Since there are many different operating influences and parameters that determine fretting (chapter 6.1), the number of different remedies is correspondingly high. These remedies are specifically designed for the problem they are intended to solve, but there is no universal remedy. Fretting remedies can be divided into several main groups:
The following sentence must be taken to heart: where there are no micro-movements, there is no fretting. Since fretting fundamentally reduces dynamic strength, fretting must be avoided in highly stressed zones. An example is disk damage that can be traced back to a crack initating in the region where a metal sheet layed against the hub (Ref. 6.3-2). The designer should always question whether a detachable connnection is necessary, or if a permanent connection (e.g. welding, soldering, one-piece assebly) is feasible. This requires comprehensive reconsideration, since compromises must usually be made. One need only think of the difference between a welded and a threaded compressor rotor or the difference between a blisk and a conventional disk/blade system. Anti-twist devices with bolts on rotor flange that are braced across the shaft, rather than screwed, have been shown to be susceptible to fretting (Fig. "Fretting damage at turbine rotor"). Therefore, threaded connections are preferable to the former type of bolt connection. If a detachable connection cannot be avoided, the design of the attached components in the contact zone should be optimized. Because micro-movements occur even in tight positive fits (when the parts in the connection area elastically deform in different ways under operating loads), emphasis must be on creating tuned, equal elastic behavior. Of course, this is not necessay, if the operating loads are so low that no noteable elastic deformation occurs (Ref. 6.2-2). This task is made easier by computer-based deformation calculations, which should by this time be standard in design layouts. In connections whose function entails glide movments (e.g. plug tooth systems in shaft couplings), sufficient lubricant supply must be ensured.
Reducing the oscillation amplitude can be accomplished by the aforementioned optimization of stiffness in the contact parts, and also by weakening the vibration stress. Before a targeted reduction of vibration stress is possilbe, a theoretical and/or practical analysis (e.g. modal analysis) of the vibration types and their causes must be completed. Only then can targeted changes to the elasticity, mass, and damping be undertaken with any hope of success.
Cylinder supports for pipes must be equipped with appropriate intermediate layers. It is important that warping of the pipe during manufacture and operation is safely avoided.
The design of contact surfaces and their transitions can make a decisive contribution to operating safety. Important factors include surface pressure, position of the surfaces with regard to the transmitted forces (e.g. dove tail angles), as well as strain-relieving fillets in the transition to the contact surface (Fig. "Reducing wear and stress in contact zone", top diagram).
Putting a pattern into the contact surfaces (Fig. "Blade foot wear symptoms") can also improve their operating behavior, but it must be designed for and be proven effective in the specific application.
Figure "Reducing wear and stress in contact zone" (Ref. 6.1-10): If the load levels of the blade root, especially the shaft, allow it, a form change can result in a very effective load-capacity increase in the lay-on surfaces. An undercut in the shape of a flat axial cyma (top diagram) increases flexibility at the edge of the lay-on surface. This causes the expansions of different part areas to be more equal, minimizing micro-relative movements. This also relieves stress in the blade`s lay-on surface that is caused by flexural modes (compare Fig. "Fretting load parameters in dovetail"). The load levels in the lay-on surfaces can be reduced considerably in this manner (top diagram).
The bottom diagram indicates the possible influence contact surface topography can have on the load capability of a part (Fig. "Blade foot wear symptoms"). This is also assumed to be true in shot-peened surfaces, where a targeted increase in the load capability of contact surfaces has been achieved. Proper application of this effect still requires considerable fundamental research work.
Overhaul and maintenance
Determining and adhering to proper overhaul intervals is important. A determining criterion is the change in the contact zone due to operating conditions. If changes are expected to increase stress levels to the point that they are not covered by the design, then the part must be reworked or regenerated. Dangerous changes include local metal removal accompanied by an increase in or local relocation of the stress in the contact surfaces. Glide coating failure that increases the coefficient of friction or a drop in the pressure residual stress due to creep (e.g. in shot-peened surfaces) requires repair.
The technology required for regeneration must be developed, proven, and made avaliable in time.
Overhaul guidelines are necessary in order for overhauls to be suitable, workable, and effective. These should be given in engine-specific overhaul manuals. Problems can arise from foreign-language technical terms or translations of these terms in the guidelines. If implementation results in implausible results (e.g. unusually frequent damage), one should consult a specialist and/or if in doubt, consult the original foreign-language document. It is reccomended that clear damage descriptions accompanied with pictures and diagrams are included, as well as clear specifications of the allowable tolerances for re-installation, repair, and replacement; i.e. parts that are no longer usable. Meaningful pattern tables for the damage types are useful for classifying damages. However, it must be remembered that assessing damage (i.e. loss in dynamic strength) on basis of the external damage symptoms is always problematic and requires sufficient experience. Additionally, repair measures (specifications and working plans) must be provided. Repairs are usually orientated according to the production process of new parts.
