PDA

View Full Version : Howden compressors with different oil injection



Mark VI
09-01-2008, 01:44 PM
Hi everybody,

This is my first post to the site!!

I have recently arrived to India for the start up of one quite big NH3 installation with Howden 321 machines.

The oil injection circuit is made in quite a tricky way: the oil is directed straight from the oil separator up to a dry oil cooler located at the roof of the plant and then the oil goes back to the unit, where it is divided in two streams: One stream going directly for oil injection (without pump) and other stream going for bearings lubrication & slide valve control & balance piston pressurization through a small gear pump.

The system makes sense, but we are finding problems with cavitation in the pump and consequently pressure lost in the oil pressurised line, and the system tripping.

Have anybody deal with a similar system before? I have always seen the pump moving all the oil from the separator...:confused:

Mark VI

Tycho
09-01-2008, 07:47 PM
Hi everybody,

This is my first post to the site!!

I have recently arrived to India for the start up of one quite big NH3 installation with Howden 321 machines.

The oil injection circuit is made in quite a tricky way: the oil is directed straight from the oil separator up to a dry oil cooler located at the roof of the plant and then the oil goes back to the unit, where it is divided in two streams: One stream going directly for oil injection (without pump) and other stream going for bearings lubrication & slide valve control & balance piston pressurization through a small gear pump.

The system makes sense, but we are finding problems with cavitation in the pump and consequently pressure lost in the oil pressurised line, and the system tripping.

Have anybody deal with a similar system before? I have always seen the pump moving all the oil from the separator...:confused:

Mark VI


Sounds kinda odd, on our units we usually have the oilpump suction directly on the oil seperator, making the flow like this: oilpump -> oilcooler -> oildistributor pipe (slightly larger pipe where all the pipes for compressor lubrication are fitted) with a throttle valve for the oil injection.

I have no problem understanding you have problems with the pump if the pipe goes from the compressor and up to the roof then back down to the oilpump.

if the oilpump starts at the same time as the compressor it would take a little while for the headpressure to build up and refill the pipe with oil and the pump will run dry for some time.

Tycho
09-01-2008, 07:51 PM
Sounds kinda odd, on our units we usually have the oilpump suction directly on the oil seperator, making the flow like this: oilpump -> oilcooler -> oildistributor pipe (slightly larger pipe where all the pipes for compressor lubrication are fitted) with a throttle valve for the oil injection.

I have no problem understanding you have problems with the pump if the pipe goes from the compressor and up to the roof then back down to the oilpump.

if the oilpump starts at the same time as the compressor it would take a little while for the headpressure to build up and refill the pipe with oil and the pump will run dry for some time.


Oh and I would have moved the pump so that it's suction is directly on the oil seperator, and made it so that it starts up a few seconds before the compressor to prelube it, or put a non return valve in the rising oil line to prevent the oil draining back to the seperator during standstill (the valve is the easy solution, but then you also risk the NRV siezing if you get a little dirt in the oilsep, as it will be before the oilfilter)

US Iceman
09-01-2008, 09:45 PM
...the oil is directed straight from the oil separator up to a dry oil cooler located at the roof of the plant and then the oil goes back to the unit, where it is divided in two streams: One stream going directly for oil injection (without pump) and other stream going for bearings lubrication & slide valve control & balance piston pressurization through a small gear pump.


Time to get the package manufacturer involved if these are new skids. Someone forgot to include the pressure losses in the pump suction piping, so of course the pump cavitates. The pump is pulling oil through all the piping, the oil cooler, plus the elevation difference. It's no wonder the pump cavitates.:eek:

Like Tycho mentioned, if the oil pump pulls the oil directly from the separator then pumps the oil to the cooler and injection manifold it's OK. I have done several like this with remote air-cooled oil coolers in a gas plant and it worked just fine.

TXiceman
10-01-2008, 05:36 AM
Like above, the line losses are killing the oil pump. This draw through system is used on some close coupled systems with a local oil cooler of liquid injection.

The line losses in the oil pump suction are creating flash gas as the pressure reduces on the oil. They are not cooling the oil enough to get the refrigerant to recondense in the oil and starving the oil pump. They need to do a NPSH calculation on the pump suction system.

