Today is...
Monday, June 18, 2018
Welcome to Control.com, the global online
community of automation professionals.
Featured Video...
Featured Video
A tutorial introduction to programming using the QuickBuilder Programming Environment.
Our Advertisers
Help keep our servers running...
Patronize our advertisers!
Visit our Post Archive
IGV Control Trouble Trip
A gas turbine MS 5001PG (MKV) tripped on the alarm "INLET GUIDE VANE CONTROL TROUBLE TRIP".

Greetings all,

We have a gas turbine MS 5001PG (Mark 5).

During normal operation, it was loaded with 10MW.

Firstly, a diagnostic alarm:


P DROP DESCRIPTION
S 1340 TCQA LVDT Position Diff. High Reg#5
R 1340 TCQA LVDT Position Diff. High Reg#5
T 1340 TCQA LVDT Position Diff. High Reg#5

Then the turbine tripped on the alarm:

P DROP DESCRIPTION
Q 0141 INLET GUIDE VANE CONTROL TROUBLE TRIP

We did the following:

- Check the solenoid coils of servo valve by measuring the resistance of each one. All 1 kilo ohms: OK

- Check the IGV filter: OK

- Check the excitation and the feedback of the LVDT: found them at the range of 0.7V @ 42deg (Min. angle of IGV) - 3.5V @ 85deg (Max. angle of IGV)

- Tried to manually stroke for IGV from 42deg -> 85deg, its feedback at the control was OK and the IGV was at the position (We physically checked it).

The turbine was ready to start, so we started it up again. It ran for about 30min at 4.5MW load. IGV was at the required position (56deg) and the actual position was:


<R> core -> 55.9
<S> core -> 55.59
<T> core -> 55.91

After 30min of normal operation, the same diagnostic alarms occurred again:

P DROP DESCRIPTION
S 1340 TCQA LVDT Position Diff. High Reg#5
R 1340 TCQA LVDT Position Diff. High Reg#5
T 1340 TCQA LVDT Position Diff. High Reg#5

and the turbine tripped after 10min on the alarm:


P DROP DESCRIPTION
Q 0141 INLET GUIDE VANE CONTROL TROUBLE TRIP

What is the cause of the trip?

2 out of 2 members thought this post was helpful...

Mohamedibr752,

The IGV Control Trouble Trip occurs when the difference between the actual IGV angle (CSGV) differs from the IGV reference angle (CSRGV) by more than a predetermined amount (approximately 5 DGA (DeGrees Angle)--the actual difference for any machine must be checked and confirmed). So, either the IGVs are opening more than the reference OR the IGVs are closing more than the reference OR there is some problem with the feedback OR the IGVs and/or the IGV actuator are physically sticking/binding and will not open (or close) when being commanded to.

Based on the information provided, it would seem there is some problem with one or both of the LVDTs that is causing the Mark V to believe the actual IGV angle differs from the IGV reference angle at some point during the operation of the unit. Sometimes, when LVDTs get worn by the movable core rubbing against the stationary armature at some midpoint of stroke the can occur. Just measuring the LVDT feedback voltage at min and max stroke won't show this problem--and the voltages you provided are very suspicious as most LVDTs that produce exactly 0.7 VAC RMS at min stroke don't also produce exactly 3.5 VAC RMS at max stroke (that exact range is the maximum allowable range for any typical LVDT used on a GE-design heavy duty gas turbine, but IS NOT the commonly found actual range).

If the LVDT feedback was calibrated using AutoCalibrate, and the control processors haven't been re-booted since the AutoCalibrate was performed, then you can use 'Verify Position ' and 'Verify Current ' of AutoCalibrate to check the LVDT feedback and the servo-valve current, respectively, as the the IGVs are being automatically stroked over their entire range of travel. If while using 'Verify Position' the plots of LVDT positions have any spikes or dips that would be an indicator of a failing LVDT.

If while using 'Verify Current' there are any spikes or dips in the plots of servo-valve currents that would indicate a sticking actuator or servo-valve. There will be a "zig-zag" in the plot when the IGVs reach maximum and the current changes to start closing them--that's normal. But otherwise the current plots should be virtually flat lines under normal circumstances.

Hydraulic actuators can become very worn over time in a certain spot if the unit is operated for long periods of time over many years at a particular load or small load range and will leak ("pass") oil and be unable to hold position--particularly when the unit is running and, in the case of the IGVs, there is a lot of air flowing through them and developing forces on the IGVs and the Mar V is trying to counter those forces.

