Problem on DGS for H2 Recycle Gas Compressor 2

[zozo21]

Hi everybody;

Please help us; we have a new (03 years old) H2 recycle gas compressor for reforming unit equipped with dry gas seals (DGS). For 3 times, when we shut down the plant and we restart again, the DGS fail (high pressure in the primary vent) it mean the primary seal is failed, so, we stop the compressor and we change by a new DGS. As per the vendor of the DGS when he made his repair report, there is some dirty between the two faces of the first seal witch they conduct to deteriorate them. Our question from where this dirty comes and how to avoid this.

Thanks.

 

[JJPellin]

You have not provided enough information. The “dirt” most likely came from the process. This suggests that you did not use correct procedures for shut down, lay-up or start up. But, I would need more information to say more:

1. Is the seal a tandem or double configuration?
2. What gas is used for primary seal gas, secondary seal gas, buffer gas and seperation gas?
3. If purified process gas is used for primary seal gas, do you have an alternate gas supply for start-up and shutdown?
4. Do your procedures maintain positive differential pressures during the pressure/depressure process at unit start-up? (O2 free process)
5. If you use a gas conditioning unit, what level of filtration is used?


Johnny Pellin

 

[georgeverghese]

Dry gas seals operate with seal gas pressure in the primary seal area > max compressor suction pressure. This should also mean that seal gas pressure should be greater than compressor settleout pressure. It does sound like your primary seal gas pressure is less than max possible settleout pressure, which is why you have dirty process gas flowing in the reverse direction into the seal on shutdown.
Also, check that your seal gas pressure is adequately dewpointed, ie Tdew < min ambient temp at the required operating pressure of the primary seal.

 

[rotw]

I agree with JJPellin post, quite relevant questions to troubleshoot this; I would like to add some questions/comments myself:

- During shutdown, do you keep the compressor pressurized? Do you have long stand still period or short (<12 ~ 24 hours)
- Regarding process gas, do you have heavy hydrocarbons (e.g. C5 and beyond) mixed with main hydrogen gas and has the DGS vendor or OEM some seal gas condensation risk analysis in the past based on normal conditions and settle out (if applicable) ? has the gas composition changed between design / actual?

In some cases, when the gas is clean (no risk of condensation), negative flow through the primary seal (assuming a tandem configuration, which you need to confirm), can be tolerated - temporarily.

What is the driver type ? steam turbine ? do you have slow roll condition?
Lastly, is it a bidirectional or unidirectional type of seal ? reverse rotation during shutdown, could occur before pressures stabilize and could damage unidirectional seal.

Do you monitor performance of DGS? i.e. flow to primary seal ?

 

[JJPellin]

We have six hydrogen recycle compressors in hydrotreater service and two more in reformer (Powerformer and Platformer) service. Both of the machines in reformer service are prone to the formation of water soluble salts. In my experience pure hydrogen in these services is far from pure. One of these two machines is equipped with dry gas seals in a double configuration. This makes it very sensitive to reverse pressure during shutdown, layup or startup. We seem to have the most difficulty when the unit is pressurized and depressurized repeatedly in order to achieve the required O2 free conditions needed for startup. We had to write very specific procedures and provide technical oversight in order to be sure that the seals did not experience reverse pressure during the O2 free process.

We have also seen problems during layup when nitrogen for separation gas was not available. Running the bearing lubrication system without separation gas creates a risk of lube oil getting between the seal faces. This oil would tend to coke off in operation which could be a source of “dirt”. From your original post, it is not clear if the failure is on the inner seal or the outer seal. A schematic of your system would be very helpful.


Johnny Pellin

 

[rotw]

JJPellin: why did you select double DGS in hydrogen service? its a question just out of curiosity.
having some hydrogen released in the vent line should not be a big issue (from toxicity / safety view point).
Tandem type DGS is "generally" better referenced than double DGS, so it should be the first selection option unless some strict standard applies on emissions to atmosphere (high H2S content, etc.).

Double DGS should be an option when suction pressure is limited, say below 6 ~ 10 bar, otherwise your plant N2 pressure would not be sufficient to buffer the seal. In addition, you need a back up source of N2 available permanently, either bottles or a small booster, is this your case?

Another question, JJPellin: when separation gas was lost, you also should have lost N2 as main DGS buffer, thereby tripping the machined immediately - am I correct ? otherwise do you apply some particular philosophy for interlock?

Back to Tandem configuration, which is also hypothetical unless OP clarifies it, contamination of the secondary seal by the lubrication oil seems to occur at low speed, e.g. during start-up, and I was told by some OEM that at rated speed, centrifugal forces associated with a specific design of the shaft are normally acting as a temporary obstacle for the oil to flow toward the secondary seal. I also think the OP has issue with primary seal, as secondary seal failure would not trigger high pressure on primary vent, typically.

