23.4.2 Sliding seals
188.8.131.52 Rotary shaft seals
The term sliding seals was choosen, because these „contact seals“ only function, when in
the sealing effective region between the against each other moving a lubrication film forms.
This minimizes wear and is so requirement for a long time service. At the same time it acts stabilizing
at the operation temperature in the sealing gap. An especial advantage against labyrinth seals lays
in the tightness during stand still. To such seals count mechanical face seals (carbon seals,
sliding ring seals, chapter 184.108.40.206) and radial seals.
In the literature we find the seal type under different markings like (rotary) shaft seal, radial seal ring, shaft sealing ring. Often it is also named after a known producer as `Simmerring'. These seals are frequently used in aeroengines to tighten against fluids. To these belong aeroengine oil, fuel and hydraulic fluid. The limited temperature stability of the elastomer, from which the sealing element (sealing lip) consists, determines the use. Main field of application are accessory gears, pumps, control units and oil lubricated accessory devices.
Up tp the 80s there was obviously a faulty imagination about the mechanism of the sealing effect. It wass supposed, that only the lubrication film (meniscus) seals (Lit. 220.127.116.11-3). Not before the following years (Lit. 18.104.22.168-1 and Lit 22.214.171.124-2) the actual sealing mechanism was recognised during tests. This is based on special pressure distribution in axial direction under the sealing lip. For this, an axial oriented micro topography acts, which elastomers typically forms at the sliding surface (Ill. 126.96.36.199-1) during the running-in phase.
With this the running-in after the assembly of the sealing ring is crucial for the operation behaviour. This finding is the cause, why especially in elder literature and such of newer literature build on it, some influences have been insufficient evaluated. Corresponding less was the success of proposals for a remedy. This produced confusion in practice. In this chapter the corrected awareness of the hydromechanical sealing effect is used.
Fig. "Sealing effect of radial ring seals" (Lit. 188.8.131.52-1 up to Lit. 184.108.40.206-5): To avoid
problems with radial sealing rings it is helpful to understand its
The sealing effect of a radial sealing ring stands in tight connection with the sealing film between sealing lip and shaft surface. The little width of the sealing face at the sealing lip is sufficient for hydrodynamic effects. So a sealing film, comparable the bearing lubrication film of a journal bearing, develops. The optimum lubrication film thickness is about 0,25 µm. A larger film thickness increases the leakage rate, a smaller rises friction and wear. For long time there was the opinion, that at the side of the medium to be sealed (meniscus), the surface tension of the sealing film is crucial (Lit. 220.127.116.11-3). This assumption however could not be confirmed in tests. Instead, these have showen the following (Lit. 18.104.22.168-1 and Lit. 22.214.171.124-2):
The special pressure distribution together with a, for elastomers typical microscopic structure of the run-in surface from the elastomer, is crucial for the sealing effect. It is not the forming of the meniscus!
The necessary conditions depend from the tribological conditions in the sealing gap. So a sealing ring is similar to a micro pump (Fig. "Conveying effect in the sealing gap"). Its conveying effect/sealing effect is, according to test results (Lit. 126.96.36.199-2) depending from several influences:
- Temperature of the medium: Does it rise, the sealing effect deteriorates.
- Rotation speed of the shaft: The sealing effect increases with the speed/sliding speed.
- Wear pattern of the sealing edge (Fig. "Conveying effect in the sealing gap")
- Deviation of the position from the shaft to the sealing ring (Fig. "Influence of shaft sliding surface dimension")
The contact pressure force of the sealing lip determines the axial width of the sealing face/sealing edge. It must have a minimum height, to prevent, that the lubrication film is bridged.
- Surface of the shaft (Fig. "Sliding surface machining problems").
Fig. "Conveying effect in the sealing gap" (Lit. 188.8.131.52-1): A suitable
wear pattern of the sealing edge is crucial for the
sealing effect. This develops during run-in with slight removal of the elastomer. The microscopic
structure of the race track is for rubber like materials particularly oriented across the circumferential
direction (sketch above). It is similar chattermarks
and is elastically distorted (sketch below) from the
friction forces during operation in the lubrication gap. So the distortion depends from the distribution of
the friction forces. The maximum is shifted to the fluid side. It is determined from the
unsymmetrical distribution of the contact pressure (Fig. "Sealing effect of radial ring seals", diagram in the lower sketch).
