The reciprocating compressor valve failure

[hazel]

It was a terrible event that our reciprocating compressor cylinder has been heated to make the paint on the cylinder smoke found by CCTV.
We discovered the valve disk plate in the second stage inlet has broken, so the second stage didn’t work and cause the safety valve on the first-stage outgoing damper blew out. After about 10 minutes, we found the second stage inlet has smoke went out since the temperature rose extremely high.
During this process, we did not receive any alarms from the compressor and the temperature located in the damper indicated normally. We decided to add a flow meter in the discharge line to monitor all stages capacity.
Does any one have such experience and how do you prevent the problem of compressor valve failure?

 

[25362]

Although you do not disclose the fluid being compressed and the operating conditions, I suggest you visit thread1036-98536, and pay particular attention to what Montemayor had to say.

If I rightly recall an article on the reliability of hydrogen compressors in a HydProc issue of some 10 years ago, on discussing valves' failure, it concentrated on the type of valves as the main variable; other factors such as liquid carryover, process debris and cylinder lubrication, all of which may reduce valves' reliability, were considered less significant.

Trying to rank the valves on the length of their useful life, the author found that poppet valves are the best, followed by ported plate, concentric ring, both types with acceptable performance, and, closing the list, were channel valves performing from fair to good.

The article also referred to the valves' lift. It appears that the smaller the lift, the longer the useful life on lubricated cylinders. A lift of 0.08 in. resulted in about 8000 hours of service, while those at 0.04 in. experienced as long as 25,000 hours.

 

[pmover]

to supplement 25362 comments . . .

once recip compressor valves fail, the flow through that portion of cylinder (crank or head end) may decrease to zero flow. when this occurs, the gas remains in cylinder and will eventually increase in temp. monitoring gas temps at cylinder outlets &/or even cylinder ends are good indicators of some type of compression failure (i.e. valve, rings, or otherwise).

some operators use temp limit switches &/or have temp probes monitoring cylinder temps. excessive temps warrant cause for immediate shutdown of unit and activating an alarm.

please explain what is meant by "temperature located in the damper indicated normally"?

lastly, using a flow meter for measuring discharge flow may not be suitable for pulsating flow; especially if some cylinder ends are deactivated. i recommend investing in temp monitoring devices.

good luck!
-pmover

 

[Montemayor]

hazel:

I hate to tell you this, but you are absolutely correct when you state: "the temperature rose extremely high." I can assure you that had you had a temperature monitor and recorder you would have easily registered 500 to 700 ^oF. That's what burned your external paint off the cylinder and valve covers.

As 25362 notes, you don't give us the essential basic data, so we can just conjecture as to what you have left in front of you as a reciprocating machine. Is it an oil lubricated, air compressor? Is it 2,3, or 4 stages? What capacity is it? What make is it?

If you have an oil lubricated, air compressor, you are very lucky to survive the experience without an explosion within the air cylinder. Count your blessings that you may recover the machine with corrected machining and repairs; but more importantly that you didn't have any human being seriously injured or killed. Hopefully you have non-lube cylinders on this machine.

A reciprocating machine is a very forgiving beast, but you've got to start paying attention to the diagnostic and operating needs of such a reliable machine. Treat it well and it will reward you with years of faithful and diligent service - believe me! Get down to basics in understanding what makes the machine work: plain, common thermodynamics. Select the appropriate valves to give you the best and safest service - just as 25362 has indicated. If you can't do this yourself, find some one experienced and capable of doing it. Follow pmover's smart advice and pay attention to your discharge temperatures. They're trying to tell you something when they start to increase.

The best advice is: stick to the basics. Don't try to get fancy by overstudying the problem. Identify if metal failure (or something else) was the cause of the valve's rupture. Once you've identified the cause, install the correct temperature monitoring - for discharge temperatures and for bearings as well. Also concentrate on the machine's maintenance needs and logs. You will definitely see an increase in reliable service and safe opration.

If you have specific questions, you must furnish specific basic data - complete data and operating information. We can help you out if you help yourself by furnishing us with data and information.

Good Luck.

 

[quark]

When Mr Montemayor speaks of very basics I feel them like pearls of wisdom. The simple and the best things are monitoring temperature (apart from pressure) and doing a regular maintenance. A broken valve indicates that the machine might not have been touched by the maintenance guys for ages.

I always believe in feeling the machine by listening to it and touching it whenever I pass by. This helped me many times.

 

[Yorkman]

On our (York) semi hermetic reciprocating compressors every thing from 50 ton and up, we recommend that top ends be rebuilt with new valve springs and valves every 5000 hrs. I've seen these run to 7000 or 8000 and generally I find broken springs and cracked valves.. If the compressor is fitted with "wave" type springs we extend that to 10,000 hrs. It's sad that when budgets get reduced it seems that a good preventitve maintanace program is the first to get the axe.

