Aeroengine Safety

Figure "Cracks originating at the surface" (Ref. 17.2-11): The surface of a part is the potential weak point. This primarily applies to dynamic fatigue cracks, but creep fractures and corrosion-related cracks (SCC, IGC) also tend to originate at the surface. The reason for this behavior is that nearly all damaging influences such as media and heating affect the surface first, and most strongly. The part strength at the suface is negatively affected by the fact that there are larger stresses here than in the volume of the part. This means that damage-related lower strength and greater brittleness are more significant. This combination has an especially noticeable negative effect in notches. In addition, because the stresses at the surface are not three-dimensional, this area has different deformation behavior than the inside of the part (Volume 3, Ill. 12.3-3)

Stresses: In many cases, parts are subject to flexural stresses that place the greatest loads on the surface due to their triangular distribution. Force transmission with the aid of positive fitting or friction can put considerable stress on the surface. A frictional connection will result in powerful shear stress (Volume 2, Ills. 6.1-9.2 and 6.1-11).

Notches and scratches: Notches increase local stresses, regardless if they are part of the design (e.g. cross-section jumps), the result of mechanical damage, or raised weld seams.

Surface damages: These can have two different types of negative influences. They can act as notches in the form of material removed by wear or corrosion. Another possibility is surface damage resulting in a loss of strength. This includes fretting damage to titanium materials, which can cause a decrease in dynamic strength of up to 70 % (Fig. "Fretting damage loweing by shot peening").

Tensile residual stresses: These lower the usable dynamic fatigue strength by raising the mean stress. Tensile residual stresses can be created in various ways, such as during machining (Ill. 16.2.1.1-2) with or without thermal influences and high thermal gradients of the type resulting during welding (Fig. "Residual stresses of welds").

Structural anomalies can be the result of influences on the surface. Examples include changes in the material with strength losses and/or embrittlement due to diffusion or oxidation. Strength losses can be expected, for example, in tempered steels above the annealing temperature and nickel alloys above the solution annealing temperature.
During the solidification of cast parts, the temperature gradient in the cross-section results in correspondingly oriented grain growth from the surface. If the grain boundaries of these so-called “columnar crystals” are parallel to the stresses, crack-causing damage mechanisms such as creep and thermal fatigue will be promoted by corrosive and oxidative intergranular corrosion (Volume 3, Ill. 12.6.2-23).

Forces in the crystal lattice: Because the surface is only two-dimensional, micro-deformations can damage the surface. One example is local destruction of protective oxide coatings through sliding during plastic deformation. This can, in turn, make it possible for damaging corrosion to occur (Volume 1, Ill. 5.4.2.1-2).

Figure "Importance of sueface integrity": In the case of fighter aircraft, requirements regarding performance and low weight are primary. In contrast, the priorities in the case of commercial aircraft are cost minimization in procurement and maintenance, as well as low fuel consumption. In both cases, the development requirements lead to greater operating loads on the parts. This especially affects the characteristics of the rotor components.

Higher stress levels: Higher performance levels and/or weight reduction will necessarily lead to higher RPM and increased operating temperatures in the rotors. At the same time, the disks are affected by increasing static (creep in hot parts) and dynamic loads/stresses (LCF).

Cost minimization based on fuel consumption requires improvements in overall efficiency. This leads to an increase in the gas temperatures and pressure, resulting in corresponding increases in the stresses on the parts. Procurement costs dictate a trend towards less parts. This means less compressor stages with integral designs (blisks, integral stators). Less stages means wider blades. These heavier blades increase disk loads. In inserted blades, this increases the root loads, especially in the contact surfaces (Volume 2, Ill. 6.1-15.2).
An increase in the overall pressure ratio with less stages in the compressor, combined with closer axial spacing of the blade rows, results in greater aerodynamic loads on the blades. Increased high-frequency vibrational loads can be expected. In order to safely ensure the required HCF strength, the requirements for part surfaces are extremely high.
Higher stress levels require the consideration of smaller, growth-capable flaws, as well as crack growth (Fig. "Testing of high strength material"). This means moving from a “safe life” concept to a “damage tolerance” philosophy (Volume 3, Ill. 13-14). Production-dependent flaws on the surface become ever more important.

