Figure "Thermal fatigue remedies": For the problem of thermal fatigue, as well, an ounce of prevention is worth a pound of cure. In this case, prevention is accomplished through proper design and construction, while the “cure” is repair.

Evaluating operating suitability: A situation may arise in which the condition of the (high-pressure) turbine stator vanes must be evaluated while they are installed. This is generally done through a boroscopic inspection. Evaluation of the cracking is done with the aid of threshold values provided by the OEM (in a maintenance manual or bulletin, for example). Determining these threshold values demands specific knowledge of

• the operating loads
• the expected crack growth
• the failure mechanism
• the influence of parts on one another (e.g. possible connections between the cooling air system and the rotor blades).

and must guarantee:

• the operability over the intended life span. This means that the functioning of the part within the specifications is ensured. This also includes functions such as avoiding haircuts as consequential damages or, if applicable, the guarantee of an intermeshing function (braking the rotor in case of a shaft failure, see Volume 1, Ill. 4.5-9).
• no unacceptably early demounting or forced demounting at an unacceptable time.
• no spontaneous, unforeseen part failure.

Construction and design: Despite the comprehensive analytic aids currently available (finite element calculation of temperatures, strain, and stresses), experience is still indispensible for ensuring acceptable operating behavior (especially thermal fatigue behavior). If possible, proven design principles should be utilized. In addition to geometric design with the fewest possible notches (e.g. stiffness/cross-sections, cross-section transitions, radii), these are material selection (base material, coatings, connections) and ensuring acceptable operating temperatures (levels, gradients), which is accomplished through proper configuration of the thermal environment (cooling system, etc.).

Repair: The limits for repair and repair processes (e.g. high-temperature soldering, etc.) are usually provided by the OEM due to their most comprehensive experience (with other engine types, as well) and must be strictly followed. This is also true for apparently minor “improvements” such as different coating procedures. One must be sure to note whether there are restrictions on the number of repaired parts that may be installed. For example, do a certain number of stator vanes have to be new parts? This type of requirement may be necessary to fulfill certain functions that are not immediately discernable to the licensee and/or operator.

Figure "When can a crack be 'drilled out'": “Drilling out” cracks is a widespread method used in machine construction to stop cracks and prolong part life simply, quickly, and economically. However, this method is not always satisfactory.
Drilling into the crack tip results in a considerable reduction of the notch effect of the sharp crack tip due to the comparatively large bore radius.
If crack “A” (a thermal fatigue crack, for example) has already slowed down (“B”), the dynamic loads on the bore wall are so low that crack initiation cannot be expected to occur, at least not during a usable life span (“E”).
However, if the crack tip is in a highly-stressed part area in which considerable crack growth is expected (“C”), crack initiation must be expected in the bore soon after drilling. The crack will actually grow even faster, since it will have become longer with the bore. In this case, only a minor increase in life span is to be expected. In fact, there is even a risk of uncontrollably rapid crack growth (“D”).
In addition, one must determine if the bore will create a stress-increasing cross-sectional weakening or have a negative effect on temperature distribution by creating leakage flow. This could combine with increased thermal stress and contribute to crack initiation in the bore.

Figure "Preventing thermal fatigue cracks by design": Annulus cracks due to thermal fatigue (Fig. "Cracks in integral turbine wheels") are a typical problem in integral cast turbine disks, which are used especially in low-power turbines (e.g. in helicopters). In order to attain the resiliency necessary to reduce tangential stresses and thus prevent crack initiation, drained annulus designs are used (top diagram). In this case, openings are made between the blades. These are designed in a way that cracks do not develop towards the center of the disk. For this reason, a droplet shape (bottom right diagram) is preferred to a bore with a circular cross-section and an outward radial slit (keyhole bore). If the cross-section is not separated outward when making the opening, then it is assumed that the annulus cracks will develop towards the outside from the bore during operation. This behavior is promoted by the droplet shape. The outward crack development contains the risk of uncontrollable cracking, but the rough, interlocking crack edges provide a better seal against hot gas encroachment than the intentionally produced smooth slot. Hot gas encroachment could result in heating of the bore wall and increase the danger of thermal fatigue cracks.

