Causes for common gas seal failure

[reallyme00]

Hi,

Good day to everyone.

I appreciate if anyone can tell me the common causes for a Dry Gas seal failure in a Cryogenic Service.

A John Crane Type 28 VL is installed on a (newly commisioned )Ethylene pump.

Its a Vertical Turbine pump. Since the seal is a proven seal and it is running for almost 8 years without any problem on similar pumps in the same plant with same parameters.

We believe that the problem is not with the seal but with the operation. We have changed the seal twice and the Inspection report says that the symptoms are the seal face contact, which should not happen in dry gas seals.

So if anyone have previous experience in this kind of seal failure during commisioning , please kindly share with me.

Thanks a lot in advance.

 

[longeron]

Would you please go into more detail about the service, how well the pump runs, if there have been any process or operational changes, and describe the support system you are using for the seal?

How soon after start-up do you have to replace the seals?

 

[JJPellin]

I am not familiar with cryogenic service, but have a few ideas. Gas seals must have differential pressure of the buffer gas over process pressure. Do you have a tap on the seal chamber to be certain you know the process pressure? Most vertical turbine pump run with the seal at full discharge pressure. But we have a couple of them where the seal runs under suction pressure. We even have one that has an internal chamber with plugs so you can choose whether it sees suction pressure or discharge pressure.

Are you certain the seal is designed for the proper rotation? Once again, most of the vertical turbine pumps are designed for counter-clockwise rotation. But there may be some set up for clockwise.

Vertical turbine pumps, especially those with many stages are very sensitive to operation at very high flow. Under extremely high flow conditions, the shaft can up-thrust and buckle. This will knock out any seal.

Is the check valve a good tight shut-off? If the pump spins backwards when shut down, this will damage a gas seal.

Based on the description of damage indicating face contact, I would be most suspicious of the check valve, rotation or gas pressure differential. Up-thrust or buckling would be more catastrophic. Any additional details would be helpful.


Johnny Pellin

 

[longeron]

I beleive that a 28VL seal in this sort of service is usually a tandem, or dual unpressurized seal that operates a lot like a dry gas compressor seal. Unlike dual pressurized gas seals this seal requires fluid in the seal chamber to vaporize as it crosses the seal faces, hence, the VL or Vaporizing Liquid. In many installations this seal is operated with an API Plan 72/76, or with a low pressure Nitrogen or other inert gas purge and connection to a flare header or other collection system.

Since this is a start-up problem in a vertical turbine pump I'd first want to be sure that the pump and seal chamber are completely inventoried. Without adequate fluid in the seal chamber at start-up the seal faces may not be properly lubricated with fluid. Many times vertical pumps like this are installed with a vent from the top of the pump to suction as an API Plan 13. There may other connections to observe as well. Take a look at the cross sectional drawing for this pump and to verify.

It wouldn't hurt to run this problem past your seal vendor or salesperson for another perspective. It is in his or her best interest to be available to help you.

 

[reallyme00]

Dear Longeron and JJ,

Thanks for your inputs.

Yes the seal is tandem seal operates nitrogen as buffer fluid. This N2 dilutes the ethylene leakage and connected to the flare. The piping plan is 72/76. Process fluid is Liquid ethylene which is in vapour state at the stuffing box as per design operating conditions.

The seal has failed twice with in a year and plant commisioning delayed due to this failure.

1.The shaft rotation CCW and the seal designed according to that. We have witnessed the seal assembly and the direction of the spiral grooves are correct.

2. We have recently fit a pressure gauge on the stuffing box to monitor the pressure fluctuations. It fluctuates around 0.2 bar which is fine with the seal vendor.(still under monitoring)

3. We have checked the flare pressure (doubting that it would have created back pressure to the seal) but found it was less than 0.5 bar.

4. We need to check the check valve(as Mr.JJ said) as well as monitor the vibration lever frequently.

5. There are three line from the pump. One is recycle, another is balance line form stuffing box and the last is for venting. Except the balanceline, the other two are in closed condition.

However, We are working closely with the seal vendor to solve the problem as early as possible.

Once again thanks to all.