I agree on moving it back to the oil separator and making it a push through, but they are cooling all of the oil and only pumping a small portion of the oil. It is basically a poor design and they were trying to copy another design.

ken

Mr. Ammonia
10-01-2008, 10:18 AM
This is what in Europe call "optimised design". I think I saw something similar from a dutch company called SES, but that time it was working OK without any problem.

If you can not move the pump suction to the oil separator check the oil viscosity. This would help.

Mark VI
10-01-2008, 04:58 PM
Thank you for the ideas.

I agree with TXiceman: most probably they were trying to copy a previous design. Just adjacent to new packages, there are another 3 very similar with the same compressors which have been running for the last nine years without any problem. The point is that in the already running units they have no problem in the oil pumps!!!

As Tycho was suggesting, there is a NRV in the oil separator oil outlet and the units have pre-lube for 12 seconds before starting.

Anyone knows what the difference can be in the new units?

US Iceman
10-01-2008, 10:44 PM
Anyone knows what the difference can be in the new units?


Well, we don't have a lot of information to look at to see the nature of the differences!



...there are another 3 very similar with the same compressors which have been running for the last nine years without any problem...


It's not that the compressor is the problem, it's the package design. Obviously, since you say these are "similar" packages there must be some difference.

Even if the new package is similar to the previous packages there is sufficient evidence to suggest whatever small difference exists is the nature of the problem observed.

That's what you need to find. Do you have P&I drawings for the two packages to compare for differences?

TXiceman
11-01-2008, 04:22 AM
I'll bet the difference is the location of the oil coolers....

ken

Mr. Ammonia
11-01-2008, 10:56 AM
I will bet the difference is the oil pump speed (rpm)....

Mark VI
11-01-2008, 11:07 AM
Well, we don't have a lot of information to look at to see the nature of the differences!.... Do you have P&I drawings for the two packages to compare for differences?

Thank you for the interest US Iceman. I really don't see any difference in the new & old systems. :confused: If you give me an e-mail address I can forward them to you!

Tycho
11-01-2008, 10:33 PM
Thank you for the interest US Iceman. I really don't see any difference in the new & old systems. :confused: If you give me an e-mail address I can forward them to you!

Could it be the NRV valve in the oil line is not opening all the way?

Maybe they had the same problem with the other units when they were new?

could be that the welders didnt clean well enough and metal grit now has gritted up the NRV... what make and model is the NRV?

TXiceman
12-01-2008, 03:11 AM
Mike, I'd be interested in seeing anything you get...

Ken

US Iceman
12-01-2008, 07:01 AM
Thank you for the interest US Iceman. I really don't see any difference in the new & old systems. :confused: If you give me an e-mail address I can forward them to you!

If you click on my screen name (US Iceman) and then send me a private message with you email address, I can send you mine.

Once I receive the drawings, I'll forward them to TXiceman also for his comments.

How's that for service?:D

Powered by NH3
13-01-2008, 12:59 AM
I have never seen this before and some questions come to my mind:
What kind of oil are you using? Is this Synthetic or Mineral? What about the viscosity?
I would say that the pressure losses are affecting more than expected in the pump suction, however I do not understand how the "old" units are running properly if they have similar design!

Josip
13-01-2008, 01:23 AM
Hi, Mark VI :)


Hi everybody,

This is my first post to the site!!

I have recently arrived to India for the start up of one quite big NH3 installation with Howden 321 machines.

The oil injection circuit is made in quite a tricky way: the oil is directed straight from the oil separator up to a dry oil cooler located at the roof of the plant and then the oil goes back to the unit, where it is divided in two streams: One stream going directly for oil injection (without pump) and other stream going for bearings lubrication & slide valve control & balance piston pressurization through a small gear pump.

The system makes sense, but we are finding problems with cavitation in the pump and consequently pressure lost in the oil pressurised line, and the system tripping.

Have anybody deal with a similar system before? I have always seen the pump moving all the oil from the separator...:confused:

Mark VI

A lot of guestion but without any scheme...it is not easy to tell what is similar (if?) and what is not...I'm also wandering what is the real difference in between new and old units...something must be...

Mike, put me on your mailing list;), please.

Regarding the size of compressor is it maybe on buster side where we already have a small oil pressure difference...and then you have a very questionable oil cooling pipe layout, what about pipe sizes....what about oil pressure regulator you have it or not....

What about oil filters...running in type or permanent....maybe clogged.....

