7.2.1 Fundamentals of Function, Construction, and Operating Behavior

Labyrinths are the most frequently used gas seals in engines. They are used to seal air and hot gas in the main gas flow as well as for sealing oil/air mixtures in bearing chambers. Labyrinths can be closed rings, as in intermediate stage seals of the rotors, or they can also be segmented and used as rotor blade shrouds. The operating loads on labyrinths are as varied as their applications. This causes specific problems and damages. The operating behavior of many other components depends on the functioning of the labyrinth. For example, the thrust loads on the main bearings are determined by the leakage from the labyrinths.

Configuration and mechanical loads:
Labyrinth seals are only effective and safe during operation if they are properly designed.

The sealing of a rub system depends on more than just the size and geometry of the clearance gap. The surface structure of the coating on the housing and stator also plays an important role (e.g. honeycomb), as does open porosity (e.g. metal felt), through which the opposite surface can be “bypassed”.
Erosion or breakouts due to dynamic fatigue and thermal stress can noticeably increase clearance leaks over the running time (deterioration). Therefore, it is important to ensure sufficient long-term resilience, especially in soft abradable coatings.

The thermal conductivity and roughness of coatings affects the heat transmission and thermal transfer. Depending on the coating type, they accelerate or slow the heating-up of the housing walls and rotor components. This affects especially clearance changes during unsteady operation. In high-pressure compressor housings, for example, this characteristic has been noted and put to practical use.

Labyrinth fins are usually on the rotating labyrinth part. There are several reasons for this:

  • The damping effect (slowing the circumferential flow in the labyrinth chambers)of open standing coating structures.
  • The problematic bond strength and shear resistance of coatings from critical bonding techniques (e.g. adhesive, spraying, soldering) prohibit high tensile loading such as occurs under centrifugal force.
  • Minimizing the heating-up of the rotor during rubbing.

The leak flow through a labyrinth with fins usually corresponds to 60% of that which would occur if the labyrinth were smooth, i.e. without fins. The repeated restriction of the flow over the labyrinth fins as well as the relaxation over the following labyrinth chambers (circumferential grooves) lowers the flow energy by about 40%. The leak flow is a function of the labyrinth geometry and the operating parameters in front of and behind the seal. The labyrinth geometry is determined by the following important values (Ref. 7.2.1-2):

  • Labyrinth diameter in the gap
  • Number of labyrinth fins
  • Axial labyrinth width (gap length)
  • Radial distance from rubbing surface to base of groove (chamber depth)
  • Axial spacing of the fins (chamber width)
  • Fin shape (fin radius)
  • Fin angle
  • Chamber shape
  • Volume of the circumferential grooves (chambers) between the fins
  • Radial gap between the fin and rubbing surface (gap width)
  • Eccentricity between fins and rubbing surface
  • Cross-section (centrifugal force, heat dissipation, stiffness)

Important operating parameters are:

  • Gas pressure in front of and behind the seal
  • Gas temperature around the seal
  • Rubbing, material removal
  • Erosive stress
  • Coke formation

Developing and proving the operational effectiveness of rub systems requires a multitude of technological tests corresponding to the load spectrum. Since the material combinations labyrinth tribo-systems are often the same as in the seals at blade tips (see Fig. "Properties and tests of rub systems"), some similar tests can be applied. The following are properties that are important to test:

  • Rubbing behavior at slow and fast infeed rates in new and aged tribo-systems.
    • Wear
      • Distribution on the rubbing partners
      • Extent
      • Distribution around the circumference
    • Heating-up and damage from overheating
      • Strength decreases
      • Hot crack initiation
      • Ignition of oil fires (in bearing chamber seals)
    • Tendency to depositing
    • Form stability (susceptibility to catastrophic failure)
      • under normal operating loads
      • during rubbing
  • Seal effectiveness under relative motions typical during operation
    • Influence of the radial clearance on seal effectiveness
    • Influence of axial deflection
  • Dynamic behavior under engine-specific conditions
    • Sensitivity to externally initiated vibrations
    • Sensitivity to internally initiated vibrations
  • Aging at operating temperatures and run times
    • Oxidation
    • Corrosion
    • Reactions with contaminants (e.g. dust)
    • Erosion
  • Material removal
    • Damaging effect on other parts
      • Clogging-up hot parts
      • Erosive effects
      • Damage to bearing runways
  • Reparability
    • Weldability of the fin material
    • Removability of the rub coatings (also aged coatings)

Figure "Labyrinth seals in turbines": The functioning of an engine is highly dependent on the pressure differences in the different part zones and the maintenance of these differences over the entire operating period. These pressure differences are closely connected to the air system of the engine. This does not affect only the functioning of the engine, but also its operating behavior and efficiency, as well as the durability of important components (e.g. cooled hot parts or bearings). In this diagram of a triple-shaft engine type, various labyrinths are depicted along with their typical functions and arrangement.