Sensitive and “difficult” materials with high strength require that the production processes of the semi-finished parts/blanks have very tight tolerances. Melting and forging processes (Fig. "Forging process caused characteristics") do not leave much room for ensuring the required optimal structure is present throughout the entire highly-stressed part cross-section. The maximum tolerable flaw size becomes ever smaller. This also increasingly limits the process parameters of the following production steps, such as heat treatment and any soldering or thermal coating processes.
The inclusion of growth-capable flaws and crack growth places further demands on flawless workability. Unfortunately, the high strength of the material is a hindrance to this. It leads to greater cutting forces and increased loads on the surface during the production process (e.g. comma cracks, Ill. 16.2.1.1-2). Structural components such as carbides and hard phases, which are necessary to obtain the required strength, can break out or tear, resulting in growth-capable notches.
Powerful machining forces can penetrate more deeply into the surface. This affects the potential flaw size and depth of damage.

Design approaches the material limits: Computer-aided calculations make possible more exact and accurate determination of the operating loads. This is advantageously used by narrowing the gap to the minimum material data (design lines). However, this also demands that the characteristics of the serially-produced parts are sufficiently close to those of the design date. This requirement is related to increased reliability in the production processes and the material data determined by them. Of course, this applies especially to surface properties that determine dynamic fatigue strength (LCF, HCF).
If “surface integrity” is accepted as a principle, it can be used to defuse the problems of the historical development trend described at the beginning. Intensive engagement with the influences and damage mechanisms of the production processes should result in better understanding. This makes it possible to successfully and directly specify and influence the production and testing processes. An important prerequisite for production optimization is knowing the process parameters relevant to the operating
characteristics. These can be expanded without great difficulty within the framework of a value analysis. On the other hand, this makes it possible to see where no concessions can be made during actions intended to minimize production costs. Minimizing the scrap rate or amount of reworking (Chapter 17-5) simultaneouly reduces production costs.

Figure "Operation safety by surface integrity" (Refs. 17.2-1, 17.2-3, and 17.2-4): The high operating loads with relatively small safety margins in the parts of modern engines (see Fig. "Fatigue strength of machined titanium parts") require a minimal, controllable, and known influencing of the surface during the production process. This refers to “surface integrity”. The effects and characteristics that act on the part surface during the production process can be grouped using various terms. One possibility is shown in this diagram. The upper section is more related to the effects of the production process, while the lower represents the attained characteristics of the part. As far as possible, the characteristics include references to more extensive explanations or descriptions.

Figure "Effects of finishing processes on part surface" (Ref. 16.2.1-10): Three types of influence from a finishing process can affect the part surface (top diagram):

• Thermal: Heat development due to friction, arcs/spark formation (EDM), radiation (EB, laser), or hot gas (plasma, welding flame, heat treatment atmosphere). This can cause structural changes that lower strength levels. This type of influencing of the material behavior can be direct (annealing, solution annealing) or indirect. A typical example is sensitization of a material. This promotes intergranular corrosion (IGC) and weakening. High temperature gradients, as are typical for welding processes and EDM, induce damaging tensile residual stresses during cooling (Fig. "Residual stresses of welds"). If the heat development causes gas absorption or reactions with gases (oxidation, Fig. "Recognizing embrittlement of Ti-welds"), it can also reduce strength.
  

• Mechanical: If strain hardening occurs under plasticizing forces, such as during machining processes, hardening (shot peening), forming/shaping, or cutting, then residual stresses will be created in addition to hardening. Compressive stresses are desirable. They prevent corrosion cracking (Fig. "Benefits of compressive residual stresses") and lower the mean stress (Fig. "Increasing fatigue strength by compressive residual stresses"), thereby increasing the tolerable dynamic loads. Tensile stresses have an especially negative effect. During operation, a reduction in dynamic fatigue strength is especially problematic.
  

• Chemical: This includes processes that dissolve materials electrolytically (ECM) or aid mechanical material removal (EC grinding). ECM and etching do not cause plastic deformations or residual stresses. However, they can remove existing hardening and alter residual stresses through material removal, usually reducing or shifting them. Additional influences are reaction zones. For example, chlorine can be bound into a thin surface layer by titanium materials. Later heating can then cause stress corrosion cracking (Fig. "Chlorine in process baths causing stress corrosion"). Chemical reactions can cause selective material removal and damage grain boundaries or structural components (e.g. glycolic acid on carbides), and also cause corrosion notches (pitting corrosion).
  

The bottom diagram shows the typical influencing of a part surface by a machining process with a defined cutting edge (turning, milling, reaming). One can see, that several zones with different, characteristic changes can be defined. The development of the individual zones is clearly dependent on the machining parameters. They influence the strength behavior of the part, especially under dynamic loads.

