Control.com - Nerds in Control

Experiencing high vibration on bearing 1 and 2 (up to 0.99 inch per second) at about 95% full speed to fsnl. Alsthom unit of MS6001B [frame 6]. This always caused us to shut-down before trip [set-point of 1.0 inch per second].

We could do with a lot more information, when did this start? Have you been doing any work on the machine? First thought, are the Bleed Valves closing, are the IGVs opening?

The Bleed valves were still opens, while the IGV at it's point of closing. The vibration grows speedily from say 0.30 IPS to 0.99IPS; that is, tending towards 1.0IPS and greater, as the IGV closes.

Yes, the unit newly came back from turbine and load-gear balancing. Thanks

I guess you meant IGV opens (to minimum full speed angle) rather than closes. Since you mentioned that the machine has come back from turbine and load gear balancing, it could most probably be a problem with shaft centering/balancing.

What are vibration levels in load gear and what are bearing metal temperatures of turbine bearings? Are they in normal range?

If the shaft was removed and sent away, was it re-aligned to the Accessory Gear and Load Gear when it was re-installed? (Believe it or not, I've seen Site Managers who claimed re-alignment after re-installation of the turbine/compressor shaft was not necessary, and blamed vibration problems on the Speedtronic. Believe it or not, after a couple of weeks, and an alignment first to the Acc. Gear and then to the Load Gear, the vibrations miraculously disappeared. And no action was taken on the Speedtronic or the vibe pick-ups. But, the Site Manager still claimed the problem was the Speedtronic!).

Some shops can perform a low-speed balance, but not all of them can guarantee the unit will not have vibration issues near or at rated speed after a low-speed balance.

If bearings were replaced, were they properly replaced? Sizes checked and fitted as necessary?

Unless someone changed the vibration pickups or the scaling of the vibration pick-up inputs to the turbine control system, the problem is most likely not the turbine control or vibration monitoring system. If the unit was disassembled and the shaft removed and re-installed, that's probably the cause of the vibration problems.

Has anyone stood out around the machine when it's starting to see if the vibrations are indeed high? At 1.00 ips, they should be discernible to anyone in the vicinity, especially anyone standing on the grating next to the turbine/acc/load gear compartments.

Is there a Bently-Nevada monitor in use on the unit to confirm high vibrations?

It's presumed the #1 and #2 bearings have redundant vibration sensors; are they both reading similar levels of vibration?

Has anyone checked the vibration pick-up hold-down bolts?

When you write to tell us you are having a problem, we need to know what you've done to try to troubleshoot the problem and what the results of the troubleshooting were.

I'd suggest getting someone to site with vibration/balancing experience and let them have a go at solving the problem. That is if you can't get the party that balanced the equipment to assist, or can't get the party that disassembled the unit and put it back together to assist. This is most likely--from the information provided NOT a controls-related issue.

Best of luck with your problem.

Thanks a lot. I really appreciate all your effort. In actual sense the shaft was removed and sent away, and was not properly re-aligned to the Accessory Gear and Load Gear when it was re-installed.

We actually called for an expertriate on vibration/balancing to our site and an alignment between the Acc. Gear and the Load Gear were conducted. Since then no more high vibrations till you excite and synchronize the unit. No action was taken on the Speedtronic or the vibe pick-ups.

Thanks for your big concern.

> Experiencing high vibration on bearing 1 and 2 (up to 0.99 inch per second) at about 95% full speed to fsnl.

---- snip ----

Have you boroscoped for potential "angel wing" rubbing on your last stage blading?

I have issue with GE Frame 6 MS6001B, same unit i had problem of high vibration with some years back (2 years ago).

The unit was started at 2018H and was synched at 2047H. The unit later tripped at 2107H due to high exhaust temperature with several thermocouples rejected.The unit was about attaining 25MW when it tripped. A loud bang and flame from the exhaust was noticed. Trend analsyis was done and it was noticed that there was no vibration during or after the trip.

We agreed that the unit be started later and left on SPEED CONTROL at 20MW to see the performance of the unit.

The unit was restarted at 2252H and was synched at 2308H.The unit was on 20MW target and PCD was at 125.6psi.

The unit was online for an hour when it had a loud bang and flame emanating from the exhaust.We quickly shutdown the unit. Unit is on RATCHET.

We now checked round the barge to see what was causing this problem.We drained condensate at the condensate area. And noticed CONDENSATE at the STRAINER end of the gas line on the Unit gas line header in to the unit. There was no condensate at the turbine compartment drain plug.