“A” : Labyrinths in the front disk region of the high-pressure turbine (HPT) ensure cooling air flow to the rotor blades and prevent hot gas from entering.

“B”: These labyrinths influence the cooling between the HPT and middle-pressure turbine (MPT) as well as the thrust loads on these components and the cooling of the rotor disks.

“C”: Sealing the bearing chambers against escaping oil fog. Usually in combination with barrier air and suitable pressure levels in the sump.

“D”: These labyrinths seal the disk region and prevent impermissible amounts of hot gas from leaking between the stator assembly and rotor.

“E” : Sealing the blade tips opposite the housing. These seals are integrated into the rotor blade shrouds (Fig. "Turbine blades shroud labyrinths"). Their effectiveness influences the operating behavior and efficiency of the turbine.

“F” : Labyrinths that maintain the pressure levels in the disk region and therefore the thrust loads (Fig. "Labyrinth seals influencing thrust bearings"). This makes these labyrinths necessary in order to reach the expected bearing run times.

“G” : A labyrinth that influences the ventilation of the LPT disks and thus also the disk temperatures, which are directly related to the life span of the disk. It additionally maintains the pressure levels behind the MPT and therefore also the thrust loads required by the specific design.

Figure "Labyrinth seals influencing thrust bearings": The condition of the seals and, with it, the pressure distribution and thrust loads, depend on the operation time and the start-up/shut-down cycles. The top diagram shows the conditions in a high-pressure rotor. The position and length of the arrows corresponds to the size and direction of the axial forces resulting from the labyrinth-induced pressure differences (“piston forces”).
The left diagram describes the typical behavior of labyrinth seals on a high-pressure compressor in a multi-shaft engine (top diagram). The axial bearing force corresponds to that from the “piston forces”. Evidently, the bearing forces in worn seals and higher compressor exit pressures are about twice as high as those in new seals. The bearing force over the compressor exit pressure does not run linearly corresponding to PE and P1, but has a pronounced maximum. This behavior can be observed in connection with the labyrinth seals in the turbine area.
The pressure P1 behind the compressor exit labyrinth is considerably higher in worn seals than new ones. P1 (bottom right diagram) increases linearly with compressor pressure PE. However, the increase is steeper in worn seals than in new ones.

Figure "Bearing loads influenced by gas-seals" : The condition of the seals influences the air- and gas pressures in the different engine zones. The compressor builds up air pressure towards the back. Without the compressor end seal, it would experience not only the forward-acting forces corresponding to the exit cross-section and total pressure ratio, but also “piston force” in a forward direction on the compressor end disk. In the turbine, the pressure drops in the direction of flow, subjecting the turbine to a high axial force backwards. The designer of the air system makes the shafts compensate as much as possible for the forces between the compressor and turbine: the diameter of the labyrinth seals at the compressor exit and turbine intake are tuned to one another in such a way that an easily controllable axial bearing force acts in one direction throughout the operating time. It is undesirable for the bearing force to be too small or for it to change direction depending on the operating conditions. If the labyrinth play changes during operation, e.g. due to wear or erosion, the leak rates change along with the pressures and bearing forces.
Therefore, unusually heavy labyrinth wear can cause fatigue damage in the bearings.

Labyrinth vibrations (compare Ref. 7.2.1-15):

Labyrinths are susceptible to vibrations. Their usual filigreed design as thin-walled cylinders or cones makes them tend to vibrate. The opposite table is an overview of causes of labyrinth vibration. Vibrations are promoted by operating conditions such as: circumferential speed; RPM; axial clearance flow and circumferential flow in the chambers; pressure gradients in front, behind, and inside the seal; and vibrations of the suspension of the rotating and/or static parts. High mean stress caused by thermal stress from temperature gradients promotes dynamic cracks and fractures.
The vibrations of thin-walled cylindrical or conical shells usually occur with n nodal circles and m nodal diameters (see Fig. "Vibration modes of thin-walled shells").
The nodal diameters in a resting cylindrical shell are fixed, but, relative to an observer rotating along with them, the nodal diameters of a rotating cylindrical shell run around either in direction of rotor rotation (referred to as a forward-running wave) or against the direction of rotor rotation (referred to as a backwards-running wave).
Labyrinth vibrations affect labyrinth clearance and, therefore, sealing. If a vibrating labyrinth rubs, certain vibration movements leave typical rub marks, which allows conclusions as to the cause of rubbing (see Figs. "Acoustic incited resonance of labyrinth seals" and "Mechanical labyrinth resonances with rotor").