But, you need to do more, better troubleshooting and determine whether the IGVs are sticking, or the actuator is unable to maintain position, or one or both of the LVDTs is failing. You haven't provided enough information to make a more definitive statement. Unless the oil is old and dirty and the IGV servo-valve is old, it's not likely the IGV servo-valve--again unless the oil is old and hasn't been tested/maintained properly.

If the unit just recently was returned to service after a maintenance outage when work was performed on the IGVs, please tell us what work was performed on the IGVs.

Please write back with more information and we can try to help. If you resolve the issue, we'd like to know what was done to resolve the problem.

Dear CSA,

Thanks a lot for your reply. I would like to share with you some readings we took by manually stroking the IGV from the Mark 5, after the first trip.


IGV at 42deg:
<R> <S> <T>
LVDT #1 Voltage -0.843 -0.827 -0.846
LVDT #2 Voltage -0.836 -0.821 -0.823

Required Position 42.00 42.00 42.00
Actual Position 42.10 41.79 42.16
Servo Current -2.05 -3.49 -0.37


IGV at 60deg:
<R> <S> <T>
LVDT #1 Voltage -1.770 -1.752 -1.772
LVDT #2 Voltage -1.780 -1.771 -1.768

Required Position 60.00 60.00 60.00
Actual Position 60.16 59.99 59.95
Servo Current -1.56 -1.95 -2.00


IGV at 85deg:
<R> <S> <T>
LVDT #1 Voltage -3.075 -3.060 -3.080
LVDT #2 Voltage -3.059 -3.050 -3.046

Required Position 85.00 85.00 85.00
Actual Position 85.01 84.71 85.10
Servo Current -2.59 -4.00 -1.37


There is something noteworthy, at stroking the IGV from 42deg to 60deg, the following alarms appeared:

P DROP DESCRIPTION
Q 0141 INLET GUIDE VANE CONTROL TROUBLE TRIP
Q 0142 INLET GUIDE VANE CONTROL TROUBLE ALARM
Q 0143 INLET GUIDE VANE POSITION SERVO TROUBLE

The turbine was stand-by from Dec 2015 till we operated it at 3rd of August, 2016, 2 days before the trip occurred.

Can you mention how exactly to get the plots you mentioned about?

mohamedibr752,

The plots I mentioned are automatically generated when using AutoCalibrate's 'Verify Current' and 'Verify Position' features, and I believe both of these are documented in the Mark V Maintenance Manual, GEK-5980.

What's not documented in the manual is the fact that to run either of these features one has to perform an AutoCalibrate of the LVDTs first. So, if you want to run either or both of these AutoCalibrate features you will need to perform an AutoCalibrate of the IGV LVDTs. I STRONGLY suggest you use the EEPROM Downloader to upload the IOCFG partition from <R>, or <S> or <T> to a generic IOCFG_Q.DAT (for a unit with an <I>), or IOCFG_Q.AP1 (for a unit with a GE Mark V HMI (running MS-Windows and CIMPLICITY and TCI). After performing the AutoCalibrate, you can then use either or both 'Verify Position' and 'Verify Current' to produce the plots and analyze them.

Once you finished with the Verify features, you then need to re-boot all three control processors one at a time, waiting about 5-6 minutes before re-booting the second- and third control processor. Be sure the LCC/SLCC Display shows I/O Status A7 for a few minutes before re-booting the next processor (this allows all of the signals to get "caught up" with the other two processors, particularly any integrating signals). This will restore the as-found, prior LVDT calibration scaling I/O Configuration Constants so it will be as though nothing has changed. (Unless you are going to be using a machinist's protractor to measure IGV angles prior to and after the AutoCalibrate, it's best just to re-use the as-found values--unless you have good reason to suspect something is amiss, or you need to replace one or both of the LVDTs.)

Should you decide you need to replace the IGV servo-valve, it is NOT necessary to re-calibrate the LVDT feedback. Changing the servo-valve DOES NOT change the physical stroke of the IGVs or the adjustment of the LVDTs, and AutoCalibrate ONLY calibrates LVDT feedback. It does nothing to change the gains and response of the hydraulic system, contrary to popular belief.

Basically, after AutoCalibrate is complete, you can then choose 'Verify Position' and AutoCalibrate will automatically stroke the IGVs and produce the plots mentioned above. Then, when that's done, you can then stop AutoCalibrate (I forget if the button reads ABORT or STOP), and then choose 'Verify Current' and it will automatically stroke the IGVs and produce the plots mentioned above.