Figure "Operating influenced by surface damages": This diagram shows typical examples from the many possible surface influences. They are not always damaging. It is also possible that a local improvement in operating characteristics will result. Hardening and induced compressive residual stressses in a plastic impression can locally increase the dynamic fatigue strength. This can also have a negative effect on surrounding areas. Tensile stresses in the surrounding area that are in balance with the compressive residual stresses in the impression can act as a weak point.

The affected surface zone in this diagram is only a few tenths of a milimeter thick.
Different finishing processes can induce comparable effects (Fig. "Process parameters influencing residual stresses"). Shot peening and machining can be used to induce compressive stresses and hardening. Etching and electrochemical processes can cause corrosion damage such as intergranular attack. The surface roughness is affected by virtually all material-removing processes (Fig. "Influence of surface topography and finishing").

Figure "HCF strength influenced by surface treatments": One of the most important material characteristics, which is also one of the most sensitive to finishing influences, is the dynamic fatigue strength in the fatigue resistance range (HCF, dynamic fatigue strength at high load cycle numbers). This is most crucial for parts that vibrate at high frequency and experiences a large number of load cycles in a very short time. These parts include compressor and turbine blades. In general, hardening due to plastic deformation increases HCF strength if the material is sufficiently ductile (processes shown above). Simultaneously occurring compressive residual stresses are made use of through shot peening (Fig. "Increasing fatigue strength by shot peening") and suitable machining (Fig. "Residual changes during finishing processes"). A similarly positive effect can be obtained by hardening the surface (e.g. nitriding or case hardening). In this case, as well, compressive stresses will be created in addition to a strength increase (Fig. "Residual stresses from hardening coatings").

In contrast, the breakdown of compressive residual stresses and hardened zones (e.g. electrochemical material removal), even to the point of creating tensile residual stresses (e.g. unsuitable grinding process) and softened areas (e.g. around local structural changes caused by heating), can dangerously lower the dynamic fatigue strength. If the surface is protected from corrosion by tough coatings, it can increase its dynamic fatigue strength considerably relative to the unprotected material.

Brittle coatings with tensile stresses (processes at bottom of diagram) can always expect to lower dynamic fatigue strength (Ref. 17.2-5, Fig. "Fatigue strength of machined titanium parts"). In this case, cracking during plastic strain is especially dangerous.

Figure "Machining parameters influencing HCF strength" (Refs. 16.2.1-2 and 16.2.1-6): The attainable dynamic fatigue strength in the surface zone is clearly dependent on the production processes and their parameters. The above diagram shows this using the example of the titanium alloy Ti-Al6-V4. In this case, ground surfaces, especially, show pronounced sensitivity (decrease in dynamic fatigue strength by about 70%) to “careless” grinding processes.

Milling and turning generally result in higher strength levels than grinding and other non-chipping machining processes. In general, “rougher” machining results in relatively lower dynamic fatigue strength losses. This could be related to plastic deformation in the surface zone and the compressive residual stresses it induces. The hardening influence should not be very pronounced in titanium alloys due to their lower tendency to harden.
Electric discharge machining (EDM) shows the known problems with dynamic fatigue strength (Fig. "Quality of boring processes").
The low dynamic fatigue strength of chemically machined surfaces relative to turned or milled surfaces can be attributed to the lack of residual stresses resulting from this non-deforming material removal process. This effect must be considered when changing a process from turning or milling to chemical material removal (e.g. electro-chemical milling = ECM). This type of change may be undertaken in the context of cost reduction in blisk production. In this case, subsequent shot peening may be necessary to harden the surface.

The bottom right diagram evidently confirms the lowering of the dynamic fatigue strength during ECM reworking of surfaces that were created by chip-removing processes. In this case, the surface is the optimal ground surface of a titanium alloy. However, it can be assumed that a ground surface created with less optimal parameters and with tensile residual stresses may actually show an increase in dynamic fatigue strength following ECM material removal. The prerequisite for this is that the chemical treatment does not create any local notches through selective attack, e.g. at the grain boundaries and/or through corrosion pitting. This effect could explain the low dynamic fatigue strength of ECM bores.
The process-typical damaging effect (in this case about 50%) of electric discharge machining (EDM) can be seen in the bottom left diagram, which applies to a high-strength forged nickel alloy of the type used in blades and disks.

Figure "Damages defused by reworking" (Ref. 17.2-7): This diagram can only provide an impression of the influence of various boring processes on the dynamic fatigue strength of cooled cast turbine blades. The tests were done on flat-bent fatigue test specimens with a bore (top diagram). The load conditions corresponded to those of a cooling air bore in a vibrating turbine blade.
The following is an analysis of the text results.