I'm not a real engineer, but I play one on T.V.
A.J. Gest, York Int./JCI

 

[hazel]

I appreciated for every one’s help.
The followings are the operating and machine data:
Construction type: vertical
Number of stage: 2 , No lubricated.
Drive: motor direct
Compressor capacity: 733 KW, 11400 Nm3/H
Speed: 400 rpm
Medium handled: 93% hydrogen
Suction pressure: 1st stage 17 kg/cm2 , 2nd stage 30 kg/cm2
Discharge pressure: 1st stage 30.5 kg/cm2 , 2nd stage 54 kg/cm2
Suction temperature: 1st stage 30 deg. C, 2nd stage 35 deg. C
Discharge temperature: 1st stage 99deg. C, 2nd stage 110 deg. C
Safety valve set pressure: 1st stage 32.8 kg/cm2, 2nd stage 59kg/cm2
Valve type: concentric ring


Situation 1: the 1st discharge safety valve blew out.
Cause: the central part of the 2nd inlet valve seat had a leakage.
Activity 1: renew a valve.

Situation 2: the 1st discharge safety valve blew out.
Cause: the copper gasket between valve and frame destroyed by the cage locking the valve. This valve is very difficult to install and often causes the seat failure.
Activity 2: reinstalled the valve.

Then we have run the compressor about 20 hours before the event happened. After the compressor started up about 8 hours , the 2nd inlet valve temperature measured by operator was normal about 50 Deg. C. During the period, the temperature records for 1st and 2nd stages kept constantly. Why did the temperature not climb before the event? We thought the location of temperature sensor (PT100) is too far away from valve. The 4 inches pipeline distance from valve to sensor is about 2 meters. We have discussed with electrical person how to measure the real temperature on the valve, but did not get a final decision. Now we request operators to measure valve temperature every two hours by portable temperature tool. The other way we added a flow meter on the 2nd stage discharge to monitor the condition of the compressor. Since there was originally a local flow meter in the field, it is easily to transfer the signal to DCS and set a low flow alarm to alert operator. From this low flow alarm, we can know the compressor shall have some problems immediately. Definitely this is not the final solution, we will find a good position to install temperature sensor monitored valve temperature. Do you have any recommendations for the sensor position?

 

[Monagas]

Hello my friend,
Did you take look to the packing seal of the compressor piston?.
I think it could be completely destroyed and the main reason for leaks and high temperatures in the valves.
You should adjust the temperatures set points of the discharge in order to shutdown the engine in that cases.

Good luck & regards

 

[Montemayor]

Hazel:

You don’t say it, but I would guess your machine is a Burckhardt compressor and has a labyrinth ring design on the pistons. This doesn’t have a direct impact on the valve operation, but it would speak well for the design and experience put into the valve selection and operation. Hydrogen service is specially critical for valve design and I would urge you to download the following two .pdf articles:

http://www.dresser-rand.com/e-tech/PDF Files/tp102.pdf

http://www.dresser-rand.com/e-tech/PDF Files/tp015.pdf

Although you probably are not operating an IR compressor, IR has done extensive work on this application and has come up with a revolutionary poppet valve design. The aerodynamics and the valve design is all the same, regardless of the compressor make. The reputation of the manufacturer goes a long way in placing confidence in the expected performance.

If you are interested, you can look into custom-designed compressor valves as offered by such specialist fabricators as France and Hoerbigger valves. I have used both with satisfied results.

Your compression ratios are very conservative and well-balanced at 1.75 – 1.77. This falls in line with the conservative discharge temperatures. Although you don’t give your piston stroke, the 400 rpm indicates a conservative piston speed. I would not have expected a direct-connected electric motor drive as you indicate, but I’ve seen a lot of different ways to drive a compressor – including direct, integral steam engines.

With the exception of pulsation drums and everything else considered, I would expect excellent valve service and performance. Something is definitely wrong with the results you have experienced and outside of dirty gas and excessive pulsations, you should be getting good, stable operation. Have you consulted with your compressor’s manufacturer and presented the facts to him? The manufacturer deserves the first chance to resolve this critical and important problem. If they can’t/won’t help, then you may have to consult with a valve specialist company like France or Hoerbigger.

I hope this analysis helps you out.

 

[hazel]

You did really give me a lot of valuable knowledge and help, thank you.
There are some Burckhart compressors with labyrinth piston in our plant, but not this one. A company in Germany, which we don’t name it in the forum, makes this hydrogen compressor.
We have checked the piston ring and rider ring, it seemed available to use, so we didn’t replace any one of them. The condition of this compressor after service about one week, the 2nd stage intake valve temperature remained about 50 Deg. C
The articles downloaded from Dresser-rand are not easy to understand, especially the valve dynamics I stuck it for a time, however, the valve should be improved.
To change a valve type is a way to do and poppet valve is one of the better solutions, because it composes of many independent opening holes rather than the concentric valve has only a moving plate to seat. If one of the poppet valve opening holes failed, the others could still do the job well. If the concentric valve moving plate failed, it almost loses all its full capacity because it is only a plate and causes the discharge flow to approach to zero and then cause the heat problem. Do you think I am right or not?