Compressor Tandem Dry gas Seal : Failure of inboard and outboard seal 2

[ed16]

Hello !

We are just performing a HAZOP for a Compressor Tandem dry gas seal. One question raised if the inboard (primary) and outboard (secondary) seal fails at the same time. To evaluate the risc i would be interested if this is a theoretical scenario or if someone has had this failure in reality.

Best regards
Ed16

 

[JJPellin]

I had this happen once on a John Crane 2800 double dry gas seal. This one was a very unusual case in an old seal design. I don't think it would be possible in a properly designed new system to fail in this manner. I will describe the failure that we experienced.

This seal was in a single stage overhung Roots machine running about 12,000 rpm. The seal was supplied with nitrogen gas using a very crude system. They had a regulator and a local flow meter with no instrumentation to alarm on loss of nitrogen flow.

The nitrogen connected to the seal using 1/2" SS tubing. A few inches away, on the side of the case, there was another piece of 1/2" tubing that was in impulse line to a panel mounted instrument that read process pressure. After an overhaul of the compressor, the mechanic reversed the two pieces of tubing. The area between the two seals was connected to the pressure gauge on the panel and the nitrogen was dumping into the case connection. After some time in service, they started leaking process gas out along the shaft. Both seals had failed catastrophically because there was no nitrogen flowing to the seal. The stationary seal face shattered and pieces of tungsten carbide dropped down onto the shaft and machined through the seal sleeve and about 1/4" deep into the shaft before we got the machine shut down.

At any other similar installation, we would have instruments to control the nitrogen buffer gas to a certain differential pressure above process pressure. We would monitor the buffer gas flow rate, as well. As long as nitrogen DP was maintained it should be impossible for process gas to leak directly to atmosphere past both seals. I don't believe we would trip the compressor off on loss of buffer gas DP. But we would definitely alarm and take immediate action to restore buffer gas flow or shut the unit down.

So, from a HazOp standpoint, I don't believe a double dry gas seal properly instrumented for differential pressure and buffer gas flow rate should ever fail catastrophically enough to blow process gas directly to atmosphere past both seals. From a HazAn, LOPA standpoint, this should be in the category of 1/100 years or less.


Johnny Pellin

 

[flexibox]

It is highly unlikely to have both seals fail at the same time in a tandem seal configuration. As you are asking about Tandem seals (Dual Unpressurized), I have seen a condition whereby due to poor design of the gas panel,(lack of instrumentation) the secondary seal had been operating in a failed condition. Some Tandem seal installations do not have an intermediate labyrinth and therefore no Nitrogen to the secondary seal. The separation gas had failed, allowing Lube oil into the secondary seal cavity. The oil entered the seal and eventually the faces made contact and failed the seal. The primary seal was still intact and the compressor was still operational. The gas panel was not designed correctly and there was no alarms to tell the operator what was happening to the secondary seal or at the N2 supply to the separtion gas. It was unknown how long the secondary seal had been running in a failed condition but upon inspection of the seal the Silicon carbine face was found shattered. So although it is highly unlikely to have both seals fail at the same time, without a properly designed gas panel and depending on the circumstances, you can end up with Process gas through a seconary seal.

It may be worth inviting your seal vendor to the HAZOP

 

[longeron]

Flexibox mentioned the intermediate labby. I believe that even in a tandem seal without N2 purge that this device is important. If something were to happen to your inboard seal faces the labby would help to shield the outboard faces from debris that would likely cause it to fail immediately. This would give your operators more time to properly shut down your compressor.

You'll find this same device (or at least a bushing) in 682 (3rd ed.) seals with a dry running outboard as well.