Fretting Wear in Tooling and Tool Steel Parts
Fretting is what happens to two surfaces that were meant to stay still relative to each other and do not. Vibration and a small oscillation move them, the oxide that forms cannot escape the contact, and the damage that results is measured in fractions of a micrometre while the fatigue strength behind it falls by a factor of two to five. This page covers how fretting is recognised on a part, the slip amplitude at which it starts and the amplitude at which it stops being fretting at all, where it does real damage on machinery and tooling, and which measures actually reduce it.
What fretting is, and how to recognise it
Fretting is adhesive wear that needs relative movement to happen, and the movement is small. Two tightly fitting surfaces are subjected to a cyclic relative motion of extremely small amplitude, usually the result of vibration in the machine around them, and the contact between them has to have some slip for damage to occur. Fretting is frequently accompanied by corrosion, and on ferrous parts in air the product is the reddish brown ferric oxide that gives the process its other name. Mixed with oil or grease, that debris is known in workshops as cocoa, blood or red mud.
The practical tell is where the debris appears. In a joint that is lubricated well enough that ordinary corrosion is not expected, a reddish brown powder is evidence of fretting, and that reading holds whether the joint is a press fit, a keyway or a bolted flange. In an unlubricated component the same powder may simply mean wear, because there the metal was free to oxidise anyway. Fretting also occurs in materials that cannot oxidise in the ordinary sense, including gold, platinum and cupric oxide, and in an inert atmosphere it produces very little debris at all, which is why the presence of debris is evidence and the absence of it is not proof.
| Term | What it describes | Where it is met |
|---|---|---|
| Fretting wear, or fretting corrosion | Adhesive wear driven by small amplitude oscillation, with oxide debris as the visible product | Any joint that was meant to be fixed but moves |
| Fretting fatigue, called contact fatigue in the source | Fretting damage at a loaded contact that also starts a fatigue crack | Press and shrink fits, blade roots, rivet holes |
| Impact fretting, also called impact fatigue | The normal force oscillates as well, so the surfaces lose contact in each cycle and the contact hammers | Tube supports, snubbers, loose assemblies |
| False brinelling | Craters in a raceway shaped like Brinell impressions, produced by vibration while the part is stationary | Bearings in transit or on a vibrating machine |
| Fitting rust | Fretting debris trapped between clamped gages or shims | Gage blocks and shim packs |
| Reiboxydation, and corrosion de trepidation | The German and French names for the same phenomenon | Non English literature |
The names fretting damage is published under, and the case each one describes. Compiled from the opening section of the article Fretting Wear Failures.
The amplitude decides everything
The small amplitude of the movement is the reason fretting damages a joint that a normal wear calculation would call safe. In a contact analysed as a ball on a flat, the centre of the contact stays locked and slip occurs only in an outer annulus, so damage begins as a ring. As the amplitude rises the ring widens, but the outer radius stays fixed, and it takes a larger amplitude again before slip covers the whole contact. The measurements below are the thresholds the source reports on steel, and they set the scale of the whole problem. Damage is detectable at an amplitude of 0.06 micrometres, which is a movement small enough to be invisible, and the amplitude at which the process turns into ordinary reciprocating sliding is over a thousand times larger.
| Slip amplitude | What was observed at that amplitude | Condition of the test |
|---|---|---|
| 0.06 µm (2.4 µin.) | Damage is detectable on steel | Ball on flat, 88 N load |
| 0.1 µm (4 µin.) | A narrow annulus of damage appears | Ball on flat, 88 N load |
| 0.53 µm (21 µin.) | Damage is measurable by profilometry of the flat | Ball on flat, 88 N load |
| 2.8 µm (110 µin.) | Damage is described as severe | Ball on flat, 88 N load |
| about 25 µm (1000 µin.) | Slip becomes total across the contact, and above this the damaged radius grows linearly with cycle count | Ball on flat, 135 N load, 8 to 25 µm amplitude, 50 Hz |
| above 100 µm (0.004 in.) | The specific wear rate becomes constant and matches unidirectional or reciprocating sliding | Several contact geometries, plotted as specific wear rate |
Slip amplitude thresholds measured on steel contacts, from the first detectable damage to the amplitude at which the process stops behaving as fretting at all. Source, the amplitude of slip section of the article Fretting Wear Failures.
The low relative velocity that comes with a small amplitude changes the contact too. At a frequency of 25 hertz and an amplitude of 25 micrometres the average velocity of the two surfaces is 1.25 mm per second, against sliding speeds of the order of 1 m per second in a bearing, so the debris that forms stays where it was generated instead of being thrown clear. It can be compacted under the contact load and it may never be exposed to the atmosphere at all, which is why the debris inside a fretted joint often looks different from the loose powder outside it. High spots also print themselves onto the opposing surface where that surface has the better finish, so the roughness of one part of a joint shows up as a pattern on the other.
Where it does real damage
The common sites are the ones where a fixed joint carries a load. Contacts between hubs, shrink and press fits and bearing housings on loaded rotating shafts are particularly prone, and on a shaft the problem is more one of fatigue than of wear, because the movement that damages the surface comes from the alternating stress inside the shaft and the damage then becomes the crack starter for the fatigue failure. In thin shell bearings the interference fit between the shell and its housing is what keeps the assembly still, and where the contact pressure is too low the shell moves, frets, and the bearing fails from the outside in. Splines and flexible couplings that exist to accommodate a slight misalignment fret severely by design. In turbines, the disk to shaft joint and the dovetail or fir-tree fixing of the blades are the sites where fretting starts fatigue cracks, while the snubbers at the blade tips lose material and allow the blades to vibrate harder.