Figure "Vibration modes of thin-walled shells": Vibrations of thin-walled cylindrical and conical shells usually occur with “n” nodal diameters and “m” nodal circles:

  • Vibrations with n=0 nodal diameters are axial, radial, and torsional vibrations of the entire shell.
  • Vibrations with one nodal diameter are flexural modes of the entire cylindrical shell.
  • Vibrations with 2 or more nodal diameters are radial vibrations and ovalizations.

The nodal diameters in a resting cylindrical shell are fixed, but, relative to an observer rotating along with them, the nodal diameters of a rotating cylindrical shell run around either in direction of rotor rotation (referred to as a forward-running wave) or against the direction of rotor rotation (referred to as a backwards-running wave).

Figure "Labyrinth seal effectiveness" (Ref. 7.2.1-2): Assuming that the flow is turbulent, the following relationships apply to the labyrinth geometry and the seal effectiveness:

Clearance width: leakage losses are approximately the clearance width to the power of 3/2.

Chamber width (splitting): Even a relatively weak structuring of the labyrinth increases the flow-resistance compared with a smooth surface. This is also true for large splits. There is a weakly pronounced optimum at about 2 mm chamber width in the region between about 1 and 3 mm. Only with large splits does the seal effectiveness approach that of a smooth gap. Compromises must be found if pressure is not steady. Pressure ratio and chamber width do not affect one another much, so there is considerable freedom for design.

Gap length: In general, the larger the (axial) gap length, the greater the gap's resistance (even at a constant chamber number), therefore the splits increase.

Chamber depth: The flow resistance of weakly profiled labyrinth gaps is strongly dependent on the chamber depth even in very flat chambers. This behavior is most likely important for flat profiled coatings, as well (e.g. ceramic coatings with ground-in grooves). The optimum is around the ratio chamber depth/width = 1/4.

Fin angle: Generally, fins should be angled against the direction of flow. This increases the flow resistance in the seal. An angle into the direction of the flow still provides a better seal than a flat wall. It is important that the angle is between 30°-60°. A pronounced optimum can be observed in this range. The flow resistance in angled labyrinths also increases as gap width shortens (left diagram).
With small gap widths, there is no usable advantage over labyrinth fins with perpendicular “weather sides” (pressure side). In this case, the relatively high expense of angled labyrinth fins is not justified.
Angled labyrinth fins have larger contact surfaces during rubbing (Fig. "Labyrinth seals properties") and therefore create more friction heat and heating-up of the labyrinth.

Chamber shape: Angled “lee sides” create several times the flow resistance of angled “weather sides”.
Especially strong flow resistance can be created by structuring the opposite surface (abradable surface). This explains the considerable influence of the abradable/rub-tolerant coating surface. For example, porosity, roughness, rub-in grooves, alterations due to erosion, material removal, fatigue, etc. are important factors for the seal effectiveness of labyrinths. Partner surfaces, such as honeycombs, have been in use for a long time (Figs. "Honeycomb seals advantages" and "Honeycomb seals damping effect"). The seal is worsened by destruction of the open cell structure. This is the result of depositing during rubbing (e.g. lid formation, Fig. "Honeycomb seals damping effect"), breakouts due to oxidation, or wear due to rubbing and erosion. Filled honeycombs are more resistant to erosion and oxidation, but do not seal as well.

Fin shape: The narrower the fins and the sharper their edges, the better the labyrinth usually seals. Of course, along with the expense of manufacture (especially with armoring), the reaction to operating factors (e.g. thermal sensitivity, erosion, oxidation) must be taken into account. This limits the freedom of design considerably.
Results from 2D- and 3D-seal testing rigs indicate the following trends for the influence of the partner surfaces:

  • Porous rub-coatings have higher leakage than smooth, solid partner surfaces.
  • Increasing the gap size does not increase leakage to the same degree. If the gap size is small, the leakage in porous abradable coatings is greater than that in smooth, solid partner surfaces. With large gap sizes, this behavior is reversed.
  • Ni/C coatings have less leakage than smooth, solid partner surfaces if the gap height is between 0.13 and 0.25 mm.

The right diagram (Ref. 7.2.1-5) shows the connection between the gap height and the abradable surface, depending on the pressure ratio in the labyrinth. Relatively small gap heights show considerably more leakage with porous coatings (as are typically used as thermal abradables, e.g. NiC) than smooth, solid ones. Astoundingly, in this labyrinth configuration, honeycomb seals perform even better than smooth surfaces (influence of cell size, see Fig. "Honeycomb seals damping effect").

