Table of Contents
15. Material Flaws
Problems with Blanks and Semi-finished Parts
When discussing problems with semi-finished parts and blanks, engineers will first think of material flaws. However, less-experienced personnel are especially prone to overlooking the fact that even the procurement of semi-finished parts or blanks can be related to major problems (Fig. "Producing semi-finished parts"). For turbine engine parts, extremely comprehensive requirements exist for quality, documentation, reproducibility, and monitoring of the detailed, specified technical steps in order for the latter to be accepted. This reduces the pool of potential suppliers considerably.
An alloy that conforms to specifications does not necessarily mean that a semi-finished part qualifies for use in an engine part. There are a large number of additional criteria that must be met before serial implementation can be considered (Fig. "Material operating properties").
The problems with procuring raw parts can necessitate temporary solutions in order to bridge overly long delivery times in the development and prototype phases. For example, one may be forced to use a raw part with a less-than-optimal grain orientation, despite the high risk of doing so (Fig. "Strength and grain orientation").
Specification-conforming blanks are not only a requirement for safe operating behavior. They are also indispensable for safe, specification-conforming, problem-free, and reproducible finishing with acceptable rejection rates. For example, traces of undesirable alloy components make welding more difficult, resulting in considerable production problems, cost overruns, and delays. In the same way, machining properties are dependent on the structure of the raw parts (Fig. "Specific part structure").
There is extensive technical literature that discusses materials science and the problems and flaws that can occur (see “recommended technical literature” at the end of this chapter). For this reason, this chapter limits itself to engine-specific problems that are not satisfactorily described in the general technical literature.
In the following, material flaws are defined as those flaws that are outside of the specification limits and were causally created during the production of the blanks or semi-finished parts. Specifications include drawing specifications, material standards, and test findings (also see Volume 1, Ill. 3-1).
The term “weak points“ (in welding technology, the term “imperfections” is used) is applied if their effects are taken into account in the design of the parts. This means that weak points are covered by the specifications.
Flaws in raw parts are not limited merely to material separation and inhomogeneities with dangerous notch effects. Collections of alloy components or phases (e.g. carbides, nitrides), i.e. segregations, can be equally unallowable. The grain size and distribution, as well as the grain orientation, can determine the usability of a raw part.
According to the definition, flaws are expected to result in unallowable impairment of the operating properties (life span, safety). Therefore, they are unallowable and must be prevented. Flaws and irregularities can affect very different properties, including:
- Strength: dynamic, static
- Material behavior: brittleness, corrosion behavior
- Fracture-mechnical specific values
- Corrosion properties, also crack growth
- Behavior during production processes (work, costs/rejects)
Due to the availability of sufficient technical literature, this text will not include a comprehensive list and description of flaws from the primary forming process (Fig. "Finishing semi-finished parts"), which includes casting and forging. Experience has shown that a few typical terms are frequently used erroneously or misleadingly outside of technical circles. For this reason, these terms will be addressed. They include the typical flaws in cast parts, i.e pores and cavities (Fig. "Types of hollow material flaws"), which have differences that are decisive for implementing solutions. In forged parts, “grains” cause confusion in technical terminology. Another complex concept is segregation, including its causes and effects (Fig. "Mechanisms of hot cracking").
The term heat treatment, which is sometimes mislabeled as “hot treatment,” is used in the following primarily to describe annealing processes. However, according to the definition provided by DIN, it also includes processes such as hardening, carburization, and nitrogen hardening. Forging and HIP processes, for example, are not included in the category of heat treatment, even though their temperature sequences have a comparable influence on the material.
In the case of production procedures, typical flaws with technical terms occur. In welding technology, especially with high-alloy steels, nickel alloys, and titanium alloys, hot cracks or thermal cracks (Fig. "Mechanisms of hot cracking") are especially problematic and sometimes unavoidable.
During raw part production and the subsequent manufacturing process until the part is finished, weak points and flaws can change, becoming more or less dangerous. This can be specifically influenced during the production process with a sufficient degree of understanding.
