Why Mechanical Seals Fail

Mechanical face seals should last until the carbon face wears away. If the seal starts leaking before that happens to the point where the pump must be shut down and the seal replaced, then the seal has failed.

Examine the faces 

Chipped edges on either or both faces 

Chipping is caused by a large separation of the faces and consequent breaking when they slam back into each other. It is most often associated with FLASHING. It is most common in hot water systems or in fluids that may have water condense in them. Water when it changes from a liquid to a gas expands thousands of times in volume and can cause a large face separation. 

Severe cavitation of the pump coupled with a hung up seal may also cause the problem. Usually small vibrations, misalignments and the like cannot cause the breakage, because they do not separate the faces enough. 

The cure in the case of flashing is to reduce the face heat. This is done by using carbon versus tungsten carbide or other cool running face combinations, by using pressure balanced seals, by cooling the stuffing box fluid area, by ensuring that the seal spring tension was not excessive due to installing it wrong or by using a double seal or an outside quenching fluid to keep the faces running in a cooling fluid so that it cannot occur. A cooling flush to the stationary ring by the use of a special gland can also be used for this problem.

Flaking or peeling of the hard facing

Hard facing of stellite, ceramic, and a variety of other materials are often used in seal designs with a rotary hard face. Flaking or peeling is generally a sign of either a defective coating or a chemical attack at the bond. The attack was probably caused by the intense heat that is often found at the face of a seal. It should be noted that when materials are plated you usually retain the chemical properties of the substrate due to the fact that most facings have some degree of porosity. This type of problem is solved by using solid face materials.

Pitting, blistering, corrosion of the carbon face

The carbon used in mechanical seals is selected for the particular application and should not be subject to these problems. This occurs when the wrong carbon is being used or where carbon faces are machined locally. Most seal carbons use an impregnated face and this is not obtainable when a carbon is machined from tube stock. Hot oil service carbon has been formulated especially to prevent blistering and pitting and this is easily cured in oil by using these carbons. Corrosive attack of carbon can be stopped by selecting carbons which are relatively binder free. In the few fluids which attack a pure carbon or carbon graphite such as nitric acid, oleum, chlorosulphonic acid and some exotic highly oxidizing acids, the alternative to use is a TFE or filled TFE face. Faces made from PTFE, TFE are a poor substitute for carbon but are appropriate for the few fluids where a pure carbon will not withstand the fluid.

Check the springs

Spring(s) or bellows breakage (metal)

 

Springs and bellows break usually because of chemical attack at the same time the device is being stressed. The phenomena of stress corrosion cracking is explained by many different theoretical methods. It is commonly seen in seals when stainless steel springs and bellows are used in certain fluids. When the fluid being sealed contains chlorine, bromine, iodine, fluorine and irons or compounds of these elements, they often will attack the chrome oxide layer that protects most grades of stainless steel. While the oxide layer is being attacked. the flexing will open up small cracks. If the oxide particles wedge into these cracks, a sudden failure can occur. For this reason, spring materials and bellows plate materials should be chosen from alloys, such as Hastelloy, Carpenter 20, Monel and the like, in the presence of the elements listed. One common fluid which also causes spring breakage is caustic soda. Stainless steel springs and bellows should be avoided in the presence of caustic.

Spring breakage and bellows breakage accompanies flexing of the device, but repeated axial compression of a bellows or spring will not cause fatigue failure. This happens when a portion of the spring or bellows is extended too much or flexed in torsion.

Clogged springs

Springs usually clog when the product is dirty and the seal is not moving axially. Some multi-spring designs clog very easily and should not run in a dirty fluid without some type of clean flushing fluid. On bellows designs and single spring type seals, clogging usually can only take place when the pump is not rotating. If this symptom is seen, it is not very important. It is only if the fluids lock up the two faces or cause the body to stick to the stuffing box that a problem is likely to develop.

