8 replies to this topic

I have come across failure case for Lubricated Reciprocating Process Gas Compressor having Hydrogen >95 % with additions of Methene,Ethene, Water Vapour, and is operating with Valves having Carbon filled Peek material for valve sealing elements. These valves are earlier operated for life of 2 years without breakdown. Now we found to have valve sealing element getting melted after 12000 / 14000 operating hours.

These valves typically fails at Crank End only of cylinder and simulataneously two valves fails. No major observations seen on Head end side valves / other parts of valve assembly.

I would like to know if I have to investigate into Valve sealing material analysis or how should I go ahead into compressor problems investigation if existing at crank end only?

Pl advise, thx
Sunny

[iyer]

Well some things from the top of my head are:

1. Is the quality of the new valve seats same as that was when the valves were originally supplied. You may wan to ask the vendor for a certified supplier.
2. Has the lubricator flow been ok
3. the compressor jacket is not fouled as a result of which you have preferential flow of cooling water.

Hi
Seat quality ahhered to drawing standards and Lubricating flow was within limit per equipment manufacturers specifications.
Point of surprise is that why Head End side valves not getting affected with only Crank End sides. No indications seen for reflecting the failure pattern for mechanical overloading of valve. The failure resulted into thinning of non metallic valve sealing element and brittleness induced in sealing material.

Is there any possibility of hot water / liquid coming into cylinder from crank side?

BR
Sunny

[Art Montemayor]

Sunny:

If what you describe is true, you have serious problem in your compressor. However, you seem to lack any background or knowledge regarding reciprocating, double acting compressors. You state that the PEEK valve material "melted", but then you don't mention it again. You also do not give us other useful characteristics of the compressor:

• Is the compressor "mini-lubed" (Teflon or other non-metallic piston rings & minimum cylinder lubrication)?
• How many stages in this compressor?
• What are the normal, design suction AND discharge temperatures of ALL stages?
• What are the normal, OPERATING suction AND discharge temperatures of ALL stages?
• Is your cylinder handling ONE stage of compression? Or is it handling TWO different stages? An engineering sketch would make a BIG difference in describing this application.

You force me to assume a lot of the information because you don't furnish it. What I DO know is that with 50 years of experience and a lot of it with reciprocating compressors (lubricated hydrogen units included), is that your case is a simple one that should be quickly resolved. You simply have not supplied a lot of vital field information, and that can only be explained by your lack of experience in operating these type of compressors. This is not an error or mistake; it is simply a lack of experience.

Your double-acting cylinder is being obviously subjected to extremely high temperature in the crank end. The fact that you are not experiencing the same effect on the head end means that the piston rings have held and are probably integral. I would not trust that assertion and would check them out before putting the machine back into performance. I presume also that you have removed the machine from service; if you haven't, you should do this until you resolve the problem.

I am not at my home at this time and don't have my files with me, so I don't have access to the melting point of PEEK. However, I do know that it is quite higher than the 300 to 350 ^oF that I would set as the top maximum discharge temperature out of your compressor. I suspect that the design discharge temperature should be around 250 ^oF. You fail to tell us that useful information. You should be getting 5 to 7 years of useful life out of you PEEK material valves - as a minimum. This is extremely good material for reciprocating compressors and it is quite reliable - much, much better than Teflon. The discharge temperatures of the crankend side of your cylinder must be exorbitant - but yet you fail to even mention them. If you are operating a PEEK-supplied reciprocating compressor without any indication of suction and discharge temperatures, you are wasting a lot of very expensive investments in your compressor, the PEEK material, and the rest of the equipment.

Be aware that the only factor that could affect one side of a double-acting piston is a damaged or faulty discharge valve in that side of the cylinder. This, of course, assumes that the suction temperature to that side of the cylinder is as normally designed. You are experiencing "re-compression", which is the compression of hot, discharge gas that leaks back into the cylinder due to a badly seating discharge valve or a damaged discharge valve. This can happen at any time and is the reason your suction and discharge cylinder temperatures should always be monitored - whether manually or automatically.