Oluwafemi Ogundaisi,

I don't understand "with several thermocouples rejected." What control system is being used for this unit? What do you mean by "rejected thermocouples"? There were older GE control systems (Mark II and earlier) that used a method of rejecting thermocouples that I've seen duplicated on PLC-based turbine control systems that weren't exactly duplicated properly. Further, even when a Speedtronic control system's thermocouple rejection system is used it can't be indiscriminately used on adjacent thermocouples. Please explain more about this 'thermocouple rejection' statement.

If you are experiencing loud "bangs" and flames coming from the turbine exhaust, something is definitely not right.

When you say you find condensate at the condensate area, what do you mean by "condensate area"?

When you say your found condensate at the [fuel gas] strainer, that could be hydrocarbon liquids which are condensing or have condensed in the gas fuel supply line. Hydrocarbon liquids making their way into the combustors of a GE-design heavy duty gas turbine can cause high load spikes and exhaust overtemperature trips. When you examine the data for these events, you say you aren't seeing any evidence of vibration, but do you see any evidence of load (MW) spikes immediately prior to the bang/flames?

You should consider checking the dew point temperature of the gas fuel supply and see if the problem is related to hydrocarbon liquids being blown into the combustors and causing high exhaust temperatures and load spikes.

Back to the "rejected thermocouple" question, if you have "rejected" adjacent thermocouples and if there is some loss of flame in one or more combustors because of some non-hydrocarbon liquids (condensate) momentarily extinguishing the flame and then if the fuel suddenly reignites there could be problems. This would be most likely evidenced by load drops (decreases) because of the loss of flame and then load spikes when the fuel is reignited.

There can also be entrained water moisture in gas fuel which can condense if the dew point temperature is reached. "Slugs" of water can momentarily extinguish flame, and if multiple thermocouples (particularly adjacent thermocouples) are "rejected" from the combustion monitor protection then something like what you are describing could happen. If the combustion monitor doesn't sense a cold spot caused by loss of flame in a combustor without a flame detector (not every combustor has flame detectors) then the unit will not be tripped. It's not normal for flame to be lost in a combustor and then reignited when the unit is running at rated speed and on load (the airflows past the cross-fire tubes are very high and can prevent cross-firing), it has been known to happen.

But something is not right. Loud bangs and flames coming from a gas turbine exhaust should be cause for concern and investigation.

"With several thermocouples rejected." -Means more than 4 exhaust thermocouple values or readings were rejected.

What control system is being used for this unit? - PLC-based turbine control systems - GE control systems (Mark IV)

What do you mean by "rejected thermocouples"? - Their values were too high for the control system to "calculate" with {Please explain more about this 'thermocouple rejection' statement}

When you say you find condensate at the condensate area, what do you mean by "condensate area"? –Means that LNG scrubber or gas/Liquid separator area

When you examine the data for these events, you say you aren't seeing any evidence of vibration, but do you see any evidence of load (MW) spikes immediately prior to the bang/flames? _Yes, higher than selected target

After further inspection and investigation, no damage found on the turbine blades. And the bleed valve integrity was confirmed okay.

Unit restarted, this morning: At about 23MW, the unit goes on temperature control even with higher targets selected. Also the PCD seems to dip seriously just before trip. All these happened at reduced grid frequency (because we are tied to the national grid). Once the frequency drops, the PCD takes a dip, accompanied by a drastic drop in load and then comes the pulsation (with the air inlet filter house vibrating seriously). Then the unit trips on "High Exhaust Temperature".

Meanwhile, our IGV is not the variable type (it is either fully open-87o or closed-34o) and the control system cannot 'tell' when the IGV is partially open.

What is your advice please?

Well, this is very interesting. I'm looking at a Mark IV+ Speedtronic Elementary for a Frame 6B unit, and it only "rejects" thermocouples less than a certain value--and only for calculating the average exhaust temperature. I remember that some early Mark IVs used on some Frame 5s did not have a combustion monitor function at all, but I never recall seeing a Frame 6B equipped with a Mark IV (or Mark IV+) that did not have a combustion monitor function.

It seems you're using LNG for the gas fuel, and while I've never worked on a unit running on LNG I would imagine the liquified natural gas has to be "expanded" in some manner (through some kind of nozzle or by using some heat source) to convert the liquids to vapor for burning. If you are finding condensate in the gas fuel system then it would seem the LNG is NOT being properly vaporized OR it is condensing as it passes through the system. Pressure drops across various elements (strainers, valves, nozzles) cause temperature drops which can cause condensation.

And if natural gas liquids are being blown into the combustor then you will see load spikes (increases).