Figure "Reducing high temperature gradient stress levels": The transition of welded-on eyelets and flange tubes to the housing walls or supporting walls of hot parts (Ref. 12.6.2-14) is a typical area for thermal fatigue loads (especially in combustion chambers and turbine housings; Fig. "Weak points at housings by thermal fatigue", Example "Low pressure turbine stage ruptured"). This is most likely related to deformation-retarding stiffness changes and differences in temperature between the housing wall and the part that is welded on (heat inflow and dissipation. Specific increases in the elastic resiliency (top left diagram) and a weld seam that is the optimal distance to the supporting wall (Ref. 12.6.2-13) are two types of improvements. However, there may well be drawbacks due to increased elaborateness or more complicated production.
If seal demands allow, disk-shaped structures that are heated from the outside (e.g. turbine disks, heat-treatment equipment and structures) can have their annulus zone “drained” with radial slits (Fig. "Preventing thermal fatigue cracks by design"). However, it must be ensured (safeguarding through specimen tests and sufficient engine testing) that no new cracks, possibly with higher growth rates, are initiated in the base of the slots or bores (Fig. "When can a crack be 'drilled out'").
Stiffness jumps exist around flanges (disk and shaft sockets, labyrinth cones, housings) and other supporting structures. The transition to the flange should be elastically constructed. In labyrinth cones (Volume 2, Chapter 7.2.2), this is accomplished by having a sufficient decay length to a fatigue-threatened zone such as a circumferential weld seam (bottom left diagram). The decay length is determined by the diameter of the cone and the thickness of its walls (Ref. 12.6.2-13).
In housings that are subjected to high temperatures and temperature gradients (e.g. turbine exit housings), it is possible to prevent typical brace cracks (Fig. "Preventing thermal fatigue housing cracks") by angling the braces in the direction of the circumference (bottom right diagram). The strain differences caused by the different diameters of the hub and outer housing ring, as well as the thermal strain of the braces, can be absorbed by rotating the hub relative to the outer ring.

Figure "Preventing thermal fatigue housing cracks" (Example "Low pressure turbine stage ruptured"): Due to the special arrangement of the injection nozzles in two concentric rings, the installation of a low NOx combustion chamber led to an unfavorable temperature distribution in the turbine exit housing struts, which were originally arranged radially (right diagram). The tangential arrangement of the struts allowed higher resiliency to the thermal stress and was introduced as a corrective measure (also see Fig. "Reducing high temperature gradient stress levels").

Figure "Remedy for thermal fatigue cracks by repair": Successful repair of parts with thermal fatigue damage, remedies for said damage, and life span extensions are all very complicated and demand part specific approaches. This is true even though repair soldering is successfully used on a large scale in turbine stator vanes.

Closing cracks: If larger individual cracks have already formed but have stopped growing (oxidation, especially in the crack base), joining methods such as welding or soldering can be used to close the cracks. Experience has shown that these seams will not provide the strength and toughness of the undamaged base material, let alone that of a new part. This holds even if the joining zone has been completely cleared of oxides. The structure of long-run parts is generally not in an optimal condition for welded joining. Therefore, micro-cracking, cavities, and structure-dependent low-strength (e.g. because optimal hardening cannot take place) are to be expected.
The strength of high-temperature solders reacts especially strongly to large soldering gaps (bonding flaws, formation of brittle phases). Remaining oxides are virtually impossible to remove completely, even with special treatments (aggressive reductive annealing atmosphere).
Repair welding or soldering of thermal fatigue cracks should always be seen as a sealing measure intended to prevent air leakage or hot gas encroachment. These repairs are often referred to as being “cosmetic”.

Therefore, it is extremely important to understand how many repaired parts may be combined with new parts, and in what configuration, in order to guarantee the required strength levels, i.e. load capabilities, of the total system (e.g. of a turbine stator assembly).

