Heavy Mist/oil leaking from GE frame 9 mist Eliminator


Thread Starter


I need help on GE frame 9E gas turbine mist eliminator leakage.

One of our unit is producing too much mist which then condensate back to oil in the accessory compartment and the outside the vent. It looks like a leakage somewhere but I have not been able to trace the place. but the unit indicates that lube oil is reducing.

Anybody with similar experience.Please kindly assist with your knowledge.
Hello there,

The job of mist eliminator is to maintain a slight negative pressure on the lube oil system and capture any vapour before it is vented to atmosphere.

Now, it is not clear from your post that whether you have changed this negative pressure value to achieve process flexibility or not. Now as you are saying that heavy mist is formed then that can be addressed by reducing the vacuum pressure of operation slightly. But, this will reduce the life of filters used in the vent line. Basically, even after this much troubleshooting your problem is not solved then I know that your OEM will advice you to increase the size of the vent piping and shall install a new orifice plate in LO piping to reduce the flow to design level.

But I do believe that by installing a new orifice plate you can achieve what you want. This operation shall require Mechanical and Instrument Engineer.

Similar problem was troubleshooted in a Siemens GT at one of my friends Power plant site.

Hope this will help you.
fluid flow

When a situation like this comes up on a unit which has been running properly in the past, the first thing one should always do is ask:

What has changed? When did this problem start?

If the mist eliminator has been working properly for some time, and then it stops working properly something has changed. The task is to find out what.

There are two basic types of mist eliminators: electrostatic and centrifugal fan-driven. Most of the electrostatic precipitator types have been replaced over the years, and the original poster did not tell us what type of mist eliminator is in use--which would be most helpful to understand. My explanation below refers to the centrifugal motor-driven type.

Yes, the mist eliminator's job is to maintain a slight negative pressure (vacuum) on the L.O. reservoir and that is usually set when the turbine is operating at Base Load. When the unit is operating at Part Load or FSNL, usually the vacuum is higher than it is at Base Load. So, it's important to understand that the vacuum maintained on the L.O. reservoir is not constant, but varies with load (and cooling and sealing air pressure). At least on GE-design heavy duty gas turbines.

Most GE-design heavy duty gas turbines DO NOT have a gauge on the L.O. reservoir to monitor the pressure/vacuum on the reservoir.... Which makes it difficult to know if the butterfly valve is working properly. Does the site have a pressure/vacuum gauge on the L.O. reservoir? If so, what does the gauge read at FSNL, 50% load, Base Load?

Next, the vacuum is usually set using a manually-positioned butterfly valve, either on the suction to the centrifugal fan of the mist eliminator, or on the discharge of the centrifugal fan.

Next, and very importantly, there is a connection between the bottom of the mist eliminator vessel (which contains filter elements through which oil vapours are drawn by the centrifugal fan) and the L.O. reservoir which is called the "loop-seal drain." Usually there is a sight glass in the loop-seal drain, and quite often there is NOT an isolation valve in the loop-seal drain piping. The loop seal drain MUST be full of oil at all times; if not, then problems can occur. AND, the flow between the mist eliminator pressure vessel and the L.O. reservoir must be uninterrupted and free-flowing.

Lastly, the filters and internals of the mist eliminator do require periodic inspection, maintenance (cleaning, usually) and replacement.

An excessive vacuum on the L.O. reservoir can cause vapours drawn through the vessel to not condense and to be visible at the discharge of the centrifugal fan. I have seen orifices in the cooling and sealing air lines to bearing housings not installed after maintenance outages causing excessive pressure on the L.O. reservoir which reduces the vacuum on the L.O. reservoir and pushes heavy vapours out of the centrifugal fan discharge, which can condense and make an awful mess--especially if they are drawn into compartments by compartment ventilation fans.

But, a common issue is that the loop seal drain line gets blocked, or if there is a manual isolation valve in the line that it gets mistakenly closed while the unit is operating. If a manual isolation valve is present in the loop seal drain line (or if there is more than valve) they should be locked in the open position at all times, except during maintenance outages when work is to be performed on the mist eliminator and/or the L.O. reservoir.

I have also seen a loose coupling between the mist eliminator motor and centrifugal fan make it seem like the fan is working when it's really not. A quick check for this problem is to use a clamp-on ammeter to see if the motor is drawing near rated current (based on the motor nameplate) when the turbine is running. The butterfly valve position does affect the current drawn by the motor, but in general, if the fan is moving air the current draw will be near rated, and if it's not (because the coupling is loose) the current draw will be low, very low.

