Mechanical Seal Lifecycle 2

[Andy49]

Do mechanical seals fail gradually (more and more leaking of fluid as days or months go on) or do they fail abrupptly? Thanks in advance

 

[flexibox]

Andy 49
Mechanical seal leakage is dependant on process fluis they are sealing. Just for your information they do leak between 5 to 10 cc/hr. Seals can operate for many years trouble free and one day start to leak. Leakage could be in the form of drops per minute to a large leak. If they fail abruptly as you put it, something is wrong. It could be a process upset condition, fractured bellows (this is a seal type) bearing failure etc. Depending on the industry to you are in seal life can go from average of 10 years Oil and Gas industry to 3 or 4 months in the mining industry. Do you have a specific application you are trying to assess?

trust this helps

 

[Andy49]

Thanks. Your answer does help. What I am wondering is that if I collect leaking liquid from pump (if it is leaking at all) and measure it at regular time intervals, will I be able to predict the impending failure or maintenance needs for that seal? Assuming all other conditions remain the same (crude oil flow, pressures, etc)

 

[flexibox]

You may depending on the application. This is called API Plan 75 (see attached). If you have single seals however this may prove difficult. You really should have dual seals to capture what you want to achieve. what are your current seal typs/arrangements? see attached it may assist you

 

[turbofanatic]

whats your seal type and configuraton also application

 

[Andy49]

Turbofanatic, It will be for a horizontal centrifugal pump, single mechanical seal. Thanks for asking. Anything you can add would appreciated. I have come out of retirement after 10 years and trying to catch up/update/reconnect with the profession.

 

[Airsporter1st]

If you are careful to collect the leakage from a single mech seal on a regular basis you can observe the leakage trend and this will tell you much about the seal performance.

Generally a new seal should leak a little more and the leakage should trend downwards as the seal beds in - ideally to the point where there is no visible leakage. Thereafter if a seal starts to leak a little and this leakage trends upwards then you can be fairly sure of impending failure.

Bear in mind also that there are numerous other factors which can affect seal leakage on a day-to-day basis, so trending needs to be carried out over a long period to smooth out any short term variations.

As has already been said, many seals simply run and run and then suddenly fail, in which case trending is of little or no value. I have seen seals which have run ostensibly leak free for years, then one day just start leaking heavily. On examination I have found that the 'nose' of the carbon has worn away such that a larger face width now comes into play and seal balance changes. This is quite common on clean light hydrocarbons where the seal runs on a marginal film.

By the way, be sure that you collect only the seal leakage and not bearing oil leakage from the flinger, or rainfall! The best way to do this is to install a pipe into the seal drain port, lead it down say 150 mm, then turn through >90 deg (i.e. slightly upwards) for say 100mm then vertically downwards again into a collection bottle or can. This ensures that anything that comes down the outside of the pipe will drip off at the low point and not make it into the collection bottle or can.

 

[Andy49]

Airsporter1st. Thanks. It makes much more sense now. In the past couple of days, I have read quite a bit of information.