Figure "Producing semi-finished parts": If an alternative material must be used, it can lead to unexpected problems. Less experienced personnel may assume that a material that is approved for use in aviation is automatically available in the form of raw parts with the required properties. This assumption can prove to be very dangerous, if the time interval between ordering the raw parts and their delivery has been underestimated. It is also possible that the desired raw parts are not available in the market due to insufficient demand. For example, a turbine disk requires a forged blank with a suitable grain structure and grain orientation. Due to supply problems, it seems practical to use a disk cut off from bar stock (Fig. "Strength and grain orientation"). However, the axial grain orientation of the bar stock does not have the required strength properties. This is especially true for LCF behavior under tangential stress. Therefore, it is not a matter of course that a material with special, part-specific required properties is readily available in the marketplace that is approved for use in aircraft engines.
If the material must first be approved for the (special) aviation application, a major investment of time and resources must be expected. This is true even if the technical steps of the raw part production (e.g. melting, forging processes) have been sufficiently tested, fixed, and specified. The expense derives primarily from the determination of safe design data (statistics) and part-specific ratings (e.g. cyclical centrifuging tests). The certification of a material for rotor parts in aircraft engines should take roughly a year and cost about the equivalent of 20 single-family homes.
The risk of raw parts being unavailable is also present if the time interval between completion of development and serial implementation of a program is too great. In military projects, especially, this time period can be several years. In this case, expensive unused investments by the manufacturer of the semi-finished parts cannot be recouped. This may mean that, for example, instead of a hitherto highly-loaded powder-metallurgical part, the only available part is made through conventional forging. In this case, even with identical alloy composition, a cost- and time-intensive certification, perhaps even with extensive testing, may become unavoidable.
Even if the desired raw parts are available, it is necessary that the supplier is approved for the production of parts to be used in aircraft engines. This certification includes not only the individual production steps, but the entire quality management process. It demands acceptance by the proper authorities, as well as the approval of those responsible for design (e.g. OEM in licensed production), the customer (being supplied), and the operator (e.g. in military projects).
Even if the raw part supplier is certified for aircraft engine part production, this does not mean that the certification applies to the production of a specific part. Special comprehensive certification may be required. It is possible that the certification process must be repeated for every single disk of a rotor. The expense does not only affect the raw part supplier, but also applies to the manufacturer of the finished part. At minimum, one must expect reject pattern tests and limited strength verifications. This may require extensive cyclical centrifuging tests.
Even if all of the above requirements for raw part supply have been met, long time periods between ordering and delivery (delivery times) may be unavoidable. This may take several months or even several years.
So far, only technically-based problems have been mentioned. However, these are overlayed by costs. As mentioned above, these do not only apply to the raw part supplier. Any further certifications and optimizing steps must also be included. Additional aspects of procurement include delivery reliability, quality, and demands for double sourcing. The last of these is the demand for two independent supply sources, which reduces costs and increases the reliability of the supply.
Experience has shown that excessive fixation on minimal raw part prices can lead to problems with additional costs that greatly exceed the savings. For example, in bad economic times, a supplier may reluctantly agree to prices that do not cover his costs. However, as soon as the market recovers in his favor, he might halt delivery for ostensible reasons such as high reject rates. In addition, if the prices are not sufficiently high, the motivation for quality improvement and engagement in projects with development risks is understandably low.
Figure "Material operating properties": A metallic material is determined by its composition (alloy). This refers not only to the main alloy components, but especially the basis metal. Important factors include both minor alloy components within the specifications, as well as limited traces and fouling. Next to the specification-conforming composition, other characteristics are of decisive importance for the operating behavior. The chart shows the diversity of influences of material characteristics and operating behavior. Operation itself can also change materials. This makes the concerned problems very complex, and this depiction does not make claims to completeness or universal validity. It should, however, sensitize the reader to this situation.