In crack fields or surface damages (orange peel effect, etc.), area-measured application may be the only option. Whole part sections such as bladings in stator assemblies made from sheet metal used in older engine types, or entire inlet edge sets, can be replaced through welding or soldering. It must always be ensured that the joining is done outside the thermally- and TF-stressed part areas and only in “healthy” material. Joining a blade edge in the area near the transition radii to the shrouds will result in rapid crack initiation and short part life spans.

Drilling-out cracks (Ill. 12.6.2-18): In this procedure, a bore hole is used to reduce the notch effect of the crack, and possibly to remove a damaged zone at the crack tip. Drilling-out is a relatively simple procedure and has the additional advantage of retaining the stress-reducing elasticity increase of the crack (breathing, Fig. "Reducing high temperature gradient stress levels"). Naturally, bores are only possible in areas where sealing requirements or overstressed residual cross-sections allow it. A further problem may arise if the fracture-mechanical effective crack length is increased by the bore and a new crack is initiated in the base of the bore despite the reduced notch effect. If this new crack is in a zone of increased tensile stress, it can unallowably accelerate crack growth.

Increasing the part elasticity in the cracked zone. Elastic construction of the housing walls with corrugation (Fig. "Reducing high temperature gradient stress levels"), or “draining” the turbine disk annulus with radial keyhole bores (Fig. "Preventing thermal fatigue cracks by design") can reduce stress levels and prevent TF crack initiation.

A corrective measure that can even extend part life relative to new parts is the avoidance of dangerous temperature gradients. Only in a few cases can this be accomplished with a protective cooling air film. A more effective option is the application of a thermal barrier coating on the hot gas side. This coating can reduce the TF loads on the underlying repair welds and/or solder far enough to ensure a sufficient life span to renewed crack initiation. Of course, it must be ensured that the coating does not impede the functioning of the part (e.g. through coating thickness or gliding properties). However, this requires time- and cost-intensive part tests and engine trials in order to safeguard operation.

Figure "Representative specimens for thermal fatigue": It should be self-evident that specimens for determining material data for engine part design correspond to the actual parts in their specific material properties in their location in the engine (also see Fig. "Representative samples for part quality"). However, this requirement is difficult to meet, especially in the case of TF and TMF.
It is very difficult to obtain a specimen that can be used for determining construction values (i.e. one which can be put under defined stress) from the highly TMF-stressed transition to the shroud of a turbine stator vane (top left diagram). For this reason, these specimens are often cast separately. If the thickness of the specimen is greater than that of the actual part at the critical point, different grain sizes, shapes, and orientations will be created during the casting process. The thicker the cross-section, the larger and rounder the grains (bottom right detail). Subsequent drilling will not change this problem (Ref. 12.6.2-15). Even with hollow cast specimens, it will hardly be possible to satisfactorily simulate the grain structure in the critical part edges. Due to the temperature gradients during solidification, stem crystals that run parallel to the shroud will develop in the blade edges (top right detail). This means that the long grain boundaries experience cross-stress from centrifugal forces. This results in earlier crack initiation and faster crack growth relative to optimal structures. The wedge-shaped TF specimen (middle right diagram) seems to be more suitable for testing part-relevant structures. In this case, the load cycles slow themselves due to the restricted thermal strain (Fig. "Cracks protecting from thermal fatigue"). For this reason, large deviations between the thermal fatigue behavior of the specimen and part are to be expected.
The situation is even more problematic with coated parts. In this case, the coating behavior (ductility, thermal strength) and thickness play an additional role. Therefore, construction data should be determined with the use of samples with part-relevant coatings.
Phase-shifting and dwell times at the maximum and minimum have a much stronger effect on the crack initiation life span than a linear damage accumulation would suggest (Fig. "Lifespan verification by cyclic spin test "). For example, dwell time in the relaxation phase may increase life spans considerably by enabling part recovery, while in other cases it may reduce part life. This is also true of dwell times at high temperatures and stress levels.
Different phase-shifting can lead to very different crack initiation life spans, depending on the material. A sufficiently safely controllable cyclical strain for FEM-supported design does not seem to be present in this type of test. Rather, this seems to be a comparative material test.