Figure "Honeycomb seals advantages" (Ref. 7.2.1-5): Seals with a honeycomb structure made from oxidation resistant Ni-based sheeting are often used as rub-tolerant coatings for labyrinth fins in the hot part areas, especially as a rubbing surface for turbine rotor blade shrouds and rotor spacer ring labyrinth fin shrouds. Even several millimeters of radial and axial rub-in do not damage labyrinth fins seriously (Ref. 7.2.1-1). This is due partly to the relatively low heat development during rubbing. They seal well even with larger labyrinth gaps. The relative leakage air flow in the diagram shows the influence of the structure of the rubbing surface on the leakage air flow when compared with a smooth wall.
The following results from a 2D- and 3D-testing rig (Refs. 7.2.1-5 and 7.2.1-10) are typical behavior for honeycomb seals used as partner surfaces in labyrinths with four fins (diagram):

  • Honeycombs decrease leakage considerably compared with smooth surfaces at gap widths typical during operation.
  • The cell size of the honeycomb has a strong influence on the leakage rate.
  • Honeycombs are especially effective for large gap heights. However, they may increase leakage rates at small gap heights.
  • With smooth surfaces and abradable coatings, increasing circumferential speed of the labyrinth fins decreases leakage. This effect is not pronounced in honeycomb seals.

The slightly porous Al-polyester and nickel-graphite abradable coatings commonly used in modern compressors have lower air leakage rates than honeycomb seals at gap widths up to about 0.2 mm. From 0.2 mm on, however, the honeycomb seal performs considerably better.
At all gap sizes over 0.1 mm, the more porous abradable coatings used for comparison (but not named or described in detail; metal felt?) had considerably higher leakage rates than honeycomb seals with cell sizes under 1.6 mm.
Honeycomb structures with very small cells (0.8 mm) have extremely low leakage rates at all gap widths over 0.1 mm. These seals require very thin walls, which should make them very oxidation sensitive.
Very long operating times, as are typical for modern engines in cargo aircraft and industrial use, and/or corrosive conditions can damage honeycomb walls through oxidation or hot gas corrosion. In extreme cases, the entire honeycomb structure embrittles and breaks out, decreasing the seal effectiveness accordingly. This results not only high leakage losses and drops in efficiency, but can also cause impermissible heating-up of the rotor spacer rings due to the hot gas leakage. For this reason, the manufacturer should be able to demonstrate the long-term durability of the seal systems in the hot gas region under operator-specific conditions.
In the high-pressure turbine area, honeycombs filled with Ni-based sintered materials are used opposite shroudless rotor blades. This is intended to minimize the danger of honeycomb walls breaking out due to oxidation.

Figure "Honeycomb seals damping effect" (Ref. 7.2.1-9): A pressure disturbance, running around the circumference in the direction of rotation and moving at the speed of sound in relation to the flowing medium in the labyrinth chamber (ring space between two neighboring labyrinth fins), causes dangerous self-increasing acoustic resonance if the whole-numbered nodal diameters of the air vibrations equal those of the labyrinth components and both waves move in the same direction. With smooth chamber walls, the wave speed is equal to the speed of sound in the air flow. The phase speed of waves moving in the direction of rotation is increased by the circumferential speed of the air flow. The phase speed of pressure waves that run against the direction of flow is decreased.
If the wall opposite the fins consists of periodical hollow spaces (as in honeycombs; bottom diagrams) that are closed at the end, the advancing speed of the pressure wave (phase speed Vp) is lowered (diagram).
The larger the cell depth S, and the shorter the representative cell length (circumferentially), the more the pressure wave is slowed. However, this deceleration occurs underproportionally. A smooth opposite wall (S/L=0) has no decelerating effect on the pressure wave. The diagram depicts the ratio a/L of 0.95, i.e. the wall thickness of the honeycomb structure is about 5% of the cell length L. With a typical cell length of 2 mm, the wall thickness is about 0.1 mm. Of course, a seal structure that has been heavily damaged by oxidation does not provide the beneficial damping effects of open honeycombs.
Oxidation of these thin walls will limit their long-term use considerably.

Figure "Honeycomb seals drawbacks" (Refs. 7.2.1-5 and 7.2.1-10): As the diagram shows, the given labyrinth configuration (large honeycombs with small gaps) has improperly high leakage rates. With larger gaps, however, the leakage rates drop below those of smooth surfaces or small honeycombs.

The influence of honeycombs on the leakage rate can be explained with the following diagram:

With small gaps or direct contact between the labyrinth fin and honeycomb structure, the airflow can use the large honeycombs to “avoid” the fin (right diagram), because the fin covers considerably less of the cell than it would with a small honeycomb.
Larger gap heights evidently do not disturb sealing aerodynamic effects in the honeycomb as much (vortex formation, air vibrations similar to those in a casing treatment).