Structural characteristics: This term primarily refers to structures created by forging and casting. The structures of powder-metallurgical (PM) materials can be treated separately. In general, the typical coarse-grained and, due to the dendrites, inhomogeneous cast structures have high creep resistance (Volume 3, Ill. 12.5-11). For this reason, they are used in parts exposed to high temperatures, such as turbine blades. The disadvantage of these materials is that their LCF and HCF strength is low relative to that of forged materials. The relatively fine-grained forged materials, on the other hand, are especially suitable for rotor disks and compressor blades due to their comparatively high dynamic strength. Powder-metallurgical materials typically have fine, non-directional grains, and reach high strength levels with good ductility. Usually, they are available in a re-forged state (hot isostatic pressing (HIP), Fig. "Safety of parts by manufacturing"), and are therefore very expensive. Their resistance to crack growth is of increased importance in the case of growth-capable flaws in parts with limited LCF life (Damage Tolerance Concept). This resistance, which increases safety, is relatively high in the inhomogeneous cast alloys. Forged alloys that contain pre-cast materials (Ills. 15.2-17 and 15.3-7) have a greater resistance than the extremely homogeneous, fine-grained PM materials (Volume 3, Ill. 14-10). The higher crack growth rate of the latter makes it more difficult, if not impossible, to catch cracks through periodic inspections during operation.
Grain size and distribution: In general, the coarser the grain of a material, the greater its creep resistance/creep strength. This behavior can be explained as follows: the finer the grain, the greater the total sum of grain surfaces, i.e. grain boundary lengths. This means that the concentration of impurities on the grain boundaries is lower in fine grains. This means that fine-grained materials tend to be less sensitive to damaging influences on the grain boundaries. These influences include chemical attacks on the grain boundaries, as well as hot cracking (Fig. "Mechanisms of hot cracking").
Grain structure: These are structural characteristics within the grains. Typical examples are the structures of steels and titanium alloys. The effect of structure, such as the hardened or heat-treated structure of steels, is generally known. This is also true of the different behaviors of austenitic and ferritic steels (e.g. magnetism and corrosion).
a+b-titanium alloys can have very different structures, depending on forging and heat-treatment. These structures can have a risk-increasing effect on the crack initiation phase and crack growth (Ills. 15.2-19 and 15.2-20).
Grain orientation: The position of the grain boundaries is extremely important for strength and failure behavior (Fig. "Strength and grain orientation"). Loads across the grain orientation will generally result in lower ductility and strength than loads acting in the direction of the grain. The crack growth under cyclical loads and/or corrosion can be accelerated, which is, of course, a negative development. Creep behavior, with its typical pore formation and cracking, is especially negatively influenced by grain boundaries that are perpendicular to the direction of stress.
Phases: Desirable (e.g. carbides, hardening phases) and undesirable (such as the brittle sigma phase) must be distinguished. They are also influenced by operating conditions and have an especially negative effect on strength and ductility.
Fouling, trace elements: Tiny amounts of material-specific fouling can change operating behavior considerably. One example is bismuth (Bi) in nickel alloys (Fig. "Heattreatment batch as finishing problem indication"). This primarily effects the behavior of the grain boundaries, on which the damaging elements collect.
Figure "Strength and grain orientation": The grain orientation, i.e. the dominant orientation of the grain boundaries, influences part behavior in many ways (Fig. "Material operating properties"). If the loads act perpendicular to the grain orientation, failure along the grain boundaries will result in worsening of the following properties:
- static strength, especially at high temperatures (creep)
- dynamic strength (LCF/thermal fatigue, HCF)
- crack growth (rate, critical crack length)
- crack toughness
- fracture strain
Notch impact toughness is evidently an exception. An impact in a direction perpendicular to the grain will cause a fracture perpendicular to the grain boundary orientation. However, the flexural stress will result in tensile stress along the grain boundaries, with high energy consumption.