Maybe you have too big pipes to oil injection and need to throttle that line there should be some regulator...be careful not too much (you will have too high discharge temp)...


Best regards, Josip :)

US Iceman
13-01-2008, 05:47 AM
Mike, put me on your mailing list;), please.


Sure thing Josip. As soon as I receive them.

TXiceman
13-01-2008, 05:49 AM
Similar is not the same as exactly the same. I am thinking there is a difference.....

I have seen may times where a machine is applied or designed in a similar application and it does not work....They are not EXACTLY alike.

Ken

US Iceman
14-01-2008, 07:39 PM
I forgot about that. Since you are a new member I think some of the features are limited until your post count increases.

I'll send you and email, and in the meanwhile I am deleting you message so that your email address is not public (no spamming).

Gerben
17-01-2008, 10:41 AM
Hi Mark,

could it be that the discharge temperature is below 73 C? Lower than 70 Cyou will get too much liquid in the oil. The liquid ammonia turns into gas shape and makes the oil pressure drop/hunt.
We have had the same problem in Belgium with also a WRV321, and an oil cooler 12 metres higher.

Gerben
RM-Support.nl

Mark VI
17-01-2008, 07:51 PM
Hi Gerben,

This is a good idea. When I started to feel the problem, I reduced the oil injection until I reached a discharge temperature of 95º C, but the problem is still the same.

Now I am thinking that perhaps a big bubble is already up there and there is no way to take it out of the system. One interesting point is that in the last unit I started, everything is going fine :D, but in the previous two I can not recover the oil pressure; it is even going to lower values :(!

In your system, the oil pump was placed before or after the oil cooler?

Mark

Gerben
18-01-2008, 04:24 PM
Hi Mark,

are all units the same?
Maybe it is possible to make a very small line from the top to the suction side, to get the gas out. If it works, you at least know what the problem is. This small line, for example 6mm, will not have any effect on your power. Gas will go through fast, oil will go very slow.

We allways place the oilpump before the oilcooler, so you also cool down the generated heat of the oilpump.

Good luck,

Gerben
RM-Support.nl

US Iceman
21-01-2008, 05:24 AM
Dear Mark VI,

Thanks for the P&I drawing. I sent it to TXiceman and Josip for their review also.

Based on what I see you have the following items in the pump suction line; 3-way temperature control valve, filter, & several service valves. That seems like a lot of pressure drop in the pump suction line assuming only normal operation.

As to the oil cooler it appears to be not connected to the pump discharge line at all. It looks like the oil supply line (to the oil cooler) comes off of the oil separator sump (on the lower right side of the separator sump). Then, the oil return line from the oil cooler feeds back into the non-pumped oil manifold. This is also the same line that is connected back into the 3-way temperature control valve on the pump suction.

When the oil is cool the 3-way temperature control valve would be closed so no flow should occur through the oil cooler. However, when the oil is hot the 3-way temperature control valve would close the bottom port off, which would force the pump to suck oil through the oil cooler (from the separator sump). This would of course cool off the oil prior to it being sucked into the pump.

If as you say all of the compressor packages are IDENTICAL, then I would investigate the design pressure loss of the oil cooler and piping. The oil cooler may have different coil circuiting that has a greater pressure drop than the other coolers. The same could be said for the 3-way temperature control valve, filter, or piping also.

Just based on what you have described and provided my guess is the oil pressure may be OK until the 3-way temperature control valve opens for oil cooling. When the oil gets warm, the excess pressure drop of the oil cooler, piping, and/or 3-way temperature control could create a cavitation problem for the pump and cause a loss of oil pressure to the oil pumped manifold. A pump should always be pumping through a device, never sucking through it.

If indeed there is a filter in the pump suction line as it seems on the drawing you might also check the differential pressure across this filter also.

That's the best I can do with this based on the available information. I hope that helps.