The effect of the boring processes on the dynamic fatigue strength differs in incubation time and crack growth (shown as an example in the middle of the diagram). There are no indications regarding the size of the dynamic loads.

The faster crack growth of the EDM bores is striking. This behavior can be seen in the steeper rise of the crack growth curve relative to an electrochemical boring process (STEM = Shaped Tube Electrochemical Machining). The cause is subject to speculation. Higher dynamic loads on the EDM specimens are probably not a factor, because otherwise crack growth would occur noticeably earlier. It is possible that the typically high tensile stresses in the melted/re-solidified zone (recast layer) were not sufficiently broken down by the shot peening. This effect may be supported by the brittleness of the recast layer. It is a result of the absorption of carbon from the process-specific dielectric (petroleum). During shot peening of the brittle coating, particles can break out and/or cracks form. An indication of this type of effect would be slightly earlier verifiable crack growth in the peened bores.

The generally very late (after 10⁸ load cycles) crack growth, as determined on the basis of the fracture surface, could be related to characteristics of the cast material.

Independent of the boring process and crack growth, verifiable crack growth (incubation period) began in the range of 10⁸ to 10⁹ load cycles (Volume 3, Ill. 12.6.3.2-1). It is not clear why the abrasive flow machined (AFM) bore showed slightly later crack growth in two cases, but not in another case with comparable treatment. It is possible that the abrasive flow machining created deeper damages that could not be removed. Therefore, these test also show how difficult it is to conduct sufficiently reproducible dynamic fatigue tests that are suitable for the selection of a finishing process.

Figure "Protection of dynamic fatigue by surface hardening" (Ref. 17.2-8): The strain hardening, (i.e. induction of compressive stresses) of parts subjected to powerful dynamic loads, such as rotor disks (LCF), also serves to minimize the influence of surface damages such as scratches, etc.. In the case of typically minimally hardening titanium alloys, the positive effects will primarily be due to compresive stresses. In nickel alloys and steels, strain hardening is of greater importance.
The diagram shows a relative dynamic fatigue strength (HCF) depending on the depth of scratches in a titanium alloy used in disks. Naturally, these values depend on the specific test conditions and should therefore be understood primarily as a trend.

A surface that is only machined using chip-removing processes showed a considerable decrease in dynamic fatigue strength when the scratch depth was greater than about 0.05 mm. This is the depth at which the “1-line” (uninfluenced surface) is breached. Evidently, this is related to the surface zone that is positively influenced by the machining process.

A shot-peened surface showed a damaging effect at a scratch depth greater than about 0.1 mm. It is interesting that this process provides the greates dynamic strength at shallower scratch depths.

Cold roll forming is a deep-acting forming process. This can be seen in the relatively deep “damaging” scratch depth of about 0.2 mm. However, in the case of shallower scratches, the dynamic fatigue strength attained by this process is less than than of shot peening.

Figure "Importance of surface hardening for life span" (Ref. 17.2-8): The top left diagram shows the pappern of the circumference component of machining-related residual stresses near the surface of a disk made from an Ni alloy. The top right diagram shows the effects of these process on the dynamic fatigue strength. It must be noted that these measured curves are very process- and material-specific and should therefore only be understood as trends.

In this case, chip-removal led to noticeable tangentially acting tensile stresses near the surface, which then transitioned into a compressive stress zone. The tensile stresses will make the disk surface sensitive even to minor damages in a radial direction. This is apparently confirmed by the relatively low dynamic fatigue strength and its pronounced scattering. However, the relative dynamic fatigue strength is still greater than 1, which means that there is still a positive effect compared with an uninfluenced surface.

An abrasive material removal process such as oxide blasting created a strikingly pronounced compressive stress zone. It evidently removed the thin tensile stress zone created by the chip removal. This process results in considerably improved dynamic fatigue strength behavior, especially in the HCF range. This is true not only for the height of the strength, but also its scattering. This behavior can be explained by the lower loads in the HCF range, where shallower surface flaws evidently do not tend to grow.

The shot peened surface showed a deep-acting protective compressive stress zone even after abrasive blasting. Therefore, shot peening has the most pronounced positive effect on dynamic fatigue strength. This is also true for high dynamic loads into the LCF range. It can also be used to “defuse” deeper damages within the pronounced compressive stress zone.
The bottom diagrams show the behavior of a disk made from a titanium alloy. The applied machining processes show pronounced compressive stress zones. This is especially true for cold roll forming (Fig. "Protection of dynamic fatigue by surface hardening") and is confirmed by the positive influencing of the dynamic fatigue strength.