Sudden drops in compressor discharge pressure (usually called CPD in most Mark IV systems) are indicative of loss of flame in multiple combustors. There is a pressure increase when fuel is burned in a combustor and when the flame is suddenly extinguished the pressure in the combustor--and axial compressor discharge pressure--will suddenly decrease. The more combustors that lose flame, the more pronounced the CPD decrease will be.

If the grid frequency is swinging wildly then axial compressor discharge pressure will also swing. As grid frequency increases, CPD will increase; as grid frequency decreases, CPD will decrease.

It would seem you have a lot of things going on at once. Based on the information provided I would say there is something amiss with the LNG system that's causing there to be insufficient superheat of the fuel being sent to turbine. (GE usually requires (strongly recommends) at least 50 deg F of superheat for gas fuel. This means the temperature of the gas fuel should be at least 50 deg F above the dew point of the gas fuel--this to prevent condensation of liquids.)

I don't know where you're getting your information about "rejected thermocouples" from, but I've never seen a Mark IV reject any thermocouple value other than those less than a certain value (usually 500 deg F when the unit is running). This is done to prevent failing or failed thermocouples from adversely affecting the average exhaust temperature and causing excessive fuel to be put into the unit. (In Speedtronic systems, open thermocouple circuits (and most thermocouples fail open circuit) read low, sometimes negative.) And this is just for calculating the average exhaust temperature (TTXC or TTXM).

When it comes to calculating exhaust temperature spreads in Mark IV, no exhaust thermocouple values are rejected (unless there is a communication problem between <C> and one of the control processors).

So, I'm pretty confused about this whole thermocouple rejection business.

I think it's unsafe to continue to run the unit until you find out what's causing the loud bangs and flames coming out of the exhaust. If the flame is extinguished in one or more combustors when the unit is loaded and if the unit is not tripping on loss of flame (because every combustor does not have a flame detector and it doesn't sound like the combustion monitor is working correctly for some reason) then unburnt fuel will flow into the combustor, through the turbine and into the exhaust and once the concentration gets high enough it will burn/explode. This is very unsafe.

Dear CSA....MVP,

Point of correction (few errors) please, we are actually using natural gas not LNG (not in liquid form but pure natural gas). Lets leave out the issue of thermocouple rejection. ALso, our PLC control system is not Mark IV but Speedtronic systems. Frame 6B (MS6001B)

Base on this please advise.

Thanks

You may be using natural gas, but the same can still occur if the temperature is at or only slightly above the dew point temperature. Whatever those liquids are (and it would be good to have them tested to see what they are--just plain water, or what--I've seen lube oil (from compressors), gasoline (yes!), and water entrained in gas fuel supply) I would think from the information provided they are at least part of the problem.

I was wondering about the PLC Mark IV business. While programmable, Mark IVs are not really PLCs. And, the only time I've seen thermocouple rejection other than on Mark II units was on a PLC that was being used for turbine control that replaced Mark II controls.

I think I've provided as much help as I can based on the information provided.

Please write back to let us know how you resolve the problem(s).

The greatest challenge of all is that the unit can generate 31.0MW NDC. but when we run it @ 20MW selected target or below on speed control, it doesn't experience this pulsation. But pulsation occurs any moment you increase the load target above this. The Unit then act as if there is air starvation with loud bang and thus respond quickly when the grid frequency drops lower with sudden CDP and then trips or call for immediate shutdown. Please can you advise further.

The unit was raised to 23mw load target during which the pcd value raise to 127.8psig with the highest thermocouple at 1014 0F. A further increase to 26mw load target after 10minutes saw the unit making 25MW, but this quickly dropped after a few minutes to 22.9MW at lower frequency and lingered on for some time before the loud bang on the unit that made us shutdown the unit @ 1113H. We actually noticed and saw the air duct line to compressor vibrates with the loud bang. Unit is ratcheting while investigation is ongoing.

This must be a very old Frame 6, as I've never seen one with un-modulated IGVs. I would have to review the Piping Schematics (P&IDs) and the sequencing in the control system to understand when the IGVs move from closed to open, and if there was any intermediate position.

From this latest information it would certainly seem there is some problem with the axial compressor and/or the IGVs. I've only experienced compressor surge/stall once, and it was only when the compressor bleed valves didn't open during a shutdown. I can only imagine what a surge/stall condition would sound like when the unit was actually running at rated speed with significant load.

I would suggest you have someone come to site familiar with the unit to observe the problem and provide assistance.

If it is a GE frame 6, the IGV's are modulated. If it is simple cycle, they may effectively be 2-position - full open or full closed (or maybe 3-position, 34 degrees, 57 degrees or 84 degrees).

the IGV's it's is a simple cycle, they are effectively be 2-position - full open or full closed (34 degrees, or 83.7degrees).