That a conveying to the inner side i.e. the sealing effect develops the contact pressure maximum, respectively the maximum friction forces must be shifted to the inner side. Then the orientated surface topography of the elastomere bends so, that at the air side the at the shaft surface moved material is `caught' and transported to the inner side (arrows in the sketch below).
The developing conveying of the rotating shaft transports continuously, independent from the sense of rotation, air to the side of the medium to be sealed (inner side). This process is called pump effect or conveying effect of the sealing ring. It determines the sealing effect.
It is influenced from different effects:
- Shifting of the maximum of the contact pressure force. Is it not far enough from the inner side or even far to the outer side (to the air side), fluid will be transported to the outer side and a leak develops. Such a situation exists during reversed installation of the sealing ring or an axial shifting of the pressing on spring.
Increase of hardness and stiffness of the elastomer, caused from overheating/hot run and aging (Fig. "Problems of radial sealing rings"). Hinders a higher stiffness of the elastomer the necessary elastic deformation, the risk of a leak increases.
- At bad run-in conditions the necessary micro structure at the sealing lip can not form. Then it is changed unfavourable by wear processes or hot run/overheating. With this the sealing effect drops.
Fig. "Influence of shaft sliding surface dimension" (Lit. 184.108.40.206-1): Deviations from the aligned centrical position of the shaft against the sealing ring can obviously have very different consequences.
- Deviation of the concentricity/lay/revolving bending up to a magnitude of tenth millimeters can improve the sealing effect against expectation. The effect of a so called `lay' of the sealing lip differs from an offset. A lay only occurres during rotation. It stresses the sealing lip throughout the whole circumference.
- Radial offset, often only named as `offset' (sketch in the middle), exists contrary to the lay also during stand-still. The sealing lip is at the circumference locally deflected, as well during stand still as also during rotation. This can lead here to deteriorations from thermal overload and/or plastic deformations (creep).
- Misalignment of the shaft (sketch below): In the range of an angle degree the sealing effect seams not to decrease. This is also true for grinding marks, which slightly deviate from the circumferential direction.
- The axial movement of the shaft is enabled from the bearing clearance and supports the conveying effect. With this it increases the sealing effect.
Note: The run-in process is of crucial meaning for the sealing effect of a radial sealing ring in the later service. Therefore specifications, which influence this process, must be kept most exactly.
Fig. "Sliding surface machining problems" (Lit. 220.127.116.11-1): The
topography of the shaft surface can be of considerable
meaning for the sealing effect. Usually grinding
surfaces of the plunge cut method are concerned (last
sketch below). Anyway it must always be reckoned with angular oriented machining marks. The twist effect
of theses microscopic traces was frequently overestimated. This changed with the awareness, that
the conveying effect acts by the structure of the elastomer sealing surface as the actual sealing
mechanism (Fig. "Conveying effect in the sealing gap").
Advantage can be taken from the conveying effect of the twist from the grinding marks also if not crucial. For this, the direction of the pitch at the rotation direction and sealing is adjusted (sketch above). In such a case, the twist direction must be indicated unmistakable with a sign on the part.
Fig. "Problems of radial sealing rings" (Lit. 18.104.22.168-3 and Lit. 23.4.1-5): Also radial sealing rings are
exposed specific failure mechanisms (Fig. "O-ring failures and its causes" and Fig. "Inner crack formation at O-rings").
Overheating of the sealing lip: The sealing function must be kept over sufficient long operation periods. For this the sealing lip must slide on a lubricant film over the shaft surface (details left). If this sealing film drops out only a short time (seconds), too much friction heat develops (Lit. 22.214.171.124-5). Embrittlement and crack formation in the sealing lip decrease the thightness. An unsuitable microscopic structure and the shift of the contact pressure distribution (creep, wear, Ill. 126.96.36.199-1) act with this especially unfavourable. An increase of the hardness with a decrease of the elastic expansion drops the deflection of the sealing lip, which is necessary for the sealing effect (Ill. 188.8.131.52-1.2). This can already lead to a leak without crack formation or wear. The typical crack appearance of an overheating is a field of parallel radial cracks. Often is the crack formation limited to the circumference segment in which the `dry run' was most intense (sketch above left).
This risk is from experience highest during the first run up/sealing run after the installation of a sealing ring. The reason is an insufficient lubrication of the sealing surfaces (lip, shaft) during assembly.