Decisions regarding type and extent of damages depend on a sufficiently reliable damage report. This requires suitable non-destructive trials and/or destructive trials of random samples. It is vital to be aware of the limitations there are to assertions made based on evidence from these trials. For example, it is very difficult to find oxide-filled micro-cracks in lay-on surfaces.
When combining broken-in parts with new ones or combining parts with different wear rates in their contact surfaces (Fig. "Fretting wear induced fir tree cracks"), it is important that no unallowable surface pressures or tensions are created
by the changes. A suitable reworking may be necessary in some cases.
If it becomes known that unusual operating conditions (e.g. windmilling) lead to increased fretting stress in certain engines, then these engines must undergo special inspections.
It is important to ensure that fretting damage is not promoted by engine maintenance. Constantly lubricated shaft connections (e.g. multiple splining), for example, must have a sufficient grease reservoir. Problems can occur due to oxidation products bleeding out or increased play, for example. Re-lubrication must be done only with approved greases. Cleaning fluids should not be used in such a way that necessary glide materials or coatings are dissolved, changed, or washed away. This includes cleaning materials used for compressor washing that can enter the gap at the blade roots if the engine is stopped or running at low RPM.
Avoiding Unfavorable Operating Conditions
Some operating conditions promote fretting wear. In many cases, these conditions cannot be avoided, but in individual cases they sometimes can be avoided. For example, windmilling (free rotation of the rotor of a shut-down engine caused by the airstream) is known to cause heavy fretting stress in the fan blade roots. This is caused by the weak centrifugal force due to low RPM. This can result in the blade roots undergoing large relative movements with contact surfaces seperating and hammering wear occuring. This is an unusual damage mechanism for these parts.
Testing Fretting Systems
A sufficiently certain statement as to the suitability of a fretting system during operation can only be made based on operating experience or, as the case may be, testing under conditions that simulate those during operation. As a basic principle, testing in the engine should be preferred. This requires that the trial runs sufficiently correspond to conditions in serial operation. For example, a corrosive or eroding influence that occurs during normal operation with marked downtimes (e.g. military use) is almost impossible to sufficiently realistically simulate in a testing rig. Shortened trials with increased loads or single-stage trial instead of load collectives similar to those present during operation should be viewed very sceptically.
Selection of suitable material combinations
There are various methods of ensuring that material combinations in contact zones act to minimize fretting damage:
Base material combinations in compressor disks and compressor rotor blades are determined by operating requirements such as heat resistance, dynamic strength, and thermal strain. This makes after-treatment of the contact zones possible.
Metallic surface coatings can be applied in various ways: with the aid of a galvanic process, streamless treatment in a coating bath, thermal spraying, vapor deposition techniques (PVD), or chemical gas reactions (CVD) and diffusion treatments. Many factors must be considered when selecting coatings. These include optimal coating thickness, sufficient heat proofing, plastic deformability (evens out imprecise application; e.g. Ref. 6.3-1, CuNi or CuNiIn coatings on titanium alloys), a low coefficient of friction with the opposite surface, wear resistance, low sensitivity to damage caused by micro-movements, sufficient oxidation resistance, and no formation of damaging oxides (which increase the coefficient of friction, initiate cracks, and ccelerate wear).
Thermally sprayed WC-Co coatings, the tungsten carbide particles of which are embedded in a cobalt matrix, have proven themselves as wear surfaces under high surface pressure and hammering stress. These coatings are typically used in areas 1, 2, 4, and 9 of Fig. "Fretting threatened engine parts".
Welded armors (e.g. Co alloys, stellite) have shown to be effective for turbine rotor blade shrouds. Hard cast-in or soldered-on wear plates or plasma spray coatings (chrome-carbide/nickel-chrome base; Ref. 6.3-1) are also in use.
The wear-protection layer must be inspected for possible negative propeties. In hard layers, especially, the dynamic strength in the contact region of the surface may be at risk for dropping suddenly. Problematic coatings include galvanic Cr and Ni coatings. The coating process itself must not damage the base material. With thermal spraying and PVD coatings, for example, a high substrate temperature decreases protective pressure-internal stress in peened coatings to a dangerous level. Overly high operating temperatures cause the coating elements to diffuse into the substrate, resulting in possible property changes (loss of dynamic strength, embrittlement) in both the coating and substrate. This effect limits, for example, the use of silver coatings above about 250°C, even though these coatings seem to perform very well on compressor blade roots, especially in russian engines. Silver coatings are used successfully on bearing carriers (silvered rings) and screw shafts (and, as the case may be, mating surfaces).
