Findings that allow conclusions as to the damage sequence
The external damage symptoms of an engine can already provide important information regarding damage sequences and special stresses. Engines are subject to many damage-relevant external and internal induced stresses (Fig. "Forces") that can influence the development of typical damage symptoms. Knowledge of these stresses is a valuable tool for the accident investigator.
It is also important to remember that engines and engine parts that, at first glance, seem beyond analysis due to complete destruction can often yield important clues upon closer inspection (Fig. "Fragments"). Therefore, regardless of the state of destruction, a complete and thorough inspection of all fragments should be made with the goal of finding characteristics relevant to the damage. This provides the opportunity to clear up even apparently “hopeless” cases (Fig. "Hopeless case").
The resulting findings are part of the collection of facts, the first step in a systematic damage analysis (Fig. "Damage analysis").
Almost every investigator will relatively have a first working hypothesis regarding the damage sequence. However, it is important to be aware that this hypothesis is merely temporary and to self-critically question it. Otherwise, hastily formed opinions can even prevent recognition of important facts (Fig. "Facts").
This type of situation is more likely if the investigator is not aware of important effects relevant to the damage, e.g. special characteristic engine part stresses (Fig. "Component stress").
Further findings that can be relevant when evaluating mechanical regulators or analog mechanical indicators are discussed in a separate chapter (4.2, Operating Data).
Figure "Forces": The forces that affect an engine and its components can be roughly grouped into two categories: external and internal. These forces can influence one another and act in combination.
These are primarily forces that are transferred across the engine suspension into the wing or nacelle and correspondingly stress these parts.
Propulsive forces: these are created by engine thrust. In flight, these forces act in the direction of flight, at least in standard commercial transport aircraft.
During landing with thrust reversers, these forces act against the direction of flight. The forces that occur during thrust reversal are usually considerably lower than the propulsive forces during flight. In military aircraft, adjustable jets and flaps can redirect the exhaust gases and create propulsive forces in different directions. Examples are VTOLs and aircraft with thrust vector control.
If aircraft are equipped with afterburners, then the activation of these increases thrust by up to approx. 40% of normal (“dry”) thrust.
If engines must be shut down in flight, then their face surface creates considerable resistance even though the engine is windmilling. In addition to the worsened flight conditions (asymmetrical thrust), this may require a thrust increase from the remaining engines.
Aerodynamic forces: The engine nacelles of externally mounted engines are subject to noticeable forces from the airstream. With modern large fan engines, these forces can be around several 10,000 N. In this way, the nacelles can contribute significantly to the lift of the aircraft.
In case of heavy turbulence (see Chapter 5.1.6), aerodynamic forces can combine with inertial force to cause high dynamic stressing of the engine suspensions ( ).
Inertial forces from flight maneuvers: In commercial aircraft, these forces can reach dangerous levels due to turbulence or emergency measures such as rapid descent.
In military engines, these forces are typical and result from the intended flight missions. Normally, the maximum allowable acceleration is determined by the tolerance of the pilots.
High G-forces can be expected during hard landings and extreme deceleration (e.g. landing on an aircraft carrier). The relevant regulations establish limits for this.
Gyroscopic forces: If the axis of rotation of a rotating mass is deflected from the plane of rotation, it creates gyroscopic forces. This kind of situation usually results due to flight maneuvers. Gyroscopic forces act as bending loads on the rotor and can even have an external effect if they are not internally countered by reverse rotation of the different shaft systems.
Forces due to the acceleration of the nacelle: Flying through turbulence can cause the wings to vibrate, which heavily stresses the engine suspensions affixed to them (see Chapter 5.1.6).
Gas pressure: Gas pressure creates internal forces in various ways. They stress the engine housing like a pressure cooker and the stages act in the same way as pistons under different pressure levels with corresponding axial forces. These can often cancel each other out within a shaft system (rotor). Forces that are not canceled out appear as external forces.
Aerodynamic forces: The compressor blading pushes and compresses the air flow, the turbine blading takes the energy from the hot gas flow to power the corresponding compressor. This energy transfer occurs through aerodynamic forces on the blade profile. This affects these primarily in the form of bending stress.