Bellows clogged at the inside

A metal bellows seal will only clog up and fail if the fluid hardens or particles become stuck at the inside of the bellows. This occurs when there is excessive leakage past the face or past the static shaft seal (usually an "O" ring). The normal leakage from a seal, installed and operating properly, will not cause clogging for years. The most important thing to investigate with this problem is:

1. Is the seal clogged at the proper operating length? If installed with no compression, or too little compression, the seal will start leaking and soon clog. This can be easily determined by measuring its length.
2. Is the wear track significant? If there are signs of excessive motion, this could be the cause of the leakage. If none, then the leakage came by the shaft seal.
3. Shaft seal damage. This can be caused by installation or by shaft deflection, causing a metal-to-metal contact in the region of the pump throat. This will leave a telltale ring around the shaft. This excessive heat will melt a TFE seal long enough to let it leak, but it may heal when the heat source is removed. If this is the problem, then the throat must be machined open, so the shaft deflects substantially.

 Deep wear in the hard face

This often accompanies outside seals, seals in misaligned pumps and seals in severe abrasive service. It is caused by face separation letting large particles between the faces. These particles then embed in the carbon face and grind the hard face. This can occur in crystallizing products also, where high face heat causes some products to change to abrasive crystals. The problem is often compounded by reuse of the carbon face because it shows little wear.

When lapping compounds are used to lap carbon, the same problems can occur. The compounds embed and then grind the hard face.

The problem is solved in the ways recommended for sealing abrasive products and products which crystallize. Briefly: keep the product at the O.D. of the faces so centrifugal action helps exclude the particles, reduce misalignment and vibration to prevent face separation. Use a seal with low shaft drag, such as the metal bellows or rubber bellows design, which have none. If possible, try to flush the fluid away from the seal with a clean external flush.

For abrasive service, a very hard stationary face, such as Tungsten Carbide or Ceramic, can retard the problem and a hard face combination such as Tungsten Carbide against Tungsten Carbide, can drastically reduce abrasive face wear. The hard face combination is most effective because it eliminates the grinding mechanism. The particles cannot embed in either face, so usually get ground up and pass through the faces and leak out. 

Worn drive lugs or worn drive slots

This is caused by "slip stick". If the two faces stick together, the pin drive will load up with a high stress. This is then transferred back to the face, causing it to accelerate and then stick again. Instead of a smooth rotary motion, the face is being beaten around in its circular path. Slip stick is caused by a lack of face lubrication. This can be caused by a variety of problems. You must look at the other clues to determine the most likely. Lack of face lubrication can be caused by:

1. Installing the seal with too much compression on the springs.
2. Too much pressure acting on the face, i.e. using an unbalanced seal where a balanced seal should be used. 
3. The fluid being sealed has poor lubricating properties. 
4. The face combination is bad. Using faces for their chemical resistance without regard for their ability to run as a seal face. 
5. Pump cavitation. 
6. On vertical pumps, air trapped in the stuffing box. Also, on these pumps recirculation lines, from the pump discharge, rather than the suction, causing trapped gases.

This is a very important clue because it tells you about the nature of your product. Double seal arrangements are necessary when a product is not a good lubricant. This clue will tell you about your product's lubricating properties.

Check the elastomer

Swollen, sticky, or disintegrating elastomer 

This is a sign of chemical incompatibility. It is solved by using a different material. Charts should be consulted or if none are available for the product, immersion tests can be run. If the product is a mixed solvent and no elastomer is suitable, then a TFE sealing device should be selected. This can be in a V-ring seal, a wedge seal, a U-cup seal, an all PTFE seal, or in a metal bellows seal, with a static TFE "O" ring.

Hardening of the elastomer Charring of the elastomer, cracking, burned appearance Elastomer has changed shape, "O" rings square, etc..

These are all signs of excessive heat.  Usually the source of heat is the face or a metal-to-metal contact of two parts. Excessive face heat is caused by lack of lubrication and that is caused by the items listed under drive lug wear. Look for signs of metal-to-metal contact. This is very common, yet often overlooked, because the marks look like they may have been machined onto the seal originally.

In the case of heat transfer fluids, such as Downtherm, Humbletherm, etc., the hardening can be caused by heat transferred through the shaft or by the loss of the cooling system to the pump. When pumps use jacketed cooling, it is quite common to see it clog up and block the cooling.

Check for accidental rubbing

 

In a troubleshooting approach, it is important to carefully inspect the shaft, the seal, gland and stuffing box, if possible. Look for signs of rubbing. In high temperature pumps, rubbing of parts may take place only when the pump is hot. When cooled, the worn mark may get covered over and not be as noticeable.