The after-treatment of coatings by hardening (e.g. ball peening) acts to increase dynamic strength by decreasing tension-internal stress and/or creating an advantageous surface structure.
As shown in Fig. "Minimizing fretting by low friction coefficient", the coefficient of friction, and therefore the stress in the contact surfaces, can be lowered with a glide coating (see ).
Loose metallic intermediate layers such as thin Cu plates (Fig. "Fretting processes in contact zone") between Ti surfaces are especially effective. However, foil inserts for blade roots have are not known to be effective in serial implementation. This is probably due to the fact that shifting or sliding-out of the foil inserts cannot be prevented with a sufficient degree of certainty. This would promote the failure of the tribo-system along with the risk of blade root failure and foreign object damage from foil fragments. On the other hand, the insertion of bushings in bores has found effective serial use (Ref. 6.3-2).
The most commonly used glide coatings are firmly adherent ones. These are graphite-based lacquers. MoS2-filled coatings can disintigrate and release sulfur at temperatures above about 300°C, which can damage the tribo-system. Application of a glide coating should be preceded by shot peening, in order to take advantage of the reservoir effect of the calotte structure. The operating life of lacquer glide coatings is shortened by wear, oxidation, or disintigration. Accordingly, the coefficient of friction increases with operating time, while the wear protection decreases. The stressing of the contact surfaces and risk of fatigue fractures increase with operating time. Overhaul intervals for engines with this type of coating must be set with these characteristics in mind.
The life span of contact surfaces that are not continually supplied with fluid or paste-like lubricants is greatly shortened. Multiple splining must be continually lubricated from a reservoir or oil circuit. Fluid or paste-like lubricants cannot be used in blading due to the operating temperatures, centrifugal force, and high surface pressure, among other factors.
Coatings must be simply and workably reparable. A coating that, when damaged, cannot be removed during engine overhaul in a serially applicable procedure (e.g. not damaging to the base material) is not suitable. The frequency of recoatings and/or repairs during the predicted life span of the part must be determined. For example, if every recoating requires metal removal, the allowable dimensional tolerances determine the possible number of recoatings. A further criterium is, if the customer wants to conduct the repairs himself and if the necessary requirements are available to him.
Mechanical Surface Treatment
The most important and most frequently used method (Fig. "Shot-peening as strengthening of surfaces") for improving operating behavior, especially for ensuring sufficient dynamic strength during fretting, is shot peening. This procedure does not noticeably alter the geometry and can also be done during repair. It plastically deforms metal surface, hardens the same, and induces pressure-residual stresses. Additionally, it creates a callote structure, corresponding to the geometry of the steel shot. This has a beneficial effect on the contact behavior of the two surfaces. Other hardening methods, such as rolling (in symmetrically rotating parts, such as shaft sockets), stamping ,and flaring (bores), are rarely used in fretting-sensitive areas during engine construction. A special profiling of the contact surfaces (Fig. "Blade foot wear symptoms") would be expected to be beneficial, but no serial implementation is known. Surface hardening through use of laser peening is being developed. This process utilizes shock waves in a coating on the surface to be hardened in order to achieve mechanical hardening.
Figure "Preventing fretting problems by construction": On the inside of cooled turbine stator vanes, thin-walled metal sleeves with profiled cross-sections (guide sheets) direct the cooling air along the blade wall. In the depicted scenario, the gap for the cooling air was maintained by internally-directed nubs on the blade wall (top left diagram). During operation, micro-movements occurred between the guide sheet and the nubs (vibrations, thermal strain). After a few hundred hours of operation, overheating of the stator vanes was observed. It was discovered that fretting wear caused small local break-throughs at the contact areas of the nubs (bottom diagram), causing a loss of cooling air.
The top right diagram shows a redesigned version, where the guide sheet was only affixed at both ends. This version has proven effective in operation.
Figure "Damage by measures preventing rubbing": In the case of a low-output helicopter turbine (top diagram), the basically correct design principle for avoiding wear points in the highly-stressed rotor disk hub region was to be implemented during overhaul mounting. The mounting that was usual until then resulted in the turbine-side central clamp bolt with a passband made contact with the wall of the hub bore of the disk. The relative movements between the clamp bolt and the hub bore (oscillations, thermal expansion) caused fretting points in the hub of the turbine disk, but not so much that dangerous damages occurred in these parts.
The contact between the bolt and the turbine disk was avoided by centering the bolt precisely during mounting. After a large number of engines were shipped, in the next few years frequent fractures were reported in the thinner expansion shaft of the clamp bolt, an area that is not subject to any fretting (middle diagram).
Inspections showed that the now freely mounted clamp bolts were no longer damped by wearing against the disk hub. Therefore, resonance could cause high-frequency oscillations of the bolt (most likely through gear frequency from the transmission). The antinode amplitude necessary for fatiguing the bolt was so small (due to the high frequency, supersonic), that the relatively small play between the bolt and disk was not bridged, i.e. not damping occurred (bottom diagram).
