Gas Seal Flush on a Centrifugal Comresspor 1

[planck121]

Hi,
We've been having occasional gas seal failures on a multistage centrifugal wet gas compressor in our refinery. The failure was determined to be wet gas condensation (since the eht) on the flush line from the compressor discharge does not work reliabiliy and we are up north in Canada. The compressor start up sequence calls for nitrogen based start-up to the primary, second and teritery seals and once the compressor achieves stable operation. The seal flush is then changed to compressor discharge. But because of the condensation issue we are planning to run the compressor on nitrogen the entire time (with certain modification to the start/stop permissive) we have regulated nitrogen header pressure and in case we would run into issues with this header pressure we are planning on using the emergency back-up nitrogen skid bottles to trip the compressor in a controlled manner.

I realize this is a deviation from the SOP for compressors and the associated gas seals. However, would like your thoughts on what modifications (if any) would be required on the compressor to achieve the above senerio. I trust it should be as simple as plugging or blinding the flush line from the compressor discharge to the seal. This should take care of the wet gas from not flowing to the gas seal anymore and using external nitrogen from the main supply header. Also, are there other issues using nitrogen in a such a fashion for compressor reliability.

Thanks

 

[bcs5274]

Could you use fuel gas which is dehydrated/processed to get away from condensation issues??

 

[planck121]

bcs5274,
Using fuel gas is not possible, or atleast not practical I would believe, because the flush line inot the seal gas is coming from the compressor discharge and hence uses wet case as the product from the compressor. For the nitrogen we can actually tie off the main nitrogen header....In order to consider the possiblity of fuel gas, new tie in points and other containts would need to be mitigated and also the pressure of the fuel gas would have to be regulated so that the seal does not lose flow/pressure....

Though it is a good suggestion and point your bring up...

 

[Airsporter1st]

Hi,

The normal way to address this issue would be to dry the process gas before it gets to the seal chamber. This could be as simple as providing a suitable knockout type pre-filter to remove entrained liquids and aerosols, or it could involve dewpoint control (usually a heater). The only way to know what is required in this respect is to get a gas composition and draw a phase diagram. Then comparison against all operating points of the compressor (pressure vs temperature) you can determine whether a dewpoint problem exists.

Since the seal/system is designed to run on nitrogen at start up, then continuous running on nitrogen should not present any issues.

Generally the main consideration is nitrogen mixing with your process on a continuous basis. Provided this is not a problem for you then I would say go for it.

You will have some sort of dry gas seal panel associated with this and you presumably already have a connection point where the nitrogen/process gas enters the panel. There will therefore be either a changeover valve (manual or automatic) which determines whether you are running on nitrogen or process gas or it could simply be a pair of check valves, one on each supply line, such that the highest pressure wins. In either case you need to block the process gas inlet - having first ensured that the nitrogen is available.

If you can post a P&I.D. on here it will be easier to determine what must be done.

The last point to consider - is the nitrogen always available at a higher pressure than the compressor suction? Typically you need at least a couple of bar higher pressure.

Hope this at least gives some food for thought.

Regards,

Paul.

 

[pmover]

i will second Airsporter1st recommendations.

wet gas compressors need dry, particle free seal gas for long term successful operation.

if the failure is due to wet gas condensation, then the source gas (discharge side of gas compressor) is providing the moisture. when this gas is expanded at/near the mechanical seal, condensation is formed, thus failure.

investigate with mech seal mfg about using N2, but this should not be a problem. the critical factor is to maintain N2 pressures above compressor discharge pressures. the seal mfg will provide you this information.

Oh, the seal gas should pass through a filter before entering the seals to ensure no particle contamination in the seal gas.

hope this helps and good luck!
-pmover

 

[CaracasEC]

The dew point temperature of the process gas corresponding discharge pressure during normal and start-up was surely indicated at the P&ID. This should be verified. Normal filter set-up for this seal system have two filters. Aside form process gas compatibility, normal set-up is that the buffer gas which serves the secondary seal will be nitrogen and on some design is a separation seal buffer in between the bearing seal and the dry gas seal, however to supply nitrogen including the primary gas, i have not yet experience.

 

[Airsporter1st]

I have never known for dewpoint info to be included on a P&I.D.

Dewpoint issues with dry gas seals are very common - often because a truly representative gas composition is not available at the time that the system is designed, or because the gas composition changes.

API requires the temperature of the gas as it enters the seal to be at least 20 degC above dewpoint.

Thus the pressure and temperature at the point where the gas enters the seal are crucial and this has to be looked at in conjunction with a truly representative phase diagram.

It is not unusual to use an external source of supply, whether that be an inert gas such as n2, or simply gas taken from a cleaner, dryer part of the process e.g. an export header.

 

[JJPellin]

In general, nitrogen is an excellent gas to run to a dry gas seal. But, from our experience, there are two problems. First, nitrogen is expensive. You can easily determine how much it would cost using your nitrogen cost and the expected flow rate. The other problem is that nitrogen is inert. Some processes don't tolerate nitrogen. In a wet gas machine in our plant, the nitrogen injected would end up in fuel gas. Excessive amounts of nitrogen in the fuel gas can cause problems with process heaters. Small amounts should not be a problem.

We do not have dry gas seals in any of our wet gas compressors. But, we do have buffer gas on all of them. We use treated fuel gas for the buffer gas. Could you use a filtered stream of treated fuel gas for your gas seal to reduce the cost and impact on the fuel gas quality? Otherwise, we have high pressure natural gas lines in some of our units. Natural gas could be used for the gas seal at a lower cost than nitrogen. With the proper treatment, any of these gas streams could be made to work: process gas, fuel gas, natural gas, nitrogen.


Johnny Pellin

 

[DLite30]

My suggestion would be:

1. Modify the seal gas supply point on the discharge line to include a "quill" fitting or insertion probe type of fitting, so that any free liquids on the pipe wall will not make it into the seal gas supply line.

2. Possibly add a coalescing filter upstream of your seal gas regulator (optional).

3. Add a colescing filter with dump, downstream of the seal gas regulator. This should catch most of any free liquids fromed during the pressure cut at the regulator.