Is a sealing ring run-in as a start successful, further start up cycles in opreation typic time periods are no problem. Merely after long stand still periods like a 'mothballing' of the aeroengine/airplane (Fig. "Aged O-rings") or the storage of the components (e.g., gear) attention is necessary. We should assure us in the manual, if there are for the first run certain measures prescribed to guarantee the lubrication of the sealing rings. Shaft surface, roughness and damages: The wear of the seal lip depends primarily from the roughness of the shaft sliding face. Roughness tips or notches can let the oil film get locally thicker or to break down. Do roughness tips contact, it comes to increased wear. On the other hand it must be considered, that a too smooth shaft surface can deteriorate the wettability and adherence of the oil film. Therefore the roughness should lay, independent from the circumferential speed, between 0,25 µm and 0,5 µm.
Even smallest scratches and scores in the track, running angular to the sliding direction/circumferential direction, can be responsible for an unexpected large leak. Also pronounced angular running grinding marks (at axial feed spiraling) are dangerous if they disturb the hydrodynamic pumping process (Fig. "Sliding surface machining problems"). Thereby the progression of the damage relatively to the rotating direction of the shaft and the tightening direction, plays a role (detail below left).
Especially must be payed attention at this problem during the overhaul of shafts. The race of the sealing ring must be removed, chromium plated and ground, because of wear (volume 1, Ill. 184.108.40.206-4).
Do anti friction bearings of the shaft have too much axial clearance, this can perform an oscillating movement with hydrodynamic pumping effect, relatively to the sealing lip. Movements of usual bearing clearances can even support the sealing effect (Ill. 220.127.116.11-1).
Aging of the sealing lip: The operation tempreature of the sealing lip is crucial for the aging. With this it influences the lifetime of the seal. In the most cases, the cause for a too high temperature is a too high radial pretension of the sealing lip. For this possible causes are wrong contact pressure fields or a too small diameter of the sealing edge. Thereby it is remarkable:
A too high radial pretension does not produce a better sealing effect. It rather leads to a deteriorating overheating with the failing of the seal.
A rise of the temperature exponential shortens the life time. It leads, corresponding the used elastomer, to a rise in hardness (Fig. "Conveying effect in the sealing gap"), embrittlement and crack formation. The dependence of the aging respectively lifetime of the seal, shows material specific the diagram above right. For the aging of the elastomers also can serve data in connection with O rings of Fig. "O-ring failures and its causes" and Fig. "Influences of O-ring materials".
Brittle crack formation during normal operation temperature is explained as result of shock loads (e.g. pressure shocks) or vibrations of the sealing lip. From experience brittle aging cracks occur only during disassembly and/or mechanical deformation.
Note: Therefore the visual inspection of the seal condition allows no dependable evaluation of an aging.
Wear as consequence of a contamination: Foreign particles like abrasion, coke and dust in the oil can promote unacceptable wear, like too high roughness. Unfortunately in the oil system develop coke particles with failure potential, which are so small, that they will not be separated from the usual oil filters (Fig. "Comparison of sizes from particles in oil").
Therefore clean oil is a requirement for a wear free operation of a seal ring. Also particles at the outer side can get into the lubrication gap under unfavourable condition. Abrasive acting particles can also develop inside the lubrication gap. Too high friction at the sealing lip heats the oil film in a manner, that the oil cracks and coke particles form (Fig. "Formation mechanisms and oil coke features").
Wear of the shaft race: From experience also this, comparatively very hard surface, underlies wear. Obviously sufficient hard particles are embedded in the sealing lip, to machine a narrow, polished looking groove into the shaft. As long as there is no offset of the sealing lip, a leak must not be expected. However, such a groove is a reason for a repair during overhaul. Thereby the shaft is sufficient ground and then coated, mostly chromium plated (volume 4, Ill. 18.104.22.168.3-1). Although this plating has a high hardness, intense groove formation will be observed.
Damages during assembly are the most frequent cause for the failing of a radial seal. Therefore during handling of the seals especially attention is needed. This is also true for the storage. The seals must be in containers/packages, which guarantee a clean and dry atmosphere. A treatment as `bulk good' should be unacceptable, because of the damaging danger. Before the installation every seal must be checked by an experienced expert. Special attention needs
- Damages of the sealing lip.
- Faultless seat of the garter spring (contact pressure spring).
- Celeanliness, respectively dust-free of seal and shaft.
Immediately before the installation the seal can be
lubricated, according to the instructions
(e.g., grease or oil), to minimize the joining forces.