This case illustrates how carefully alterations to proven engine configurations must be examined and considered in order to avoid a worsening of the situation.
Figure "Minimizing fretting by low friction coefficient" (Ref. 6.2-1): Because a high coefficient of friction between the contact surfaces creates high stress levels (Fig. "Coefficient of friction indicating stressing"), low coefficients of friction are desirable. The diagram shows coefficients of friction of metal surfaces that have been treated in different ways. It is interesting to note that a shot-peened surface interacting with a oxide-blasted surface lowers the coefficient of friction considerably when compared with two oxide-blasted contact surfaces.
The bottom curve shows that it evidently sufficiently reduces the coefficient of friction if just one surface has a glide coating and the other is simply shot-peened. This is an important realization if one surface cannot be shot-peened due to inaccessability (e.g. disk slots).
Figure "Shot-peening as strengthening of surfaces" (Ref. 6.3-4): Shot peening involves balls of steel, glas, or ceramic striking the surface being treated with high speed, causing plastic deformation. Shot peening is the most common way of hardening surfaces during engine construction. The dynamic strength is increased by various degrees, depending on the material and hardenings and/or pressure residual stress. When combined with a glide coating, fretting is minimized (Fig. "Fretting at titanium alloys"). Shot peening can be effective up to several tenths of a millimeter below the surface, depending on intensity and material. Aside from a lesser effective depth, the shot-peening effect is more pronounced the harder the material is (Ref. 6.3-5, Fig. "Effects of shot-peening").
In order to achieve optimal dynamic strength, it can be advantageous not to peen the base material before coating it, but to peen the coated surface. This is especially true for hard coatings of chrome and nickel with high tension residual stresses.
To ensure optimal results from shot peening, important parameters must be met. These include blast intensity (measured in almen intensity, which corresponds to the deflection of a sample strip that has been blasted on one side) and solidity ratio of 100% (i.e. the entire surface is covered with impact marks). Additionally, fretting seems to be arrested by the typical callote-structure of the peened surface.
Currently, all rotor blade and disk part surfaces subject to high dynamic loads (LCF, HCF) are normally shot-peened during production. These include typical fretting zones such as lay-on surfaces of blades and disks, bolt bores, and flange lay-on surfaces. The dynamic strength-increasing shot-peening effect is especially pronounced in notch zones.
The optimal angle for shot peening is 90°. In order to ensure this, a hard baffle plate is used for difficult-to-reach surfaces (taper pin, prism, bottom left diagram).
Operating temperatures that cause a noticeable drop in the effectiveness of shot peening through creeping limit the applicability of this process (right diagram).
Figure "Effects of shot-peening" (Ref. 6.3-5): The harder the material, the highter the pressure residual stress that can be induced (top left diagram) and therefore the greater the expected effect on the dynamic strength. This effect can be attributed to the depth in which noticeable pressure residual stresses were induced. In heat-treatable steels, this can be up to several tenths of a millimeter. Due to the greater hardening depth in softer materials (top diagram, graph), it can be be expected that deeper scratches will be mitigated. This protective effect with deepened action contributes substantially to the protection and regeneration of fretting surfaces. It is understandable that effective depth increases along with shot-peening intensity (almen intensity=deflection of a sample strip that has been blasted on one side). In accordance with the top diagram, the maximum depth in steels is between 0.5 and 1 mm, in Ti6Al4V, the standard alloy for compressor blades, roughly 0.5 mm. Protection against damage in the region of tenths of a millimeter should therefore benefit from shot peening.
Shot peening of notches can also considerably reduce the notch effect. This is especially true for notch-sensitive materials (bottom diagram).
6.3-1 J.D. Schell, K.P. Taylor, General Electric Aircraft Engines, “Wear of Jet Engine Components”, ASM Handbook, Volume 18 “Friction, Lubrication and Wear Technology”, pages 588-592.
6.3-2 J.S. Alford, General Electric Co., “Design Criteria and Configuration for Long-Life Aircraft Gas Turbines”, SAE-Paper No. 670344, Proceedings of the National Aeronautic Meeting, New York, April 24-27, 1967, pages 1-16.
6.3-3 H.-J. Böhmer, “Wälzverschleiß und -ermüdung von Bauteilen und Maßnahmen zu ihrer Einschränkung”, periodical “Materialwissenschaft und Werkstofftechnik” 29, pages 697-713 (1998).
6.3-4 P. Adam, “Fertigungsverfahren von Turboflugtriebwerken”, Birkhäuser Verlag 1998,, ISBN 3-7643-5971-4, page 110.
6.3-5 Published by the “Metal Improvement Company”, pages 1-47.