Some easily overlooked causes for rubbing are:

• Flushing lines coming into a lantern connection and extending into the stuffing box. 
• Glands which do not pilot, slip down enough to hit the seal. 
• Gaskets slip into the seal cavity. 
• Stationary rings which do not plvot and come in contact with the rotating shaft. 
• Built-in restricter bushings in pumps that are supposed to be removed for high temperature, but are not. 
• Build-up of scale in the stuffing box. 
• Stuffing box not concentric with the shaft. 
• Excessive shaft deflection caused by throttling the discharge or otherwise operating the pump at its wrong capacity. 
• Set screws back out and hit the stuffing box.

Widened wear track

A widened wear track indicates that there is serious misalignment of the pump. This can be caused by bad bearings, shaft whip, shaft deflection, a bent shaft, or severe vibrations from a cavitating pump, bad coupling alignment, severe pipe strain, or a stationary seal ring which is tilted. The widened wear track is usually associated with leakage and seal hangup. If the seal is forced to move both radially and axially on each revolution, there is a tendency for the seal faces to separate slightly on each move. This leads to leakage which can gum up the sliding elastomer. This is especially true when the seal has a TFE sliding shaft seal. These are more likely to get hung up then resilient elastomers which can often flex.

The options to cure it include: alignment of items mentioned, reducing vibration through better couplings, reducing or eliminating pipe strain, operating the pump at the designed capacity, or by reducing the sliding friction on the shaft caused by the secondary seal. Though reducing the friction will not reduce the width of the wear track, it will extend the life of the seal. If the seal can follow the vibrations and motions with little drag, many of the problems caused by face separation can be eliminated.

Broken ceramic

Ceramics are sometimes subject to heat shock or cold shock. This most often occurs when the ceramic is heated unevenly and then subject to a rapid change in temperature. In many industries the pumps are cleaned by hosing them down. If a stream of water hits a ceramic that is running hot, it will cause it to fracture. Tests run by the ceramic companies indicate that breakage is a function of several things. The more pure and smaller grain size ceramics are less likely to break. Also breakage depends on the shape of the ceramic piece. The more corners and sharp-edges (called stress raisers) the more likely to break and if there is a temperature gradient across the face, that is, the face is hot but the back is cool, the more likely it is to break. What this all means is that a square block pure ceramic raised evenly to a high temperature and suddenly cooled will probably not break. A "T" shaped stationary ceramic running with a hot face which suddenly is cooled is most likely to break.

As a general rule the cure for breaking ceramics is a material change if the problem is from heat shock. In initially selecting faces ceramic is often avoided When the fluids are in excess of 300^o F(149^o C) because there is always the possibility of rapid cooling. Ceramic is often an economical hard face which has exceptional corrosion resistance and if selected for these reasons, a block type shape would be the best around.

The other cause of broken ceramic is from mechanical shock or tension. Ceramics are strong in compression but when put into tension by clamping them against an uneven surface or attempting to press them into a shell they often shatter. In seals that use rotating hard faces that are driven by pins ceramic should be avoided. The chance of fracture when the faces stick is very high. This is because the pins start hitting the ceramic.

Worn spot in the stationary ring

In some seals the stationary ring is carbon and there circulation from the pump discharge impinges on it. When this happens, it can cause erosion. Some seal companies direct the seal flush at the faces without regard for this problem. It usually will accompany face abrasive damage and other signs of face separation. The flush line should be directed not directly at the seal, but tangent to it. That is, the flush should come in at an angle causing the fluid in the stuffing box to circulate.

Heat 

Many problems associated with seal failure such as pump cavitation, dry running, loss of flush and accidental rubbing of metal to metal cause seal destruction because of the heat which may be generated.

Heat generated by the faces cause problems only for those materials which are heat sensitive. The most heat sensitive element in any seal is the elastomer. In a rubber bellows seal, for example, the elastomer is in close contact with the face and even a short dry run will cause immediate damage.