I've also seen the butterfly valve "flapper" separate from the operator shaft/handle and open, causing similar problems.

Ruptured filter elements, or improperly positioned filter elements, or improperly rated filter elements can also cause similar problems.

So, there's lots of things to check. And some of them need to be checked while the unit is not running. It's pretty certain that the mist eliminator vessel is pretty full of oil, which it should not be--suggesting some kind of problem with the loop seal drain. It could be there is too much oil in the vessel to flow through the small piping back to the L.O. reservoir, or, again, there is some obstruction or an isolation valve is closed when it should not be. However, an excessive level is not always indicative of loop seal drain problems. There may be more than one problem.

But, again, the first thing one should do when a turbine has been running without problems and one arises is to ask: What has changed since this problem started? Were the filter elements recently replaced? Then the installation and orientation should be checked. Review manual isolation valves; look at the vendor's documentation in the Service Manuals; inspect the obvious things.

High flow through an orifice is usually what causes wear that increases the diameter of the orifice, or wears the sharp edges of the orifice. But, there isn't a lot of flow through this orifice, and it's mostly vapours, so the orifice size shouldn't be in questions--again if it was working fine previously. Was the orifice not re-installed after a maintenance outage? Was the wrong orifice installed after the maintenance outage?

These are the kinds of questions one has to ask when problems like this start. Especially if the started after a recent maintenance outage.

Hopefully the original poster will write back with answers, progress and findings.
Hello there,

That is a wonderful insight which you have provided. I referred back to P&ID and understood what you intended to convey.

While referring to P&ID, a question struck my mind. is it possible that if this mist is carried in the compressor then owing to compression it may iginite?

Hoping to hear from you soon.


That would depend on lots of factors. There is usually a temperature drop at the bell mouth/IGVs. And, I don't know what the flashpoint of oil vapors is for turbine lube oil at CPD.

I think the oil will collect on the axial compressor, making it extremely "dirty" and decreasing the unit efficiency. It could create quite a difficult mess to clean up.

But, yours is a good question. And it would be rather difficult to determine if vapors would ignite.
Thanks for your incisive answer to my concern. I was able to open the mist eliminator compartment and found that the cartridges looked clogged and dirty. I would have uploaded a pi but I can not see a link to do so.

I will check the other things mentioned today. You talked about cleaning the cartridges. Though GE manual do not mention that, we do use condensate from natural gas to wash the filters but it seems it doesn't reduce the High Dp of the unit. is there a better fluid to use to clean from your experience?

In adjusting the butterfly valve in the mist eliminator compartment, I realized that it is done once and when the unit was not at baseload. I have asked that it should be regulated now when the unit is at Baseload.

I am changing the filters today and carryout other checks as outlined and keep all posted. Once again thanks to all of you for your help.

Thanks for the feedback! Unfortunately, pictures can't be posted to control.com. Others have used free webhosting sites like tinypic.com to upload pictures and files to and then post the link to the file for people to access and see the information. It's free, but it does require some time to set up and upload and post the link. But, I think most people here would be extremely happy to see some pic's of the inside of the vessel and the filters. I know I would!

As for cleaning the filters, I don't recall what they're made of, and one has to be careful about the solvent/cleaning fluid not deteriorating the filter material. That would be a novel use for natural gas liquids, though! I've never seen liquids which didn't vaporize at ambient temperature/pressure or didn't contain lube oil (from leaky natural gas compressor seals), or, worse, gasoline (yes--gasoline in natural gas piping!). I'm used to the elements being replaced, but I know there can be a long lead time in some cases, and finding a supplier in some parts of the world can be challenging, too (because prices from the OEMs can be staggering).

Again, if you could find the time to upload some pictures and then post the link to them that would be really appreciated by many! That's one of the best things about forums like this is that many can benefit from the problem(s) at a single site; the ability to post pictures would be great, but Web security and hacking being what it is these days we all have to understand the risk it would pose to this forum--which we're all very grateful to have.

But the feedback is the best part of the threads on this forum. We like to say, "Feedback is the most important contribution!"(c) here at control.com. It's what let's others reading this thread now--and later--know what worked and what didn't. I monitor several related sites on similar and different topics and I can say that the posters on control.com are the BEST when it comes to feedback. Most people just ask for help, don't provide a lot of information to make the responses easy or concise, but they do--for the most part--reply with their results or findings. Which is both gratifying for those responding to their requests and informative to those trying to learn from others experiences and issues. So, thanks--and we look forward to hearing how this is resolved.

Good luck--and please do keep us posted!