This is all speculative, unless OP gives more insight.

 

[georgeverghese]

Dewpointed and filtered primary seal gas supply to the seal, given that its pressure is greater than compressor settleout pressure ( by say at least 1bar or so ), should be on FIC control and not on PIC control.

If there is no way of getting the primary seal gas supply pressure to be higher, then you may be forced to auto blowdown the H2 recycle compressor for all shutdown cases to avoid fouling the seal faces with process gas. Auto blowdown of the compression circuit is also the typical course of action for loss of N2 separation gas on the outer seal. We have also used compressed dry air as the separation gas in some compression trains when N2 gas or inert gas was not available.

In all our large GT drive compression trains, an automated purge sequence is used to clear the circuit of any air / O2. A dedicated automated purge gas valve is installed on a parallel line across the main feedgas supply SDV for this purpose.

 

[JJPellin]

I did not select the double DGS. My predecessor made this selection. I would have likely selected a tandem as you suggest. Our machine uses pure makeup hydrogen (PSA hydrogen) for both the primary seal gas and buffer gas. Nitrogen is only used for separation gas. My concern about the loss of separation gas was when the machine was not running during turnaround. So, trip interlocks don’t come into play. And shaft rotation cannot protect against oil migration into the secondary seal.

It is most likely that the OP is describing primary seal leakage on a tandem arrangement. Thus, separation gas would not be relevant.


Johnny Pellin

 

[zozo21]

I try to give you more details
- Seal arrangement: Tandem with int labyrinth / bidirectional.
- Driver type : steam turbine with slow roll condition
- For primary seal, we use N2 for startup and shutdown, and H2 from 2 nd stage of compressor in normal operation.
- For secondary seal, we use nitrogen (N2) and for separation, we use also N2.
- The supply lines to the gas seal is equipped with coalesce filter elements with grain size < 5µm in addition the use of gas heating (steam tracing).
- Gas composition: H2 + C1+ C2 + C3 + iC4 + nC4 + C5.
- For DGS monitoring : we have PDT for primary seal vent and FT for secondary seal supply.

Finally, I attached two pictures the first DGS failure in 2014 and the second DGS failure in 2016.

 

[rotw]

From the pictures and maintenance crew logo, it seems the installation is in more or less tempered climate country. What I mean is that low ambient temperature at site should be ~-5 deg C or that order of magnitude, but not too far on low ranges.

And still you have, by design, seal gas heating + heat tracing, which means that the gas composition was well known to be prone to condensation, so the seal gas has been overheated in order to keep its actual temperature distant from its dew point, I guess ~20 degree C as minimum margin should be enough. The very fact that C5+ exists in gas composition indicates that this potential for condensation is real. By the way, beware the uncertainty on the components higher than C6+ - trace level can dramatically shift the seal gas dew point.

Fair enough that seal gas heating is included, but we need to be clear on something, that is often an overlooked problem (its only recently that OEM & technical community in general admitted this to be an issue) which is as follows:

You could super heat the seal gas and heat trace it to keep the seal gas distant from the dew point, BUT - as soon as the seal gas enters the seal assembly / compartment, there is no mean to control its temperature. Lets skip the stand still condition and look at normal operation only at low ambient: the seal gas assembly (wall temperature) can happen to be at lower temperature than seal gas dew point whereby condensate can form when the gas contacts the metal. Its quite typical the next planned start-up, is the condition where seal deteriorates mechanically due to the formation of condensate.

I am not saying this is the explanation, but it is one plausible scenario to evaluate.

Question: has the problem same level of severity on both the drive-end and non-drive end DGS's?

 

[georgeverghese]

So is your compressor settle out pressure less or more than than backup / startup N2 seal supply gas pressure ?

 

[zozo21]

the leakage is in the both side DGS's (ND and NDE side).
settle out pressure = 12 kg/cm2G
startup N2 pressure = 6,5 kg/cm2G

 

[georgeverghese]

Since primary seal gas demand is usually low, it may be possible to use a Haskel pressure booster to increase the N2 pressure to 13-14barg and operate the seal gas supply on FIC. A small buffer vessel may also help to store this N2.
A Haskel booster in this application would use a slipstream of N2 at 6barg to boost the main supply N2 from 6barg to 13-14barg to feed the FCV.
Check that this 12barg is the max possible settleout pressure of the compressor - there may be some failure scenarios that result in higher settleout pressures - ask a process engineer to build a model of this worst case scenario on Pro II or Hysis and run the settleout simulation.