A further type of aerodynamically induced forces are dynamic forces that occur in rotating seals
due to the clearance gap flow. These forces can lead to unallowable vibration of individual components (rotor and stator), as well as the affecting the complete rotor system.
Vibrations: All parts of an engine can be made to vibrate. This is not a problem as long as these vibrations remain within the intended range. Vibrations become dangerous if they exceed the values that were designed for or not considered. Examples of this include resonance and large imbalances.
Vibrations also occur in combination with aerodynamic forces (e.g. blade flutter).
Centrifugal forces: The rotor components are subject to large centrifugal forces. These loads are important values for design and determine the life span of the parts. The most important factor for fatigue is not so much constant high RPM, but rather start-up/shut-down cycles.
Inertial forces (G-force): In the literal sense, many forces including centrifugal force can be classified as inertial forces. However, this case refers to forces that act inside the engine and are caused by external G-forces. For example, these can cause the housing to bend and rub against the rotor.
Forces from thermal strain: Impeded thermal strain, i.e. thermal strain that is impeded by neighboring components or parts, creates forces and corresponding (heat) stress.
A typical example is the high heat stress in rotating and static hot parts due to temperature gradients. If this stress changes along with a change in operating conditions, it causes low-cycle fatigue stressing of the engine parts.
Gyroscopic forces: Even if gyroscopic forces in reverse-rotation shaft systems usually cancel each other out, expected effects include internal elastic deformation of the rotor and connected housing sections, as well as considerable bearing forces.
Figure "Fragments": A less knowledgeable individual will stand helplessly before the remains of a crashed engine if he is supposed to make conclusions as to the damage sequence and causes from “this hunk of scrap”.
To the experienced expert, however, this is a challenge with realistic chances for a solution. It is not without reason that experience is listed first. Experience consists of various kinds of knowledge:
This knowledge makes it possible for the expert to determine whether or not there is a possibility of a specific type of damage having occurred. The evaluation of operating data such as RPM from the external damage symptoms is also possible and is covered in Chapter 4.2.
Fundamentally, it must be attempted to recover all engine parts and fragments in such a way that no further damage is done to them that would make analysis more difficult.
The analysis of these parts then takes place under suitable conditions.
Even if the cause seems immediately apparent, the engine should be closely inspected for anomalies in both its complete state and disassembled. This is also true for parts that are not immediately identified as relevant to the crash (see Fig. "In flight damage symptoms").
Experts can gain important first indications as to the damage sequence and causes from millimeter-size anomalies. For example, a dynamic fracture can be recognized by certain characteristics. Therefore, it is important to inspect all fractured surfaces for these characteristics. If in doubt, a microscopic inspection must be conducted.
Figure "Damage analysis": The term damage analysis is often confused with damage inspection. Therefore, it is important to understand these fundamental terms.
A systematic damage analysis is a prerequisite for understanding the causes. There are many educational systems on offer from different companies. Most of these systems include three important steps, even if some characteristics differ:
Fact collection includes the damage inspection of the hardware. However, it also includes the software, such as damage statistics, quality records, manufacturing processes, and other technical documents.
The accuracy of the fact determines the quality of the investigation results. Facts are beyond doubt. They must be tested as intensively as possible through critical questioning. If uncertainties remain, further verification and testing is necessary.
Experience has shown that the multitude of viewpoints and interests often consciously and/or unconsciously turn hunches into facts.
-Formulation of hypotheses:
This is primarily a creative process. Basically, all hypotheses from the parties involved in the investigation are documented without any judgment. It is important that controversial or apparently different hypotheses are not neglected.
Of course, very early, maybe even before the investigation begins, every expert has his or her own working hypothesis that is somewhat necessary for systematic, directed action. However, these hypotheses must not affect the process in such a way that they might falsify the results. It is important that the investigator is at all times able to critically examine his or her own hypothesis and perhaps discard it for a more plausible one if there are well-founded considerations.