Figure "Labyrinth seal clearance changes after shut down": There are many labyrinth seal configurations used in engine construction.
Selection of a labyrinth seal type should be based on proven implementations in comparable engines with comparable operating conditions.
As the table at left shows, each configuration has specific strengths and weaknesses. Manufacture, costs, operating behavior, assembly, and repairs are important criteria for seal type selection.

Manufacture: The filigreed geometry, exact tolerances, and high-strength materials of labyrinth seals require special manufacturing methods. This is intended to avoid problems such as distortion, induction of dangerous (tension-) residual stresses, and crack initiation (dynamic cracks during the manufacturing process, grinding cracks, comma cracks) in the materials, which are usually difficult to machine or grind.

A special problem is secure bonding of the rub coatings. This includes flawed solders and adhesives or poorly bonded thermal spray coatings.

The process of armoring the fins must be sufficiently reproducible and safely controllable. Complicated labyrinths with several fins can impede the thermal spraying process used to apply armoring through poor accessibility, ricochets, spray dust deposits, or poor temperature control. The hard ceramic materials used as armoring (e.g. Al2O3 or ZrO2 ) must be tuned to the rub coating (e.g. heat development, depositing, wear). This requires intensive optimizing and testing of the parameters of the thermal spraying process (e.g. pre-treatment of the bond surface, pre-warming, cycles). If the armoring must be ground over, corresponding loads (cutting forces, heat creation) will stress the labyrinth fins.

Costs: The material- and manufacturing costs of integral labyrinth components can be a significant portion of those of rotor disks or blades. The material selected for the rotating labyrinth carrier must be able to handle powerful operating loads (centrifugal force at high temperatures), as well as several repairs, if necessary. This can necessitate the selection of expensive materials. Multi-step application processes for armorings and/or abradable coatings require long cycle times and expensive manufacturing facilities with intensive and elaborate quality controls.

Operating behavior: Labyrinths must safely withstand a number of operating influences. These include:

  • Rubbing
    • local heating-up/thermal fatigue
    • Hot-crack initiation
    • Vibration incitement
    • Material removal/wear
    • Distortion, imbalances
  • Oxidation
  • Corrosion
  • Erosion
  • Vibrations (e.g. due to flow effects)

Mounting: Labyrinths should be as easily mountable as possible. This includes both detachable joining of the new parts to supporting components, as well as possible replacement as a part of the repair process.
The labyrinth gap should be sufficiently measurable. The labyrinth fins and/or the rub coatings of the labyrinth components must not be overloaded by forces created during the joining of modules or engine assembly.
It is desirable that the labyrinths are not separated during module replacement, but rather remain complete on one of the modules.
If the labyrinth must be separated during module replacement, it must be ensured that the new labyrinth components provide a sufficiently large labyrinth gap (larger than during new part assembly). This is especially true if the remaining abradable coating is assumed to have lost some of its abradability due to a long period of operation (oxidation, thermally induced structure changes).

Repairs: Due to their function, labyrinths are subjected to wear and thermal damages, which may necessitate multiple repairs. This must be taken into account when selecting materials, designing the geometry, and deciding on possible welding- and coating techniques. For successful repair, the seal must be replaceable, and parts with integrated labyrinth rings (e.g. a turbine disk in a small gas turbine) must be designed for reparability.

Figure "Rub coating types for labyrinth seals" (Ref. 7.2.1-6): The tribo-system labyrinth fin/rub coating requires a suitable selection of the rubbing partners, depending on the operating conditions. Despite specific weaknesses, the following rub coatings have proven themselves in actual operation.
Gramophone grooves: This structure is sensitive to mechanical influences (e.g. foreign objects, mounting). The thin ridges can be made to vibrate and break out (Ref. 7.2.1-11). It must be ensured that there is a sufficient vertical curve radius at the base of the grooves.
Unfilled honeycombs are widely used in the warm engine areas (high-pressure compressor, turbine). Long-term oxidation of the thin wall structures poses a problem (see Chapter 7.1.3).
Filled honeycombs are more resistant to erosion and oxidation than unfilled honeycombs, but their abradable properties are limited.
Soldered coatings made from silver, copper, and bronze are found mostly in older engine types. Problems can arise if the removed material damages the hot parts (diffusion, embrittlement, corrosion of nickel alloys). It is especially important to match the thermal strains of the solder and supporting part (Fig. "Dangers by delamination of rubcoating").
In the cold and middle temperature area (compressor), bonded or soldered metal felts are used. With these, sufficient bond strength is a problem, because the porous structure sucks in the bonding medium (Fig. "Shaft vibrations by labyrinth pressures"). This can cause the infiltrated metal felt to lose its excellent abradable properties.
Thermal-sprayed Ni/graphite coatings are widely used in modern engines. The problem with these is their limited erosion resistance and aging due to oxidation at temperatures above 350°C.
Filled silicon gum can be poured in or affixed with adhesive. However, it has been observed, that even small deviations from proven application parameters cause impermissible bond strength problems, which are very difficult to solve non-destructively.
Filled synthetic resin coatings include polyester with Al powder. These coatings have a specific damage mechanism. Shrinking leads to gap formation and coating corrosion along the deposited Al particles with blistering and serious coating delamination.
Porous spray coatings can be infiltrated by sodium silicate and become more resistant to erosion. The infiltrated coating is considerably less abradable, which can lead to catastrophic labyrinth failures.