Best regards,
US Iceman

Martin Wierbosc
27-01-2008, 08:50 PM
Hello Mark,
I am Martin. We know the howden compressor a bit.( inside and outside)
Howden delivers only bare shaft compressor. builders make the unit. Looking on the howden compressor , you find 2 lines to the bearings and one bigger line to the cooling of the compressor. I can send you details. best you can use the oil cooling line to keep the discharge temperature under control. to low temperature results in a less efficient oilseparator system. if you send me the kind lubrication oil and the temperature at the entry of your oilseparator, discharge temperature and temperature of the oil into the compressor we ca help you. Perhaps you miss the balance of you unit, it can easy be restored. If the oilseparator not operates efficient, the oil to your oilcooler and to the compressor can contain to much refrigerant. as a result you can expirience low viscosity ( less pressure of the pumps), by gear pump even damage( mostly bearing and shaftseal failure) . also by the compressor you can expirience at the shaftseal, in the long term lack on lubricatoin problems. We did a study for some years and this resulted in facts, even we did not this expected. But makes some symptoms understandable.

Best regards
martin Wierbosch RM support. martin@rm-support.nl

TXiceman
28-01-2008, 02:58 AM
By the nature of the design, a vertical oil separator is going to be less efficient than a horizontal separator. Also, the vertical does not have as long a residence time for the oil and will not allow as much time for the oil to degass.

The only way I have seen the pull through designs work properly is to have a local mounted (close by) oil cooler. The pressure differential (or pressure loss) through an air cooled external oil cooler is going to be much higher than a shell and tube type. The other problem with the design as pointed out by US Iceman in the 3-way valve also puts too much line loss when included with the extra losses through the remote oil cooler and the extra piping.

The pressure loss or reduction in the line causes the refrigerant to flash out and create a foamy oil for the oil going in to the oil cooler. The oil cooler is designed to cool oil, not a foamy mixture, so it will not properly cool the oil. The pump will get a warm and foamy oil to the inlet and cause the pump to cavitate and the compressor to have a low oil pressure and high oil temperature.

The type oil you are using could have a big effect on the compressor operation. If the oil will carry too much refrigerant, the problem is made worse. If you are not using a synthetic or at least a semi-synthetic oil, you need to check into an alternate oil. You can contact CPI Engineering in Midland, Michigan, USA.

Also, check the location of the oil return line (bypass) into the oil separator. If it is dumping right over the oil outlet, it will only make the problem worse.

I have worked with Howden and Mycom compressors for about 35 years now and have seen lots of problems caused by the improper design and application.

Let us hear how you are doing on the solution.

Ken

Mark VI
04-02-2008, 03:48 AM
Dear all,

Thank you everyone for the suggestions.

I have been working hard these days in the plant, as the customer wanted to have more cooling capacity urgently and he requested the new units to be working by first of February. I explained him about the problem I was finding in the system and how it could be that something which I don't know is impeding two units to work properly. For him the issue was clear: The only difference was the start up guy (me) :(.

So I have put the only unit which is giving good performance to work in automatic mode and I am focusing now day and night in the other two units.

After all the ideas we have interchanged here, it was quite clear that the problem is coming from an ammonia excess in the oil (or by too much pressure drop). I wanted to test that point, so I inserted a sight glass at the oil pump suction line. It was incredible: Hundreds of small bubbles were travelling alongside with the oil!! :eek:

The next step was trying to eliminate these bubbles and the problems started there. I have tried nearly any solution: Very high discharge temperature (110º C), reduction of the oil flow to the rotors, so that oil cooler pressure drop is reduced, cleaning the non return valve in the oil line, etc, etc, but the bubbles don't disappear!!!

Still I am thinking that there are big gas bubbles inside the oil cooler and I don't manage to take them out. These bubbles would be originating this phenomenon. The point is that the oil cooler construction is quite tricky: 14 meters long horizontal tubes arranged in a 4 pass configuration, with the oil entering at the top and going down. The air is going upwards, which is common sense from a heat transfer point of view.

I think the problem is just there: During normal operation, the oil is going down in the six passes, pushing the gas down until it reaches the main header, where the gas is collected as the pipe diameter is bigger, hence speed is lower. As soon as I stop the plant, the gas goes up, but it stops in the last pass (lower) of the oil cooler. The tubes are so long, that the oil can not flow upwards easily. In other words, when the gas is formed inside, it is very difficult to take it out.

If I made same mistakes during the first start up, the gas has been formed there and it will not go out in a natural way.

Just to be sure that the problem is up at the oil cooler, I am going to replace the thermostatic three way valve with a manual three way valve. I will take the oil for the pump suction straight from the oil separator and if the problem stops, it will be clear that the problem is in the oil cooler. The weather is now cold and I hope I will be able to run the plant for some hours before the oil temperature gets too hot.

I will update you with the results.