-Testing hypothesis with the facts at hand:
All hypotheses are tested based on the fact at hand with regard to their plausibility. This is done in repeatable, written procedure.
Important: if a fact, even a seemingly minor one, goes against a hypothesis, then the hypothesis is wrong! There is no limbo in this evaluation!
Example "Older factures" (Fig. "Facts"):
After several combustion chambers exited the engine of a twin-engine fighter aircraft during flight (Fig. "Facts"), tearing 1 sq. meter holes, the aircraft was still able to land and the engine was taken for investigation.
There was noticeable deformation of the thrust jet. Even though the combustion chamber housing was torn wide open as shown in Fig. "Facts", all “investigators” were standing around the thrust jet. All activity was concentrated around this area. They assumed that an explosion had occurred in the afterburner with a large pressure increase that blew apart the combustion chamber. A typical “favorite hypothesis” with an apparently logical consequence.
Of course, this hypothesis presupposed that the combustion chamber housing only exhibited fresh fractures. Either the investigators were incapable of correctly assessing damage symptoms, or they did not even make an effort to inspect the fractured surfaces, since the hypothesis only allowed for a fresh fracture.
An expert who was then called in immediately recognized that the burst combustion chamber housing played an important part in the damage sequence. A brief visual inspection of the fractured surfaces revealed characteristics of an older (oxidized) fracture. This was a sure sign that this area was the cause of the damage.“
a good, involved expert will orient his or her actions in line with a working hypothesis. For example, he or she will ensure that certain engine parts are treated with extra care, and that no unallowable cleaning or dissection procedures are used.
However, the working hypothesis must not hinder the search for all anomalies that may be relevant to the damage and contradict the working hypothesis.
It sometimes happens that certain inspections are not conducted or certain parts are not included in the investigation because a preconceived opinion in the form of a “favorite hypothesis” deemed this unnecessary.
As described in Example "Older factures", the disregarding of facts may happen unconsciously.
Sufficiently broad technical knowledge and practical sense can prevent this from occurring, at least with regard to characteristic part stresses and the interrelation of parts during operation. It has often proven to be problematic when a specialist from a single field (e.g. combustion chamber expert, strength expert) is called in to oversee an investigation, since this runs the risk of every problem being seen only in the context of this narrow field. In this instance, the following adage holds true:
“For the man who holds only a hammer, every problem resembles a nail.”
Additionally, having certain assignments that are causally related in some way to the damaged engine part can cause a conflict of interest. An engineer would much rather find a material defect to be the cause of damage than his own design. A “materials expert”, on the other hand, would be less likely to see a flaw in the material selection as the cause of the problem (e.g. misestimation of the corrosion strength).
The task is to select these specialists and to use their expertise at the right time.
Figure "Component stress": As mentioned in the description of Fig. "Facts", the knowledge of characteristic engine part stress is important for damage analysis. This knowledge alone is not enough, however, unless the investigator also develops a “feeling” for it that allows him or her to estimate the effects in practical application. For example, while a strength specialist can simulate all effects on his or her computer, it does not guarantee that he or she would be able to apply this knowledge to an actual damaged engine lying before him or her.
Examples of these important effects and part stresses:
Axial forces on compressor blades:
The pressure buildup in the compressor leads to considerable, forward-directed axial forces acting on the blading. This can be better understood by imagining that the blade annulus is a ring-shaped piston surface, the back side of which is subject to increased pressure due to the compression. In large fan blades, these axial forces are absorbed by special constructive measures, such as threading of the nose cone. These threaded connections are highly stressed.
Forces acting on turbine stator assemblies:
Similar to the conditions in a compressor, but corresponding to the pressure decrease in the hot gas in the opposite direction, the turbine stator vanes are subject to strong axial forces acting in a rearward direction. Since the pressure decrease occurs across fewer stages than in the compressor, the pressure difference in front and behind a stage is greater than in a compressor stage. The pressure differences in the high-pressure turbine stage are especially large. The axial forces is equal to several tons in the intake stator assembly. This force must be transferred by the blading into the supporting housing structure.