Figure "Operating properties by labyrinth fin geometry": The geometry of the fin cross-sections can be used to optimize and control the operating behavior of a labyrinth. The assessments in the bottom table were made by the author on basis of experience and the following reasoning, and do not claim to be universally applicable:

Heat development: Assessment criteria are the size of the potential rubbing surface and the cuttability of a sufficiently rough armoring (and resulting low friction heat).

Heat accumulation: The difference between the amounts of heat accepted and dissipated by the fin were assessed. This is most likely worse in small fin ridges of a stepped cross-section than in continually widening cross-sections. Aside from its cuttability, the armoring provides a thermal insulating effect, which should considerably reduce the amount of accepted heat.

Crack initiation: Crack initiation is caused by high rubbing temperatures and high residual stresses due to limited heat strain in the heated area. Important influences include heat development and heat dissipation, which should not cause dangerous heat accumulation.

Crack growth: The evaluated slow crack growth originated from hot cracks or damaged rubbing zones, as long as it was limited to the fin cross-section. The primary factor promoting crack growth is the start-up/shut down cycle (LCF).

Fail safe behavior: This concerns the tendency of a growing crack from the fin to spread into the labyrinth ring.
This type of crack growth can be caused by LCF as well as high-frequency oscillations. The crack growth can threaten the integrity of the entire labyrinth and lead to dangerous consequential damages after the part fails. The notch effect of a stage in the labyrinth cross-section should make the crack branch off in direction of the circumference. However, this crack behavior can not always be guaranteed.

Wear: The volume of removed material indicates the wear of the fin and therefore the clearance increases, i.e. the decrease in seal effectiveness. In this case, the size of the fin`s contact surface and the cuttability are important factors. This criterion does not apply if the material removed is from the fin.

Material removal: Large amounts of removed material can block cooling air ducts or deposit on the fin and cause a self-increasing rubbing process. Spalled armor must not enter into the bearing, where it might cause track fatigue and erosion of gas ducts.

Friction force: The forces created during rubbing should be as small as possible, so that they do not overload the labyrinth structure or initiate dangerous mechanical vibrations.

Reparability: Complex, thin-walled fin geometry with exact tolerances requires elaborate repair procedures (deposit welding, machining, coating). High-temperature resistant materials (e.g. powder metallurgic and cast materials) are difficult to weld and tend to thermal crack initiation.

Bulging: Outward bulging of the rotating inner labyrinth ring with local heating-up and heat strain due to rubbing increases the rubbing process. This presents the danger of a self-increasing rubbing process with serious imbalances.