Riveted and bolted joints are the same problem in a different form. There is no such thing as a static joint on an aircraft, and even a single riveted joint offers three separate fretting sites, between the sheets, between the rivet head and the sheet, and between the shank and the hole. Leaf springs fretted between their plates, and a wire rope frets at every interwire contact, where in a locked coil rope the debris has a larger volume than the metal it came from and forces the wires apart, letting the lubricant out and the atmosphere in. Steam generator and heat exchanger tubes fret against their supports under flow induced vibration. Fretting needs no load cycle to appear either, since a bearing carried on a truck or a train frets between the balls and the race while the shaft is stationary, which is the failure the industry calls false brinelling.
The fatigue side is the expensive one
Fretting damage and fatigue combine, and the combination is far worse than either alone. A test reported in 1941 found that fretting pits on the gage length of a steel fatigue specimen reduced fatigue strength by 13 to 17 percent. Later work on the two acting together, which is the usual case in practice, gave strength reduction factors of 2 to 5 and greater. The reason is visible in the life split. In ordinary fatigue, crack initiation may account for about 90 percent of the total life, so a crack that starts early costs little of the total. When fretting is present the same initiation accounts for about 55 percent of the life, which means that almost half the fatigue life is consumed before there is any crack to detect. That is the mechanism by which a joint that carries load and moves shortens the life of the part it holds, and it is why the roughness, coating and peening measures matter more for the crack than for the metal loss.
What actually reduces it
The first question to answer about a fretting contact is whether the movement is force controlled or displacement controlled, because the two need opposite treatment. In force controlled fretting the contact is in partial slip, and raising the coefficient of friction or the normal load reduces the amount of slip, which reduces the damage. In displacement controlled fretting the slip covers the whole interface and the amplitude is fixed by the surroundings, so the way to reduce wear is to reduce the normal load and the friction instead. Choosing the wrong one of those two levers makes the problem worse.
Beyond that, the source works through design, surface condition, coatings, inserts and lubricants, and each one has a limit worth knowing before it is specified. One detail runs against intuition and is worth stating plainly, which is that a rough surface frets less than a highly polished one, so a mirror finish on a mating face is not a fretting cure.
| Measure | What it does | The limit on it |
|---|---|---|
| Remove the relative movement | Takes the cause away instead of treating the symptom, and it is the only measure that does | A joint usually exists because two parts have to be assembled, so the movement can rarely be designed out completely |
| Change the contact pressure | Under partial slip at a constant load, a higher contact pressure reduces the area that slips | Raising the pressure adds a fatigue problem of its own at the same joint |
| Roughen the surface, by shot peening with glass beads or steel shot | A rough surface frets less than a highly polished one, and peening also work hardens the surface | The compressive stress that peening leaves does little for the wear itself, but it does a great deal if a fatigue crack is trying to start |
| Diffusion or bombardment treatment, such as carburizing or nitriding | Hardens the surface in place, and carburizing and nitriding of steel are established against fretting as well as against gear wear | The treatment has to suit the grade, and any growth has to be allowed for in the drawing |
| A hard coating applied over the surface, or a softer deposited layer | Puts a wear resistant material at the contact | A coating harder and more brittle than its substrate cracks and breaks up when the substrate yields, and a very hard coating such as titanium nitride then abrades the counterface |
| An insert between the surfaces, a soft metal shim or a low modulus polymer | The insert absorbs the movement by plastic or elastic deformation, and a metal and polymer combination carries the conductivity of the metal with the compliance of the polymer | Needs room in the joint, and the material has to suit the service temperature. Bronze filled PTFE against steel has shown very low friction and undetectable wear |
| A liquid or solid lubricant | Lowers the friction at the contact and carries debris away | Keeping the lubricant in the contact is the difficulty, and a solid film such as molybdenum disulphide wears away |
The measures the source lists against fretting damage, with the mechanism each one works by and the point at which it stops helping. Source, the prevention of fretting damage section of the article Fretting Wear Failures.
What this means on a tool steel order
Fretting is a joint design problem that ends up as a tool steel question when the measures chosen are metallurgical. Three of them belong in a purchase discussion. A nitrided or carburized surface on the parts that move against each other, because those treatments are established against fretting as well as against wear, with the case depth chosen for the service rather than for the drawing. A steel that is clean and sound enough to start a fatigue crack late, because the crack is what costs the life and the debris is only the visible part of the damage. And a hardness and toughness combination that suits the joint rather than the tool, since a coating harder than the substrate it sits on cracks and breaks up as soon as the substrate yields. The way the pits and marks left by fretting are read against other damage modes is on the damage mode identification chart, the fatigue crack that follows is covered in fatigue failure in tool steel, and the surface treatments themselves are gathered under steam treatment for tool life and in hot work tool steel heat treatment. Wear resistant grade families and their measured test data sit in abrasive wear resistant tool steels, adhesive wear resistant tool steels and tool steel abrasive wear test data, the related impact side is on impact wear and spalling of tooling, and the way damage of any kind is classified before it is explained is set out in tool and die failure analysis. Where the joint is a shrink fit on a hardened shaft, the residual stress the fit sits in is reported as described in measuring residual stress in tooling.
Before you act on a fretting diagnosis
A reference page, it is not an Aobo Steel specification, and the figures above are the values published for the steel contacts and the test conditions named. Fretting damage depends on the amplitude, the load, the frequency, the environment and the finish of the actual joint, and a threshold measured on a ball on flat contact does not transfer directly to a press fit or a spline. Reddish brown debris in a lubricated joint is evidence that fretting is happening and not a measurement of how far it has gone. Where the part also carries a cyclic stress, the crack is the item that decides the life, and it deserves its own examination before the wear is repaired.
Source: ASM Handbook, Volume 11, Failure Analysis and Prevention, ASM International, 2002.