Turbine stator assemblies are also affected by considerable circumferential forces caused by the gas flow redirection. These forces must be absorbed by massive fastening pins to prevent rotation of the stator assembly (Example "High internal pressure").
Internal pressure against the housing:
Engine housings are thin-walled pressure cookers without lids or bottoms. The lid and bottom are provided by the pressure buildup in the compressor. The combustion chamber housing is the most highly stressed (Example "Circumferential forces") and is subject to the highest gas pressure in the engine. In modern engines, this pressure can be up to 40 bar (40×105 Pa).
Excerpt: During take off, an uncontained failure of the left engine occurred, an examination revealed a fatigue fracture through a rear flange bolt hole of the combustion chamber outer case at the three o'clock position of the engine.
Comments: This damage illustrates the high internal pressure in the combustion chamber housing and its flange connections.
Excerpt:“The aircraft lands uneventfully after the right engine failed during a descent. The shutdown occurred after failed anti-rotation pins allowed gas loads to drive normally stationary third stage vane clusters in the low pressure turbine area. Rotation of the clusters in their channel allowed the clusters to machine themselves out through the low pressure turbine case. Evidence indicates that 23 of the 44 third stage vane anti-rotation pins ruptured followed by fracture of the remaining pins from shear overload. The pin stress rupture was the result of bending stresses on the vane pins from vane twist & pin material creep. A design task was initiated in 1986 to provide a more durable pin and (the OEM)…released new material, Inco 901 non-stepped, anti-rotation pin…to replace the…tinidur pin.”
Comments: Frequently are the high circumferential forces on the turbine stator caused by the gas deflection not aware (see also Ill. 126.96.36.199-8).
Figure "Hopeless case" (Example "No hopeless cases"): There are no hopeless cases, and anyone who begins an inspection with the attitude that it is hopeless is not suitable for the task, which requires a great deal of imagination and optimism.
It is important to search for characteristics that can not be explained by the impact. These are especially those characteristics that take longer to develop than the time span of the known damage sequence or impact (Fig. "In flight damage symptoms"). In some cases, symptoms that can only have occurred in flight are especially worth noting.
Example "No hopeless cases" (Fig. "Hopeless case"):
A compressor stall occurred in a fighter aircraft during flight. After several attempts, the engine could not be brought back up to normal operation without a stall, and the pilots ejected. The aircraft crashed into a cliff face at almost a right angle. The engine was compressed to a fraction of its original axial length. The rotor blading showed no sign of rotation at the time of impact (see ). Evidently, the engine was not capable of sucking in foreign objects at the time of impact. This allows the following apparently trivial conclusions.
All damage symptoms on the engine that would require a rotating rotor or working combustion chamber (see Fig. "In flight damage symptoms") must have occurred during the damage sequence in flight. These symptoms include, for example:
During the inspection, pin-sized metallic drops were found on the turbine blading that had come from the compressor blade material and could not have been created during normal operation. This was the first indication of compressor damage during flight.
Further inspections revealed that the damage had been caused by a dynamic fatigue fracture of a blade in the front compressor rotor stage, which was caused by inadequacies during repairs.
It is assumed, that experienced investigators are capable of selecting relevant recommendations from the following list for the case at hand. The order of these recommendations does not imply importance or sequence.
Figure "Parallel cases": Cost and time pressures frequently result in the argument that parallel cases no longer have to be exactly documented because they have already been investigated once. Experience shows this to be a fundamental mistake. It has been shown, that only the investigation and documentation of many parallel cases makes possible collection of detailed findings. This in turn makes realistic assessment of the damage-causing influences, risks, and the development of effective solutions possible.
4.1-1 J.W. Purvis, “Manufacturers have important resources that can be put at the disposal of investigators”, Paper at the Asia/Pacific Regional International Society of Air Safety Investigators (ISASI) Seminar, Tokyo, May 1996.
4.1-2 NTSB Identification CHI87IA100, microfiche number 37163A.
4.1-3 NTSB Identification ATL94IA097.