Other factors that contribute to excessive heat include the following. The face combination determines how fast and under how much pressure the seal can last. The best face combinations for chemical and refinery use are carbon against solid Tungsten Carbide (certain grades), carbon against ceramic and carbon against stellite. Some new silicon carbide face materials are exceptional in their ability to run dry without failure. The fluid being sealed of course is important. Viscous fluids at higher speeds and with very flat faces can cause excessive heat through shear of face film between the stationary and rotary face. Seal balance is an important determinant of face heat. Balanced seals usually run with a lubricating film while unbalanced seals can quite often become over pressurized squeezing out the film and thus increasing the friction dramatically. The amount of face load plays an important part in determining how the face will be lubricated. The same holds true for the face width. Wide faces have a problem establishing a film across the face.

Heat causes several problems that are not always obvious. When the seal contains a sliding elastomer in the face and the face is running hot, the elastomer will be hardened over a period of time . The hardening then reaches the point where leakage starts. Once leakage starts past the elastomer, there is a tendency to gum up or hang up the elastomer. This gumming up then stops the seal from following motions caused by misalignments at which time the seal starts increasing its leakage rate through face wear. When the leakage is too high to be tolerated, the seal is then changed. The underlying reason for the failure was the heat, but unless a close check is made of the elastomer, it will not appear to be the cause.

Inspect the seal drive    

Seal designs all use some way to transmit torque from the shaft to the rotary face. Quite often, it is done with pins, set screws and lugs. In a few cases it is done with the single spring. To check for this clue you must first determine for your particular seal where the drive junction is located. Seals are usually loose in torsion, that is, outside the pump you can twist them slightly before they engage. You are looking for signs of wear at the pin, drive lug, dent or spring. In bellows seals the signs are not present because they are usually a solid drive.

Hysterisis 

When a stationary ring is not square with the shaft, the sliding elastomer in the face of the seal must move back and forth on each revolution in an axial direction. The amount of motion depends directly on how much misalignment from a perfect 90 degree angle. Misalignment can also be caused by pipe strain, bad bearings, a bent shaft or shaft deflection caused by improper system operation. The seal is alternately pushed away from the stationary ring by the immovable face and back towards it by the spring pressure and by the fluid hydraulic pressure. The spring force must be high enough to overcome the resistance to motion caused by the drag of the elastomer.

Hysteresis is sometimes used to describe the amount of drag caused by the elastomer as measured in lbs., oz., etc. Hysteresis is also used to describe a delay or lag between two events. The rate of motion of the seal face axially must be the same in both directions or the seal faces will separate in not returning as fast as it was thrust away from the stationary ring. This minute separation caused by motion, drag and hysteresis depends then on not only the amount of drag, but the size of the seal and the speed. Hysteresis is the underlying reason for face separation, leakage, premature life, abrasive face damage and a variety of other ills in pumps that are not in perfect alignment.

Face separation  

The faces of a seal are normally flat to less than 20 millionths of an inch and are lubricated with a thin film of the sealed fluid. Because this leaves something less than a micron between them, they would normally act as a natural excluder of abrasive particles. When the faces are moved axially on each rotation, there is a tendency for them to separate much greater distances than millionths of an inch. .005" to .030" misalignment is not uncommon. The number of times that the seal has to move axially is over 10 million times a day on a 3600 rpm pump.

The separation of the stationary and rotating face by a few thousands of an inch causes two problems: It allows large abrasive particles to get between the faces and it allows the fluid being sealed to leak out. The leakage out can carry away wear particles causing rapid face wear and it will gum up or hang up the sliding elastomer from the outside where no self-cleaning takes place.

Proper size wear track  

This is an important sign because it tells you that the pump is in good alignment and face leakage is probably not the cause of any seal problem you might have. In a clogged metal bellows seal, for example, this is the clue that tells you the seal leaked by the static secondary seal.

Narrow wear track 

When the wear track is narrower than the thinnest face, this means that the seal has been over pressurized and has bowed away from the pressure. This bowing causes the seal to seal only on a portion of the face width. This is from improper design and the seal must be changed to a higher pressure, more rugged design if this occurs.

No wear track 

If there is no apparent wear on the faces of the seal after they have been in operation for some time and the seal is a rubber bellows type you should examine the springs and stuffing box. This means the faces may have been pressed together with the shaft rotating under the rubber. The springs will be worn and shiny if this has happened. This is because the spring remains stationary and rubs against some rotary part of the pump. This is caused by using the wrong lubricant on the rubber during installation and could be also due to an underside shaft and too good a shaft finish.