 

[rotw]

georgeverghese,
I don't think that the original design just overlooked the issue of buffering the compressor for settle out pressure condition. They should have some mitigation in place, whether it is short term or longer pressurized stand still. What I am saying is that its unlikely that such a 'big elephant' in the room would have been missed...

 

[georgeverghese]

@rotw, Agree, it is unusual that settleout pressure wasnt considered in the backup N2 supply seal gas provisions for restart - maybe the recycle compressors were meant to autoblowdown on shutdown in the original design concept ?

 

[zozo21]

Agree with you because in the initial design the gas seal start-up and shutdown was Hydrogen (82% H2 purity) at 15 kg/cm2G and for unknown reasons they changed to the N2.
Following chart, which simply shows, different operating modes of the compressor, in order to ensure availability of external seal gas (Nitrogen in this case) at all transient condition.

Operation Mode Description N2 availability for primary seal

1- Initial Slow Turning at 16 RPM Casing Pressure 0.5 Kg/cm2g (Start-up) Yes
2- Pressurization / Loading 0.5-6.8 Kg/cm2g Yes
3- dynamic operation 6.8 Kg/cm2g Hydrogen from 2nd stage of Compressor
(suction pressure / sealing pressure)
4- Shut-down / Settle Out 9.5 to 12 Kg/cm2 g Hydrogen from 2nd stage
(casing isolated and pressure balanced)
5- Shut-down / Depressurization 6.5 down to 0 Kg/cm2 g Change primary seal gas to Nitrogen
0 RPM, Venting the Casing
6- Cool Down Slow Turning at 16 RPM Yes

Another point change over from 2nd stage discharge to Nitrogen supply is manual.
My first question is with this startup and shutdown procedure there is a risk of damaging the seals?
The second question can you explain me what happens exactly in settle out in this condition (N2 at 6.5kg/cm2G as seal gas for primary seal)?

 

[rotw]

Agree with you because in the initial design the gas seal start-up and shutdown was Hydrogen (82% H2 purity) at 15 kg/cm2G and for unknown reasons they changed to the N2

1/ Can you investigate what are those reasons exactly? Is it process-related reasons ?


2/ I am also unclear about start-up from the sequence you described.
You are explicit on use of N2 for shutdown but for start-up, I can only guess as follows

High purity H2 has been used in orginal design to buffer the machine during start up? is this correct ?
But this gas has been replaced with N2 as part of changed philosophy. So during start-up, N2 is used as primary seal, then when compressor ramps up to a threshold speed / pressure level, the seal gas source is switched from N2 to H2 taken from a high pressure source [which could be of different gas composition than compressor process gas, but accounted for by original design]. Question is - is the switch from N2 to H2 automatic or done by an operator?

3/ what is the typical duration of your settle out condition ? Refer to my previous post related to gas entering seal gas assembly...
Do you have risk analysis report performed by OEM/DGS vendor for the gas composition [H2 taken from high pressure source, your so-called '2nd stage' pressure source]. A longer duration for settle out condition, poses even higher risk on dry gas seals integrity if possibility of contamination exists because of the scenario I described.

Is your machine feeding the 2nd stage ? when subject machine is shutdown, is 2nd stage operating or stopped too ? Is not a must to know this but just to complete the understanding as I assume the availability of H2 from that source is warranted anyway.

 

[zozo21]

1-In the original design, the source of pure hydrogen for start-up and shutdown was from cylinders at 110 bar and to draw from 110 to 15 bar some modifications were necessary to do.
Until now we have not done.
2- The startup procedure, is as your describtion.
3- The switch from N2 to H2 is done by an operator.
4- The duration of settle out condition is about 15 to 20 minutes.
5- The 2nd stage is the second discharge of compressor and in shutdown condition we close the suction and the 2 discharges valves.

 

[georgeverghese]

The aim should be to auto switch over from normal to backup source for seal gas on compressor shutdown. At no time should there be a potential for flow reversal. A buffer tank on the supply line should help to ride out interim pressure fluctuations during the switch over.
Reverting to normal supply H2 from the backup source can be manual, since the startup event is manned locally.

 

[rotw]

This is most likely a problem of management incompetence lying with the owner. Any DGS vendor could trace back the root cause of this easily. It requires competence and know how to have provisions included in third parties contracts that protect the end user and allow to hold parties accountable for any design flaws, non conformities, etc. in short all of the obligations making it mandatory for the OEM/DGS vendor to do their upmost to fix such kind of problem.

Because what is the OEM and DGS vendors doing? sitting back and waiting for next failure to happen?
Or maybe owner is happy to buy new DGS cartridges every now and then, so why bother.

Sorry for the direct approach, but it appears it is not machinery process aspects which would be the most worrying.