The influence of the grain direction is of considerable importance for the selection of specimens from parts and of parts from rolled raw parts such as bar stock or plate stock (top diagrams). Exceptions are rings rolled in one piece or butt-welded from rolled sections. Their grain direction is oriented to the highest loads (tangential stress). In rotating parts such as disks, especially, it is virtually impossible to avoid high cross-stressing if the parts were made from plate- or bar stock. No such case has been reported in serial operation. However, in the case of testing or development parts, which must be realized within a tight time frame, the temptation to use this procedure would be great.
A potential problem that also applies to serial applications is the poor orientation of metal sheets in pressurized housings and flanges in compressors, combustion chambers, and turbines (bottom diagrams, poor configurations; Volume 3, Ill. 188.8.131.52-9). Under circumferential stress caused by the internal pressure, there is a danger of axial cracking, rapid crack growth, and a small critical crack length due to the lower crack toughness. This means that the failure risk increases in two ways: early cracking and a short time span for discovering the cracks. Even in housing walls with containment tasks, poor grain orientation can result in unallowable operating behavior. This type of poor grain orientation can be the result of available raw part dimensions or a mistake.
Figure "Specific part structure": The different zones of an engine part can be subjected to very different mechanical and thermal loads for varying durations. These demands cannot be met by a single material structure alone, as the example of a turbine disk shows.
The annulus of a turbine disk (top diagram) is located near the hot gas flow. This subjects it to especially high operating temperatures. In this case, the life-determining loads are primarily creep and thermal fatigue. Therefore, in this part zone, coarser grain (Fig. "Material operating properties") is desirable than in the highly LCF-stressed hub and disk, which have relatively low operating temperatures. The right diagram shows the changes in LCF strength at different temperatures in the different structures of a typical disk material shown below. These structures are created by the forging and heat-treatment processes, which are specifically adjusted with regard to temperature distribution, duration, and shaping. One can see that below a part temperature of about 600 °C (grey area), fine-grained and hardened structures have a considerably greater LCF strength than coarse-grained structures.
The behavior is different in the case of LCF strength (thermal fatigue), hot yield strength, and creep resistance above 600 °C. The right diagram clearly shows that the flatter curves of the coarse-grained structure are located above the sharply dropping curves of the fine grain. For this reason, a coarse-grained structure in the annulus is optimally matched to the increased creep stress and thermal fatigue stress in this area.
In the integral turbine disks of smaller gas turbines (e.g. for helicopters), this structural tuning (dual property) has become especially advanced. With the aid of diffusion joining (diffusion welding, HIP), an annulus made from a typical coarse-grained cast material is joined with a forged or powder-metallurgical hub (Ref. 15.1-11, Fig. "Diffusion weld of a ‘dual property’ part"). The problem with these parts is always the joining zone, which must be viable for non-destructive testing methods (Ref. 15.1-1).
Figure "Finishing semi-finished parts": The quality of finished parts is influenced by production steps that are, in turn, dependent on the properties of the semi-finished parts:
- grain size and orientation
- hard phases and their distribution
- hardening state
- segregations and fouling
- residual stresses
Machining: The strength or hardness are decisive for the machining properties. Materials with extremely high strength, such as PM materials for turbine disks, can cause unacceptable tool wear. For example, it is possible that broaching the fir-tree slots of a single small turbine disk uses up broaches equivalent in value to a medium-size car. The required surface quality in the typically extremely highly-stressed part zones does not permit weak points such as chatter marks, galling, or cracking. A special occurrence is comma cracks, which are an inter-crystalline cracking-open of damaged grains at the surface (Ill. 184.108.40.206-2). They are evidently related to the grain orientation.
In very coarse-grained, difficult to machine materials (grain sizes greater than 1 mm), such as cast Ni alloys, the different orientation-dependent properties of the grains become apparent. The different machining forces become apparent in the heights of the individual grains of the machining surface. This results in a surface with a pronounced structure (orange peel effect). This effect is also an indication of high hardenings and correspondingly large, locally very different, machining-induced residual stresses.
For the strength of high-temperature materials, in addition to the hardening phase (usually g'-phase), the size and distribution of the carbides is important for a machining procedure.