In several conventional seals we have seen this symptom where the seal had run against the gland rather than the pressed-in stationary face. This had been caused by the gland slipping in one case and in another case by the gland bore being smaller than the OD of the seal.

No wear track - shiny spots on the face  

This is caused by a warping of the face with the spots. Warping is caused by too much pressure, improper bolting or clamping or a bad face on the pump where the face is clamped. This can happen easily on two bolt glands that are not thick enough; it also can happen when the face is severely out of flat before it has been installed.

Cures for the problem include checking to see if the hard face is flat prior to installation, facing off the pump so that it is a clean smooth surface, using four bolt glands or glands that are strong enough to spread the bolt force evenly, and taking pains to draw up the bolts evenly. 

This is an important symptom because it indicates the seal probably was leaking from startup. The constant leakage usually causes the elastomer to hang up and the seal is no longer able to clean- itself. This can then lead to clogged springs which might have appeared to be the cause of the failure, but was really a result of the leakage.

Collect the entire seal 

Do not try to troubleshoot a seal by using only the parts that look important. You must have both the rotating part and the stationary part. If possible, you should also be able to inspect the gaskets, O-rings or other secondary seals, the shaft sleeve and the inside of the stuffing box.

It is a good idea to have someone troubleshoot all the seals that are removed whether there appears to be a problem or not. The best way to do this is to use a procedure that is very successful in several chemical plants. When a seal is removed the stationary and rotating parts are tied together and tagged with any information that may be useful. Then they are stored in the shop until they are ready to be rebuilt or discarded. In the meantime they are available for troubleshooting and failure analysis. Though the troubleshooting may be of little help for the pump that contained the seal, quite often this type of troubleshooting turns up common problems that can be corrected. When the two parts of the seal are separated after removal, it sometimes becomes difficult, if not impossible, to determine the actual cause of seal failure.

Guides for the installation of successful mechanical seals  

Preparation of the pump

(a) Be sure seal chamber is clean and free of all foreign matter.

(b) Shaft of shaft sleeve on which the seal is to operate must be to size, must be smooth, straight, and free of all burrs, sharp corners, nicks, or excessively deep scratches.

(c) Plug all holes in the stuffing box which are not to be used in the operation of the seal.

(d) Face of stuffing boxes must be smooth, clean, and square with the axis of the shaft.

(e) Halves of boxes on horizontal split case pumps must match perfectly, with the gasket between the halves extending flush with the surface on which the mechanical seal gland-gasket is to seal. Remove all sharp corners and burrs from stuffing box face.

(f) Check the shaft for alignment with a dial Indicator. The maximum allowable runout for optimum seal performance Is .005" TIR. Excessive misalignment may mean faulty bearings or bent shafts.

(g) Keep shaft end-play at a minimum Recommended maximum end-play is .005 inches.

(h) Check pump wear rings and impeller for proper clearances. Shaft must turn freely. Vibrations caused by rubbing and improper clearances can cause seal failure.

(i) Wherever shaft sleeves are used, make certain the sleeve is properly gasketed to the shaft to prevent leakage under the sleeve.

Installation of the Seal         

(a) Always handle mechanical seals with extreme care. Cleanliness is imperative. Never place faces face down on bench or floor. Keep seal in shipping containers until ready to install.

(b) Follow seal Installations Instructions carefully, (See Installation Instructions).

(c) Be sure all seal set screws are tight.

(d) Where set screws are used as a drive between seal and shaft, shaft should be counter-sunk to receive cup point.

(e) Care should be exercised In tightening gland bolts. Tighten evenly and do not spring gland.

(f) Use four equal-spaced gland bolts wherever possible; API-ASME Code for Unfired Pressure Vessels should be followed wherever possible when selecting gland bolt size and spacing.

(g) When tightening gland bolts, check clearances between shaft and gland with feeler gauges. This is particularly important when the gland is not piloted on the stuffing box as glands must be accurately centered.

(h) Test seals statically under pressure before starting pump. Make slight adjustments in gland nuts as necessary to stop any leakage which may occur through gland-gasket.