Note: During installation of a radial seal ring, it must be payed attention, that correspondent the specifications/instructions the sealing faces (sealing lip, sliding face/race of the shaft) must be treated. If necessary, care must be taken for sufficient lubricant (oil, grease).
Fig. "Damaging the sealing lip by handling" (Lit. 22.214.171.124-3 and Lit. 126.96.36.199-5: The
assembling method (lower sketch right)
depends from the insertion direction. It is important, if the seal
first is moved onto the shaft and then into
the casing (seat). The other possibility is first to insert the seal into the
seat and then to introduce the shaft. If possible the first method should be preferred. The introducing face side of the shaft should
have a chamfer (upper detail) or a radius. For
repaired sliding faces especially attention must be
payed, that they are free of burrs (volume 4, Ill. 188.8.131.52-3) and have
soft edges. That is also to apply to small break outs at the
edges of a cromium coating (volume 4, Ill. 184.108.40.206.3-5). For the safety, a
sleeve with a favorable edge can be slided on the shaft (right sketch). This is especially advisable, if the shaft
has flutes. A slight rotation movement
can ease the inserting.
The symmetric and gentle forced fitting, with this the alignment of the seal in the casing/seat, can be assured with a bell shaped tool (sketch below left). Thereby the outer seat of the seal can be lubricated, as specified. Also the entrance edge of the locating bore in the casing should have a chamfer (detail above left). A continuous forced fitting process can be executed without overload. Thereby the sticking friction is avoided. Under unfavorable assembling conditions, it can be necessary to remove at first the tension spring of the seal and to attach it afterwards after the positioning at the seal again. That is recommended if the seal must be inserted reverse side.
Example 220.127.116.11-1 (Lit. 18.104.22.168-4):
In several cases a `teflon seal' (seal with
PTFE sealing lip?) in the inlet gearbox (= IGB)
failed. Seemingly parts of the seal
membrane separated. These blocked the scavenge oil
sieve. This lead to an oil accumulation in the gear.
The consequence was oil entry into the high
pressure compressor with Oil vapour/smoke in
Comment: From the papers available, it could not clearly seen, which type of seal is affected. Obviously primarily concerned was an assembly problem. It rather may be a problem of the material, production or design.
Fig. "Remedies at shaft seals" (Lit. 22.214.171.124-3: This chart should support the practitioner. However it
can not replace informations in manuals and specifications. That must be applied for
remedies for seal failures.
Realizing (features) the problem is premise for a constructive remedy. For this the failure mechanism with the specific causes must be identified.
Evidence of a maintenance problem: At this point begins the difficulty with the identification of the failure attribute. Not always it attracts our attention. If its occurence lies in spacial distance (oil drops , oil track) to the leak, sometimes this features can not be connected.
Next arises the question, if the observation can be seen as normal or if it is unusual. A desicion needs satisfying informations in manuals, together with expertise and experience. For example it must be decided if bleeding oil is to be seen as a leakage. The details in the manual need an interpretation if the instruction allows a discretion. In such a case enough technical qualification is necessary. We must know concerned functions and likely consequences.
For the attendant or supervisor, detectable outside features of a failures of a radial sealring are:
- Alarming high leakage rate. For this classification there should be sufficient detailed information in the manual.
- Observable damaged seals: Crack formation, mechanical damage.
- High friction forces, that arise in the sluggish behavior of the shaft. Here also experience is needed for an evaluation.
Is the indication recognized respectively classified as a failure, the cause must be
suggested from the relevant features of the finding.
126.96.36.199-1 M.Kammüller, H.K.Müller, „Physikalische Ursachen der Dichtwirkung von
Radial-Wellendichtringen” , ATZ Automobildtechnische Zeitschrift“ 88 (1986) 1, page 39-45.
188.8.131.52-2 W.Hermann, H.-W.Seffler, „Neue Erkenntnisse für den Abdichtungsmechanismus von Radial-Wellendichtringen” , ATZ Automobildtechnische Zeitschrift“ 87 (1985) 9, page 475-484.
184.108.40.206-3 M.W.Brown, „Seals and Sealing Handbook” , Verlag „Elsevier“, page 137-166.
220.127.116.11-4 P.Lironi, „CF6-80C2 engine history and evolution” , Zeitschrift „Engine Yearbook 2007“, page 80-85.
18.104.22.168-5 „Mitteilungen des Simrit-Werkes Nr. 700/60” , Publisher: Carl Freudenberg, 1960.