Carbides located near the surface can break out and/or be destroyed during machining. The sharp notches that this creates, even if they are small, can have a considerable influence on the dynamic strength of the part.
If, during the machining process (e.g. grinding, high-speed milling), local temperatures are high enough to cause the grain boundaries to soften, this can combine with the simultaneously occurring thermal stress to cause hot cracks (Fig. "Mechanisms of hot cracking"). The sensitivity of a material for this cracking of the grain boundaries depends on the grain orientation, the grain size, and the high-temperature strength of the grain boundaries. Trace elements (such as bismuth) and the hardening state (grain strength) also play an important role.
Forming, trueing: Understandably, lower ductility (plastic deformability) promotes the development of cracks and thus the fracturing of parts. The plastic deformability is influenced by proper orientation of the grain boundaries (grain direction). The greatest ductility is in the direction of the grain boundaries, and forming should make use of this factor.
Coarse-grained materials are generally less easily deformed than fine-grained ones. With pronounced coarse grains, the orange peel effect described above can occur. Here, as well, the orientation-dependent deformation properties of the individual grains are a factor.
For production steps with this type of effect, an optimal structural state should also be emphasized (solution-annealed, for example). This can vary considerably from that of the finished part, which is only attained after further heat treatments.
Welding: This production process is especially material-dependent. The main flaw type is hot cracking or stress relief cracking (Fig. "Hot cracks by heat treatment" and Table 220.127.116.11-1). The hardening state is
important. Therefore, when welding heat-treatable materials, a ductile, less creep-resistant, solution-annealed structure is preferred.
Coarse grains are considerably more sensitive to hot cracks than fine grains (Fig. "Influences at micro-cracking").
Understandably, segregations at the grain boundaries result in a decrease in creep strength and lowering of the melting point, which promotes cracking. Therefore, a material variant with especially strict alloy specifications may become necessary.
Etching, electro-chemical and chemical material removal and boring processes: Aggressive media are required for the removal of oxide coatings or the opening of cracks that were smeared shut in previous production steps. If these media act too long or attack sensitive grain boundaries too strongly, it can result in dangerous damage (Fig. "Etching problems"). The sensitivity of grain boundaries is a result of accumulation and/or depletion of alloy components and phases. In this regard, alloy type, heat influence (sensitization, Fig. "Schweißteile") and grain size (Fig. "Material operating properties") are factors. The surface quality (roughness) of chemical and electrochemical treatment surfaces, such as those subjected to material removal or boring processes (Ills. 18.104.22.168-3 and 22.214.171.124-6), also reacts to grain size and inhomogeneities.
Heat treatment: Most importantly, deformation and cracking are to be avoided. High residual stresses from a forging or casting process can lead to deformation during machining and/or cracking during heat-treatment. The level of the residual stresses depends partly on the high-temperature strength and creep resistance of the material. This limits the possible stress relief. This cracking is subject to the same influences as welding.
Quality assurance: This section treats non-destructive testing as a production step. In the case of ultrasonic methods the testability, i.e. the safely detectable minimum flaw size, is dependent on the interference of the structure on the echo (“grass”; Fig. "Limiting influences non-destructive testing"). The smaller and more even the grain size, the smaller the flaws that can be safely detected. The result of this is that despite having the same alloys, a forged part (coarser, more uneven grain) will not tolerate as high a level of cyclic loads as a fine-grained PM part.
In the case of the large and structurally inhomogeneous grains of cast parts (up to the cm range), even X-ray inspection is made more difficult through diffraction caused by the varying orientations of the atomic lattices of the individual grains. Residual stresses inside the parts cannot be measured with serially applicable non-destructive methods (Fig. "Measuring residual stresses and its interpretation" and Table 17.3.2-1)
Figure "Material flaws": Flaws in the semi-finished or raw parts, i.e. material flaws, can be created in very different primary shaping processes, and can influence one another during the course of the production process. In the depicted schematic sequence, the individual production steps are referenced to diagrams with more extensive explanations in the following chapters listed in the right column. Damage-relevant problems of important production steps are:
Melting: Even the basic materials from which an alloy is created influence the quality of the semi-finished parts and the finished parts. For this reason, semi-finished parts that were, for example, melted using chips left over from machining, are suspect because particles from the machining tools (hard metal) may have been included. Properties of titanium alloys are influenced by the quality of the titanium sponge as the raw material. Traces of dangerous foreign materials can be introduced through the primary material. These can have a negative influence on the following production steps, as well as on the operating behavior of the finished part.