(i) Never operate mechanical seals dry. Carefully follow instructions for flushing and cooling connections where specified. Be sure suction and discharge of pump is open and a positive head of fluid is present before starting pump. This applies even to that period when checking for proper direction of rotation and adjustment of motor electrical connections.

Troubleshooting         

1. Seal spits and sputters in Operation

Product is flashing across the seal faces due to vaporization. Keep in mind a definite liquid condition between the faces is required and take steps to maintain this. Check to determine if pressure, perhaps, requires balanced design rather than unbalanced and, if the seal is already balanced, it may be that pressures are more severe than indicated on the specification sheet. Determine the correct actual stuffing box pressure and temperature and also the specific gravity and vapor pressure at these conditions for the product being handled as this data may provide the clue to the trouble.

2. Seal leaks and icing appears around the gland 

Product is flashing across the seal faces due to vaporization. If icing has occurred, undoubtedly some damage has been inflicted on the stationary seat and the carbon seal ring. These faces should be inspected and repaired, if possible, or replaced, if necessary, after the vaporizing condition has been corrected.

3. Seal drips steadily 

Check to see that the gland gasket is under proper compression against the face of the stuffing box. With horizontally split case pumps, be sure to check at interface of joint and gland. Faces may be deflected or not truly flat. Improper gland-bolting or overstressing of the bolts may have caused a deflection of the stationary seat under compression. This will occur, primarily, with clamped type seats. Shaft packing on the rotary unit or on the stationary seat may have been damaged in installation. The waring faces may have been scored by abrasives or other fine particles. If a unitary assembly or mounted on a pump sleeve, it is possible that leakage is coming under the sleeve itself.

4. Squealing seal  

This indicates dry operation which may be due to lack of liquid at the sealing faces. It is possible that a circulating flush line from discharge or an external source of fluid may be necessary. Furthermore, if one is already installed, it is possible that the orifice in it is too small and it may be necessary to enlarge the orifice.

5. Carbon rotating face dusting and this wear showing up outside the seal on gland and along the shaft 

Insufficient liquid at the sealing faces. Liquid is flashing due to vapor pressure built up between the seal faces leaving a fine crystallized particle residue or is creating dry contact thus grinding the carbon away. The stuffing box pressure is too high for the seal design and, undoubtedly, some correction has to be made. A balanced seal may be the answer.

6. Seal leaks and there is nothing apparently wrong

The faces may not be flat. This can best be determined by removing the faces and examining the wear pattern as discussed previously. The stationary seat may have been distorted due to excessive gland bolting stressing the clamped stationary seat and distorting same. This can also be determined from the wear pattern on examination. Improper piping to the suction and discharge gland of a pump can actually stress that pump, distorting the seal faces in the alignment with the shaft. If this problem is encountered, it is most common on vertical end suction overhung type impeller centrifugal pumps. Many of these pumps are not of sufficient strength in design, etc. to tolerate the excessive weight which results in misalignment due to same and this will affect the seal. Pipe hangers are the only solution to this problem. Possible shaft vibration can be caused by misalignment, impeller unbalance, cavitation, and bad bearings.

7. Short Seal Life

The greatest major cause of short seal life is excessive abrasives getting between the faces and causing rapid wear. The source of these abrasives may either come from slurry condition or they may come from the supercooling of a supersaturated solution or it may occur due to flashing across the seal faces, causing the dissolved solids to crystallize out between said faces and, again, causing wear. Cooling or heating, as the situation might occur, and/or most assuredly circulation of pumpage from discharge to the stuffing box or external clear flushing will alleviate these conditions. Misalignment of equipment. Pipe strain distortions as mentioned above. Seal shows signs of running too hot when a by-pass flush or recirculation may be necessary. Check for the possible rubbing of seal components along the shaft. Throttle bushings and poorly piloted glands can often cause this condition. Attempt more effective cooling of the seal area by connecting all cooling lines, checking to ascertain that all cross drilling of flush lines, etc., are clear and unobstructed (remove all scale, etc., that may accumulate in these lines), and by increasing the capacity of cooling lines or open the orifice clearances on circulation lines. Possible improper choice of type of seal may have been selected