I seriously doubt that there is anything wrong with your new cylinder valves. If they are indeed of PEEK material, then they are expensive and should be coming from a serious and experienced valve manufacturer. Also, you should be checking their quality of fabrication and leakage when you receive them. This is a routine operation in an operating plant.
Do not be concerned with the flow of lubricating oil. Oil lubrication in a recip is meant primarily for piston ring/cylinder lubrication and has absolutely nothing to do with a PEEK valve.
Any recip cylinder jacket is merely a token attempt to supply cooling to the compression part of the process. At best, it can only affect the cylinder temperature 1 or 2 ^oF. I have operated recips without any cooling water to the jackets and never witnessed any difference in operation after years of service. I am not advising you to do the same; I am simply showing you that it is not a factor in your problem.

The answer to your problem is in monitoring and controlling the inlet and discharge temperatures to your cylinder and the condition of your discharge valve(s).

Hi, thanks a lot

Summarizing all useful information of compressor for yr reference -

Single Stage compressor with two cylinders, Full Lube, double acting
Design Condition: P1= 35.3 bar, P2= 53.4 bar, T1= 131 deg F, T2= 206 deg F
Operating Condition: P1= 34.8 bar, P2= 54.3 bar, T1=125 deg F, T2= 203 deg F
RPM - 300

Normal operating temperature lies in range of 195 to 205 deg F but we did observe temperature rise up to 220-230 deg F when valve failed to operate.

This is Ring type valve, no of rings in valve #4

As I said earlier we have used other set of these valves without breakdown for more than 18000 operating hours. Inspection of failed valve do not shows sign of spring breakages, wear of seat land areas.

Critically, valve rings observed with brittle apperance of its surface with material across its thickness got compressed and left in form of only thinner segments. Roghly you can imagine that thickness reduced from 4 mm to 1 mm. Spring marks / impressions observed heavily on ring surfaces which made my thoughts over material brittleness phenomenon.

My concerns - Even valve manufacturer used high grade material having melting temp in range of 300+ deg F, and as such operating temperatures do not exceed to limits beyond 200 deg F, is that any chemical reaction then causing ring wear / material flow away? But then again in Crank End only?
We measure valve cover temperatures once in a week and never it shown any signs of detoriation.

How do you comment on how long valve can operate in environment of re-compression process leading to high temperature? Because I suspect this is sudden failure of two valves on crank side with abrupt changes in flow/temperature/liquid/???

Valve Installation mistakes should cause pre mature failures and should not be operating for long hours with stabilized operating conidtions.

Please advise,
Sunny

[akslzf]

Have you really experienced "melting" or is it a "mechanical damage"? A photo can help.

Anyhow, your questions arent technically informative, for assessment.

Why dont u go thro these two books, ask yourself this question and then come back?
http://books.google....page&q=&f=false
http://books.google....page&q=&f=false

Hi akslzf
may be my observation perception is not in line with your theoritical terminlogy.
attached photograph for ring damage observation.

Pl advise
Sunny

Attached Files

•  DSC08128.JPG   599.53KB   19 downloads

[akslzf]

The damage appears to be more of a mechanical nature. And it is at the sealing portion of the ring (except in one place, where it is in the middle).
Having said this, my comments are, presuming this is a suction valve :
1) if there's liquid carry over to a small extent, it can erode the seats, which I've experienced
2) if the compressor is run partly loaded (can you check whether the crank end was run unloaded for an extended period?), lube oil accumulates on the unloaded side & could cause this type of erosion / cavitation damage
3) I've once had a nick on the valve seat, that damaged the rings in no time. How's the valve seat in your case?
4) how were the springs? In one of my investigations, the springs in the valve were of varying stiffness!!
5) how's the pressure drop across the demister in the KO drum & how clean is the suction pulsation pot?
Coming back to "why crank side", if your answer to #2 above is no, it is most likely #3
If it is a disch valve, liquid and/or dirt MUST be the reasons.
Having said these,
a) the leakage should've caused at least a slightly higher valve temp and instrumentation/operator monitoring must have captured the high temp.
the damage is not too bad to lose sleep over

Edited by akslzf, 29 December 2009 - 08:17 AM.

Many Thanks for few hints, i am adding some more photos with respect to yr comments for better visual

This is Discharge valve, Photo will reveal nature of seat surface, ring damage.

at present I am not having data for pressure drop across demister in KO drum and suction pot cleanliness. Springs were alomst safe without any single spring broken / deformed in any manner. Stiffness of few samples selected found in limits per valve manufacterer report.

Please reply
Sunny

Attached Files

•  photo.doc   1.7MB   31 downloads