, in several cases, attempts to restart the engine after it came out of the surge were unsuccessful. This unique and “stubborn” type of stall was observed especially during the transition from single-shaft (turbojet) engines to fan engines (turbofan) in military aircraft. The stall is promoted by the space between the fan, core engine, and bypass. Apparently, the time when the fan and high-pressure compressor come out of the stall is also important. This can be understood in connection with the pressure waves during surges in the high-pressure compressor (Fig. "Stall problems with tight clearances" and Fig. "Pressure shocks and vibrations during surges 2"). These create an underpressure pulse that causes a rotating stall in the high-pressure compressor. The stall is ended only after the RPM rate is decreased below the idling rate.
It is odd that while this phenomenon occurred during the development phase, the danger it posed to serial operation was not adequately considered.
The following measures seem to have alleviated the problem:

• (1) Lengthening the splitter towards the front in order to make it more difficult for the pressure wave from the afterburner to enter the high-pressure compressor via the bypass duct.
• (2) Adjusting the afterburner fuel regulator so unsuccessful ignition attempts in the afterburner result in a reduction of the fuel flow rate, even if the pilot has the throttle set to maximum power.
• (3) Adjusting the electronic thrust jet regulator so that the thrust nozzle opens after a failed afterburner ignition attempt.The available older literature does not indicate whether opening the thrust jet shortly before afterburner ignition was considered as an alternative solution.
  

This is the method in use today. However, at higher bypass flow ratios this method can be problematic, since it lowers the pressure in the afterburner, making successful ignition difficult.

Figure "Compressor problems by ground vortex" (Ref. 11.2.1.2-12): Depending on crosswind, the roll speed of the aircraft, use of thrust reversers (Fig. "Ground vortex caused by thrust reverser"), and engine power (see Volume 1, Chapter 5.2.1.2), a ground vortex may be created (right diagram, Ref. 11.2.1.2-17). The determining factors are the ratio of the engine diameter to the distance from the ground, and the air intake rate (see Volume 1, Ill. 5.2.1.2-5, Ill. 5.2.1.2-6 and Ill. 5.2.1.2-7)

A ground vortex can pose a threat to compressors in several different ways:

• initiation of a surge through local pressure disturbance in the inlet flow.

• the shapelessness of the inlet flow near a ground vortex (right diagram) can cause the blades of the fan and the first HPC rotor stage to experience dangerous vibrations.

• a ground vortex is capable of sucking massive foreign objects up from the ground.For these reasons, a certain amount of effort and expense may be justified to prevent a ground vortex from forming (Volume 1, Ill. 5.2.1.2-9).
  

Figure "Compressor damages due to missing 'bell mouth'": In older-model turbojet engines that were used for de-icing runways (top diagram), serious compressor damage originating in a dynamic fatigue fracture of a blade in the first rotor stage occurred with unusual frequency. An investigation revealed that the affected engines were operated without bellmouths. During a damage analysis, an engine on a testing rig was also operated without a bellmouth. It was observed that the sharp housing edge at the compressor inlet caused a serious flow disruption with turbulence in the intake flow (bottom diagram). Observation of the engine inlet with a stroboscope showed that the rotor blades of the first stage began to vibrate so heavily, that their tips appeared to touch one another. Under such extreme dynamic loads, dynamic fatigue fractures in the blades could be expected within minutes.

The conclusion: Never operate an engine on a testing rig (such as a field testing rig) if it has no bellmouth or a bellmouth that is not specifically for use with that engine type.

Figure "Fragments blown into compressor from combustion chamber": In the case of serious compressor damage, the entire blading can fail spontaneously in such a way that the expanding gases from the combustion chamber blow blade fragments forward out of the front of the engine. This phenomenon was observed in a case where a helicopter engine was struck by ice (Volume 1, Ill. 5.1.4-5).
Several influences that can explain this phenomenon may act in combination:

• the blades create a frontally directed ascending force which pulls them forward.
• the increasing pressure levels towards the back of the compressor throw the blades forward.
• destruction of the compressor blading allows the gas in the combustion chamber to expand in a forward direction.