Figure "Labyrinth seals properties": Heating-up of the labyrinth is an important factor for safety and operating behavior. It is caused primarily, though not exclusively, by rubbing. The friction heat causes local extreme temperatures in the components. These temperatures depend on the amount of heat created, heat transfer, heat capacity, and heat diffusion. Additionally, air friction in the labyrinth may have an effect. The smaller the leakage flow, the higher the temperatures in the labyrinth. Therefore, a minimum leakage amount is vital.
The amount of friction created is partially related to the size of the contact surfaces. Axial rotor deflection causes one-sided heating-up of the labyrinth fin along the entire circumference (top left diagram). Interruption of the leakage air flow and the large contact surface cause especially intense heating-up of the labyrinth fin. Radial deflection causes the labyrinth fin to form a seal with the contact surface in one sector, but air continues to leak outside of this sector.
The top right diagrams show the differences between an angled and straight labyrinth tooth. The angled labyrinth tooth causes larger heat development with wider fins (necessary for withstanding mechanical loads). Narrow angled teeth have poorer heat dissipation than straight teeth due to their different cross-sections. This makes angled labyrinth teeth more susceptible to overheating.
The different “cutting angle” (infeed angle of the axial contact surface) of angled teeth causes rubbing force to be directed into the coating. This can increase the rubbing process with soft rotors.
The bottom diagram shows two of the most frequently used labyrinth fin cross-sections. The triangular cross-section on the left has better heat dissipation than the step-shaped one on the right. Therefore, triangular cross-sections are more likely to prevent overheating with crack initiation. However, it transfers more heat into the supporting ring, which increases the danger of local deformations and uncontrollably accelerating rubbing. If thermal fatigue does cause crack initiation in the fin, however, the crack is more likely to spread into the labyrinth ring from the triangular cross-section than from the stepped one.
The narrow, right-angle stepped fin cross-section is more easily overheated due to its poorer heat dissipation. However, the lower heat acceptance and stiffening effect of the relatively massive ridge cross-section make it less likely to cause deformations in the supporting ring. Reports indicate that the crack-stopping effect does not seem to be guaranteed.
The minimum height of a narrow fin cross-section “h1” must be greater than the radial length of expected hot cracks. If the hot cracks run across the step into the thicker cross-section, effective deflection of further crack growth is no longer ensured.
The height “h2” is important not only for heat dissipation, but also for radiation of the friction heat during rubbing, and should therefore be sufficiently dimensioned.

Figure "Compressor spacer labyrinth seals": Typical intermediate stage labyrinths in compressors. In the left diagram, the labyrinth fins are integrated directly into the rotor drum or spacer ring. This construction requires considerable experience, in order to avoid impermissible damage to the rotor drum or spacer from occurring during rubbing (see Fig. "Labyrinth fin crack by rubbing"). The coating is on a ring which has been riveted onto the stator assembly. This makes replacement during repair easier.
In the middle diagram, the labyrinth ring is connected with the drum or spacer ring by means of a ring bracket. This reduces the possibility of dangerous rotor damage due to rubbing. In this case, the coating is directly on the inner stator vane shroud.
In the right diagram, the labyrinth fins are carried by rings, which are positive-fit onto the rotor. This relatively elaborate construction is prone to problems with friction wear and imbalances in case of rubbing.

Figure "Turbine spacer labyrinth seals": This diagram shows the various design possibilities for labyrinths in turbines. One must be aware of large differences in thermal strain between static and rotating labyrinth components. The intensely cooled high-pressure turbines require elaborate and effective seal systems (“A”) to lead the cooling air into the blading. In two-stage high-pressure turbines, the intermediate stage seal (“B”), which must safely prevent hot gas incursions, is a special challenge. If this seal fails, the disk annulus can overheat. In the depicted case, the axial initial tension of the bow-shaped spacer ring is maintained by an internally braced ring. The spacer ring splays open under centrifugal force. The spacer rings of high pressure turbines, which are subject to high thermal and mechanical loads, do not allow much play for rubbing damage. In this case, low-cycle fatigue (LCF) cracks (start-up/shut-down cycles) can spread from the damaged zone.
The bottom three diagrams, “C,D,E”, depict labyrinth seals for spacer seals in low-pressure turbines. The relatively low RPM of this group of components permit filigreed structuring of the seal carriers. Experience has shown that these thin-walled rings are easily made to vibrate at high frequency.

Figure "Bearing chambers labyrinth seals" (Ref. 7.2.1-12): The top diagram schematically depicts the configuration of the bearing chamber of a typical main bearing (GE Co.). This type of bearing chamber has a multitude of seals. The most common are labyrinth seals or combinations of labyrinth seals and mechanical seals (bottom diagram).

The following requirements must be met with the aid of these seals:

  • Sufficient seal effectiveness over the entire operating time in order to ensure the necessary pressure losses. The bearing chamber, its surrounding environment, and the barrier air have different pressure levels, which must not change so much that it compromises the functioning of the bearing.
  • The oil/air mixture, which is under slight overpressure, must kept in the bearing chamber without unallowable leakage losses.
  • No damaging oil coke formation in the seal.
  • In order to prevent bearing damage, no foreign particles must be created in the seal and enter the bearing. These potential foreign particles include spalled armor (Example "Aluminum oxide particle contamination", Ref. 7.2.2-3) or material removed from labyrinth fins and rub coatings. For this reason, thermal-sprayed Ni/graphite coatings are used as rub coatings, since they are very soft and do not necessitate fin armoring (Ref. 7.1.2-13). If the pressure in the bearing chamber is greater than that in the first seal, then foreign bodies will not enter the bearing chamber.
  • Axial and radial motions of the shaft relative to the bearing chamber or the attached housing must occur without causing damage.
  • If the seals rub, it must not cause sparks or temperatures high enough to cause an oil fire in the bearing chamber. This is another reason for soft Ni/graphite rub coatings in bearing chamber seals.