Best regards,

Mark VI

Josip
04-02-2008, 04:59 AM
Hi, Mark VI :)

yes that is obvious...new start up guy...the only difference for some people:rolleyes:....unbelievable...

...what about your econimizer...maybe you have to much liquid/wet gases there passing together with suction gases into oil separator and then go back with oil again...that connection is 5"....maybe to try with completely closed ECO connection...

....what is your suction superheat...

....according to scheme seems there is not any other connection with liquid ammonia or maybe it is not shown...

I have your P&I diagram (sent by US Iceman) with compressor, but nothing about that nice oil cooler of 14 m:confused: it could be some problem there....

....what about regulation of oil flow within pipes from 5" OIL NON PUMPED MANIFOLD i.e. 2 pipes x 3/4" and 1 pipe of 3"....this 3" pipe connection is quite big...for oil injection between rotors (if I am right)...

...if you have too much ammonia within oil it must came from somewhere....what about connection of compressors to the plant....suction...economizer/s some P&I scheme...

seems we ask much more questions then give answers;)

Best regards, Josip :)

Mark VI
04-02-2008, 11:17 PM
Hi Josip,

It is very difficult to tell everything you see & experience at the plant, but what you are telling reminds me about one experience I had two weeks ago:

I was checking the oil cooling of two compressor units, one "working" and one "not working" (low oil pressure).

I discovered that in the sick one the oil cooling was quite better and my surprise was big when I saw that one of two fans of this perfect oil cooler were stopped and in the other both fans were running!!!!:eek:

Then I thought: It must be liquid ammonia with the oil, is clear!!. The best idea is not to allow it to reach the oil separator.

My first step was to close the economiser, as you suggest now, just in order to avoid one possible problem. But nothing changed.:(

I checked the suction and I didn't feel any liquid there. Also with a discharge superheating of 40º C it is difficult to have liquid there. This made me forgot about the idea of liquid ammonia mixing with the oil...

The explanation I give now to the "extra cooling" of the oil is that because of the bubble inside the oil cooler, the oil flow is reduced and therefore the temperature difference is larger than in the other, regardless of having only one fan working.

I will try to find a drawing of the oil cooling coils and send you. I have never worked with such a big oil cooler.

Best regards,

Mark VI

US Iceman
04-02-2008, 11:54 PM
One thing you might check is the full load pressure drop of the oil cooler you are having problems with and compare it to the other oil coolers which do not have problems.

This information should be available in the heat exchanger data sheet.

The newer oil coolers may have been sized for the proper heat rejection and oil flow rates (same duty as the previous oil coolers), but the coil circuiting may be different which causes a higher pressure loss.



When the oil is cool the 3-way temperature control valve would be closed so no flow should occur through the oil cooler. However, when the oil is hot the 3-way temperature control valve would close the bottom port off, which would force the pump to suck oil through the oil cooler (from the separator sump). This would of course cool off the oil prior to it being sucked into the pump.


This was from one of my earlier posts. Is the oil pressure OK when the oil is still cool? And then, when the oil begins to warm up, does the problem mysteriously appear?

RANGER1
16-02-2008, 09:17 AM
Has the oil injection line got a non return valve,maybe in some circumstances like on start up the oil pump suction can favour sucking through oil injection causing cavitation?On start up close oil injection and see if oil pressure is more stable. Also take out oil pump suction strainer as i imagine strainer is upstream of pump and oil injection line.Ive noticed new howdens also have new arrangement for oil injection as it is not through slide valve but in rotor housing like mycom "V" series

CrimsonK
19-02-2008, 11:40 PM
Hi mark VI,
We have almost same problem with vmy 346 compressor. the compressor have shaft mounted oil pump and oil cooler and filters are at the suction side . the total pressure drop from the separotr to pump suction must not be more than 1 bar.

cheema85
12-05-2008, 07:38 AM
Hi all,

We are also using this package. Compressor is WRVih321 and refregerent is propnae. I m also facing a low differential oil pressure problem. I have already done all these activities, but didnt get better result. Have u guys conclude any thing. If u guys came to any result pls share with me on this forum and as well as on my e-mail adress. My e-mail adress is cheema@pakoil.com.pk

SESHowdenHolima
14-06-2009, 11:14 AM
Mark VI,

Is your problem solved. If you still need some assistance, please contact me.
Bas (of ses)