Casting: Experience has shown that the ingot can contain flaws from the casting process that were not removed through material separation (Ref. 15.1-3, Ills. 15.1-13 and 15.3-7). Flaws resulting from these casting problems can remain in the finished forged parts even after intensive shaping processes (Ills. 15.2-22 and 15.2-23).
These include accumulations (segregations) of impurities (oxides, nitrides, carbides) that can be traced back to the melting process (Fig. "Segregation induced turbine disk failure"). Uneven distribution of alloy components can also usually be traced back to the solidification process. Oxide-coated separations, such as cavities and cracks, can not be completely closed by the subsequent forging process. If the flaw location in the raw part is problematic for non-destructive testing, due to its ultrasonic contour, for example, this type of flaw will also be found in the finished part (Fig. "Correct application of non-destructive testing"). In cast parts, such as turbine blades and integral turbine disks, undetected flaws and/or tolerated weak points such as limited shrink porosity (Fig. "Types of hollow material flaws") will remain in the parts. These are followed by residual stresses from the cooling process. If a machining process disturbs the equilibrium of the residual stresses, it will be revealed by warping. Later operating loads can overlay with residual stresses and cause damage (also see Fig. "Residual process stresses in titanium rotor disk"). The thicker the cross-sections in cast parts (e.g. integral turbine disks), the greater the expected tensile stresses.
Forming: Forging, rolling, and extrusion turn the cast material into forged material. This process step is unable to eliminate or minimize some flaws from the casting process (Ref. 15.1-2, Fig. "Formation of forging flaws"). The forming process itself can also be the cause of flaws. A susceptibility to this occurrence may, in turn, be related to characteristics of the casting process. High-temperature alloys make the forging process especially difficult. Hot cracks and a low level of deformability (Fig. "Forging process caused characteristics") with undesirable grain size and/or grain orientation are typical results (Ills. 15.1-3 to 15.1-5). Hot cracks can occur at the surface and/or inside of forged parts. It is not expected that oxide-coated hot cracks that are open to the outside can be closed by the forming process. Even internal cavities can remain as weak points.
Heat treatment: Even in the procedural steps of casting and forming, “integrated heat treatments” are taking place. If these are not sufficient for the required raw part properties, special heat treatments must be added. Large cross-section differences have different influences on the time/temperature sequence. This can have unallowable effects on grain size, precipitations, and residual stresses. Depending on the high-temperature strength, dangerously high residual stresses can be expected after heat-treatment (Volume 3, Ill. 14-9; Fig. "Residual stresses in spite of stress-relief annealing"). High thermal stresses can cause hot cracks that will remain in the finished parts if non-destructive testing fails due to the unusual flaw pattern (Fig. "Failed detection of large cracks").
Powder metallurgy (Fig. "Influences of structure features"): Due to the nature of this process (sifting), the maximum flaw size can be kept small. The grain size is very small and even. However, this may not be desirable for zones under high creep stress. This problem is addressed with subsequent procedures, such as special heat treatments, which act through grain growth. Especially problematic flaws are foreign particles from the filling process or the powder (Fig. "Safety of parts by manufacturing"). The utilization of the high potential strength only permits very small weak points. Flaws that escape non-destructive testing (Ills. 17.3.1-2 to 17.3.1-4) can easily become growth-capable flaws during operation. Even re-forged (HIP) semi-finished parts can have dangerous residual flaws of this type (Fig. "Safety of parts by manufacturing").