Figure "Turbine blades shroud labyrinths": Turbine blade shrouds, especially in the low-pressure turbine, are outfitted with labyrinth fins and thus form a segmented rotating labyrinth ring. The opposite surface is the housing or seal segments set into the housing.
The shroud of seals “A” and “B” moves into a ring gap in the housing. The gap is large enough to ensure that rubbing does not occur. With “C” and “D”, the housing side has a rub coating (e.g. an abradable honeycomb structure or a heat resistant, abrasive thermal spray coating).
Even the blade root platforms can form a rotating seal ring to prevent hot gas from reaching the disk annulus. Each blade has a bracket (arrow) that forms a labyrinth ring. The opposing surface consists of the inner shroud of the preceding turbine stator vane.

Figure "Labyrinth fins design influenced heat up": Tribologic operating properties of a rub coating are determined by the rub coating structure, bond zone/bond layer, and base material. These properties determine whether the coating is rub-tolerant (material removal is primarily from the labyrinth fins; bottom left diagram) or abradable (material removal is primarily from the coating; bottom right diagram). The operating properties can be classified into the following groups:
Physical material properties: thermal expansion, thermal conductivity, reflection, permeability, density.
Chemical properties: composition, homogeneity, oxidation, corrosion.
Technical properties: thickness, layering, roughness, topography.
Structure: porosity, layering.
Strength: inner strength, bond strength, toughness, hardness.


7.2.1-1 P. König, A. Rossmann, “Ratgeber für Gasturbinen-Betreiber”, Vulkan Publishers Essen, 1999, ISBN 3-8027-2545-X.

7.2.1-2 K. Trutnovsky, “Berührungsfreie Dichtungen” (Fundamentals and Uses of the Flow through the use of Cracks and Labyrinths), VDI-Publishers GmbH, Düsseldorf.

7.2.1-3 D.V. Wright, “Labyrinth Seal Forces on a Whirling Rotor”, NASA-CR-168016, January 1983.

7.2.1-4 C.O. Nicks, D.W. Childs, “A Comparison of Experimental and Theoretical Results for Leakage, Pressure Distribution, and Rotordynamic Coefficients for Annular Gas Seals”, NASA-CR-174000, September 1984.

7.2.1-5 H.L. Stocker, “Determining and Improving Labyrinth Seal Performance in Current and Advanced High Performance Gas Turbines”, AGARD-CP-237, pages 13-1 to 13-20.

7.2.1-6 L.P. Ludwig, R.C. Bill, “Gas Path Sealing in Turbine Engines”, ASLE Transactions 23 (1980) 1 pages 1 - 22.

7.2.1-7 W.F. MCGreehan, S.H. Ko, “Power Dissipation in Smooth and Honeycomb Labyrinth Seals”, ASME Paper 89-GT-220 of the “Gas Turbine and Aeroengine Congress and Exposition”, June 4-8, 1989, Toronto, Ontario, pages 1-11.

7.2.1-8 L.P. Ludwig, “Gas Path Sealing in Turbine Engines”, AGARD-CP-237, pages 1.1-1.38.

7.2.1-9 J.S. Alford, “Labyrinth Seal Designs Have Benefitted from Development and Service Experience”; SAE-Paper 710435 (1971).

7.2.1-10 R.C.Bill, L.T.Shiembob, “Some Considerations of the Performance of Two Honeycomb Gas-Path Seal Material Systems”. Zeitschrift “Lubrication Engineering”, April 1981, pages 209-217

7.2.1-11 J.S. Alford, G.W. Lawson, “Dimensional Stability and Structural Integrity of Labyrinth Seals”, SAE-Paper 660048 of the Automotive Engineering Congress, Detroit, Mich. January 10-14, 1966, page 5.

7.2.1-12 I.E.Traeger, “Aircraft Gas Turbine Engine Technology” second Edition,Glencoe, Macmillan/McGraw-HillI, ISBN 0-07-065158-2, 1994, page 276.

7.2.1-13 C.L.Broman, “Energy Efficient Engine, Core Engine Bearings, Drives and Configuration Detailed Design Report”, NASA-CR-165376, page 16.

7.2.1-14 H.J. Macke, “Traveling Wave Vibration of Gas-Turbine engine Shells”, Journal of Engineering for Power, April 1966, pages 179 - 1897.

© 2021 ITTM & Axel Rossmann
7/72/721/721.txt · Last modified: 2020/06/25 22:43 (external edit)

Page Tools