(I have this feeling I'm going to regret this.)

If you have a GE Speedtronic turbine control system on a Frame 6 it's mostly likely a Mark II or a Mark IV. If you have a GE Speedtronic turbine control system then the name of the solenoid that controls the IGV position (for a two-position "bang-bang" (no pun intended) IGV system) is most likely 20TV-1. I would presume 20TV-1 is to be energized when the IGVs are to be in the open position, and de-energized when they are to be in the closed position.

When the unit is operating at approximately 20-25 MW, what is the IGV angle?

Can you determine from the Speedtronic elementary drawings (again, presuming the turbine control panel is a GE Speedtronic) when 20TV-1 is to be energized?

This may be a little complicated because at one time GE used servos (yes, servos) for solenoids and applied 125 VDC to the servo coil, reversing polarity to change the direction of flow through the servo, which is akin to opening/closing a solenoid-operated valve. So, the servos were always energized, but the polarity of the applied voltage changed with the intended operation of the device it was controlling the position of.

The questions I'm trying to get answers to are:

1) When should the IGVs be open and when should they be closed?

2) When the unit is operating at about the load it is experiencing the bangs and flames, what position are the IGVs at?

In other words, is it possible the IGVs are either not open when they should be, or are open when they shouldn't be? I'm asking this because from the information provided it seems there may be a compressor surge/stall problem, but it's difficult to know for sure without being on site.

The only way to know the answers to the above questions for sure is to examine the Speedtronic elementary drawing to determine when the IGVs are to be open and closed, and then to determine if they are open when they should be and closed when they should be.

I still think you should have someone come to site (barge) and help with resolving this issue instead of continuing to start and load the unit and experience the bangs/flames. At some point, something in that machine is going to break, and because it's rotating at approximately 5100 RPM when something breaks (like a compressor blade or a turbine bucket) there's going to be a significant amount of damage, and even more lost revenue whilst trying to repair the damage at great cost.

When the unit is operating at approximately 20-25 MW, what is the IGV angle? – 840

Can you determine from the Speedtronic elementary drawings (again, presuming the turbine control panel is a GE Speedtronic) when 20TV-1 is to be energized? – No

We are not on Servo – 20TV

The questions I'm trying to get answers to are:

1) When should the IGVs be open and when should they be closed? – @ 95% Turbine Speed it should be open while on shutdown or trip it should be closes (though faster at trip than at normal shutdown)

2) When the unit is operating at about the load it is experiencing the bangs and flames, what position are the IGVs at? - – Cant even tell because we weren't there to confirm physically and unfortunately the limit switch in control cant sense level/position because it is preset to sense signal @ fully open or fully closed (–– 340 and 840 respectively)

Oluwafemi Ogundaisi,

It is virtually impossible for me to be of much more help to you, or to Solomon Osayi (who has also opened a thread for this same site), because we can't see the Piping & Instrumentation Diagrams (P&IDs, or Piping Schematics) for the turbine, and we can't see the sequencing that is running on the turbine control system. From the information provided and without being able to examine the documentation and/or to witness the problem it is going to be very hard to help with your problem.

As otised has said, it would be very unusual for a GE-design Frame 6 not to have modulated IGVs--not impossible, but very unlikely. Even if the IGVs are NOT modulated, there has to be an actuator of some sort (hydraulic) that has to have high-pressure hydraulic oil ported to it to open or close the IGVs (or it could be an electric actuator, but that wouldn't likely be original equipment). So, we would need to know what that device is that is porting the high-pressure hydraulic oil to the actuator to open and close the IGVs, and then to understand if it is working correctly.

You seem to have discounted the possibility that there are liquids of some sort being sent to the fuel nozzles/combustors, even though you have said you found condensate in the condensate system of the gas fuel system. Okay; that's fine.

The conditions you are describing seem to be slightly contradictory--to me, anyway. For there to be loud bangs and explosions in the exhaust, it would seem that there would need to be some accumulation of unburned fuel that was suddenly being ignited. This would seem to mean that flame was lost in one or more combustors without flame detection (unless the loss of flame detection was being "bypassed") allowing unburned fuel to flow into the exhaust where it was subsequently ignited causing the loud bangs and flames.

If slugs of gas fuel liquids were making their way into the combustors then load would spike very high, the exhaust temperature would also spike, the compressor discharge pressure would also spike while the liquids were burning. When the liquids had finished burning the load would reduce, probably below the previous steady-state level, and compressor discharge pressure would also probably decrease below previous level. These kinds of conditions are very difficult to trace without high-speed trending capability to learn what's occurring first. But, again, you seem to be discounting the possibility of liquids.

When flame is lost compressor discharge pressure will decrease; the more combustors that lose flame, the higher and more sudden the decrease in compressor discharge pressure.

Not ever having personally experienced an axial compressor stall/surge condition at rated speed under load, I don't know what would happen to compressor discharge pressure--though I would imagine it would decrease rather quickly (the "stall") and then pulse (the "surge") until the unit tripped, possibly on high vibration or exhaust overtemperature or something similar. But you say you aren't experiencing any high vibrations, so that would seem to rule out any axial compressor surge/stall conditions. The only surge/stall condition I witnessed caused vibration readings to over-range the input scaling, and the ground shook, too. The turbine had already tripped and was coasting down from rated speed, and fortunately there was no damage to the turbine.

So, I'm completely at a loss to explain what's happening based on the information provided. Sorry; but nothing adds up with the descriptions. Either there's something wrong with the air flow through the compressor, or there's something causing an accumulation of unburned fuel in the exhaust. An axial compressor problem would most likely cause high vibration; I would expect at least one of the combustors with flame detectors would be indicating loss of flame.

You say you've "rejected" exhaust T/Cs, so it's difficult to say what the actual exhaust temperature spreads are. You also said the rejected thermocouple readings were abnormally high, which is kind of unusual for a turbine running on gas fuel because unless there are fuel nozzle problems (excessively worn orifices or leaks) the amount of fuel going into each combustor should be relatively the same (within, say +/-10% even for a poorly matched set of fuel nozzles).

The loss of compressor discharge pressure and exhaust overtemperature are also somewhat contradictory--based on the information provided. Of course, if fuel continued to flow when compressor discharge pressure was low and then compressor discharge pressure suddenly increased, that might account for the exhaust overtemperature.

I suppose it's possible that a loss of speed signal feedback to the turbine control system could cause a momentary spike in fuel flow command and fuel flow-rate. But, why would that only be occurring at one load at what should be a relatively stable speed? Although you have said the national grid frequency--which directly affects your turbine-generator speed--is not very stable when these events are occurring. Could be a poorly tuned control system, but, why would these problems start occurring now? What has changed recently? Some "tuning" of the control system?

I don't know if we've asked, but are you sure the compressor bleed valves are fully closed? That shouldn't cause the kind of problem you are seeing, but, in conjunction with other conditions it may be aggravating the problem(s).

The only other question I can think to ask is: When did this problem start? After some maintenance outage? After fuel nozzles were replaced? After another emergency trip condition?

Sorry; I'm at a loss here. Hopefully you'll write back to let us know what you ultimately find the cause(s) to be, or someone else here may have another idea.

> ALso, our PLC control system is not Mark IV but Speedtronic systems. Frame 6B (MS6001B)

All MS6001B GE gas turbines use a GE Speedtronic control system. It may be Mark IV, Mark V, Mark VI or Mark VIe, or MAYBE Mark II. The first MS6001 was built before the controls shifted from Mark II to Mark IV, so if you have an old MS6001B (say pre 1984) you may have a Mark II Speedtronic control. A Mark II panel would have a bunch of analog indicators, switches and lights and probably an ITS exhaust temperature control system. If it did NOT have the ITS system, it probably had the exhaust temperature "swamping box" which did permit the operator to reject individual exhaust thermocouples.

If there is a monochrome CRT with membrane switches on the front of the control panel, it is a Mark IV Speedtronic control. Mark V and later control have a separate PC for the operator interface.

I'm in agreement with you, otised. All Frame 6s should have modulated IGVs. I've had people tell me that IGVs were two-position because at closed they were at 34 DGA, and at maximum they were at 84 DGA, but, in fact they were modulated to intermediate positions during normal operation. (They were just reading from a Control Specification, and not reading the entire section.)

I have a feeling this unit was relocated to a barge from its original site, and in the process may have gotten a new control system. I do recall there was a company producing PLC-based turbine controls called Speedtronics or Speedotronic.

Again, it's still a credit to GE that some of these turbines survive as long as they do. Some turbines would never have survived this kind of ... operation. And, this one keeps starting (though it can't seem to be loaded past a particular point at this time).

CSA_MVP,

Investigation completed, the compressor was lifted and the consequential effect of the under frequency and bang noise, lead to bent and broken blades mostly on stage four and five. This contributed to the reason why the unit cannot be load above 20MW target. We are going to renew the rotor and get back online ASAP.

Thanks all for your effective contribution on this forum.

You guys are the best.

Oluwafemi Ogundaisi,

Thank you very much for the feedback. It's unfortunate about the damage; best of luck for a speedy return to generation!