I note from the data you provided, that the min- and max voltages are NOT 0.7- & 3.5 VAC RMS as reported in the original post. And, I note that <T> processor seems to have a larger discrepancy in the high-selected IGV LVDT feedback value than <R> or <S> (evidenced by the position values and the servo-valve current differences). This by itself is not unusual--because one usually downloads the average of LVDT calibration Zero- and 100 voltage values and there are slight differences in analog components on the TCQC cards and TCQA cards there is usually some difference, so this is not unusual.

Based on the length of time the unit was "on standby", it's entirely possible that some varnishing of the oil in the hydraulic circuit (especially the servo-valve) may have occurred. Without disassembling the servo-valve it's almost impossible to determine for certain (and disassembling a servo-valve requires special tools and special equipment for re-adjusting the fail-safe spring tension; it's not something that can easily be done in the field).

Please write back to let us know how you fare!

Dear CSA,

We did replace the IGV servo valve with a brand new one. We started the turbine, and it did successfully start taking a load of 10MW. After around 30min, we were slowly raising the load to 20MW till suddenly the 3 alarms appeared again at 12MW.


P DROP DESCRIPTION
S 1340 TCQA LVDT Position Diff. High Reg#5
R 1340 TCQA LVDT Position Diff. High Reg#5
T 1340 TCQA LVDT Position Diff. High Reg#5

We decided to put the turbine back to 10MW and put it on REMOTE PMS mode. After few hours, we checked the turbine to find that the load fluctuates between 9.5MW and 10.3MW. These fluctuations caused the IGV to fluctuate as well. FSA was fluctuating too, so we suspected the fuel control system. We found that the main gas supply slowly fluctuates between 15bar to 19bar. We did raise the load to 20MW for few minutes and then lowered it down again to 10MW. The load was stable. Fluctuations disappeared!

What do you think is happening here?

Does the main gas supply fluctuating affect the turbine load?

We did LVDT calibration while the turbine is off, but sometimes we get these two messages:
"LV1 NOT CHANGING"
"INCORRECT SERVO-NULL BIAS"
One message at one core at a time. So, we weren't able to Verify Position or Verify Current at the 3 cores.

What do these messages actually mean?

What can we do to be able to utilize the replaced servo valves? Can we order the Filter Replacement Kit and replace the filter? Do we need special tools to be able to do that at the field?

1 out of 1 members thought this post was helpful...

Mohamedibr752,

Holy Smokes!

So, one of the things that happens when a problem like this "suddenly" appears is that a LOT of people start watching the troubleshooting and turbine operation, and there is a constant chorus of, "It's never done THAT before!" or, "What is it doing?" Most of these people have never really paid attention before, because a "good start" is one that results in generator breaker closure and load, and no matter what alarms came up during the start as long as the unit didn't trip, well, it was a good start. Right? (Wrong. Every alarm, Process or Diagnostic, is important and should be understood and resolved. It's not easy, but, that should be the process. Unfortunately, it's usually not the process.)

Gas fuel supply pressure fluctuations should not cause problems with load--unless the pressure fluctuations are very fast. The SRV (Stop-Ratio Valve) should be capable of responding to most gas fuel supply pressure fluctuations so that the GCV (Gas Control Valve) can do its thing--in this case it looks like some remote signal (don't know if it's discrete or analog) is trying to control load to some particular setpoint.

It's possible that the two systems--the GT and the gas fuel supply pressure regulator--got into some kind of resonant frequency that caused the load fluctuations.

If the GT is simple cycle, or is being operated in simple cycle mode, the IGVs will try to hold a particular exhaust temperature (900 deg F, I believe) during loading so if the exhaust temp was at or near this temperature while the unit was operating then it's possible that the load swings caused the IGV swings. (If the unit is combined cycle and was being operated in combined cycle mode, then the IGVs will also modulate to try to maximize exhaust temperature at part load, so if the actual exhaust temp was at or near the IGV exhaust temp control setpoint then that could explain small fluctuations in IGV angle/position.)

As for the two AutoCalibrate alarm messages, I've only seen the 'LV1 not changing' alarm when the IGVs moved very slowly--much slower than normal. Were the IGVs moving when that alarm was annunciated? Does the unit have an Aux. Hyd. Pump (driven by an AC motor), or do you use CRANK mode for stroking hydraulically-operated devices?

The second AutoCalibrate alarm message brings up an interesting question. What is the servo null bias current setting for the IGVs? I've been presuming it's the typical 2.67%, but, it may not be. You can find the answer to this question by going to the I/O Configurator, opening the <Q> TCQA card, and scrolling to the SVO 5 (Servo-valve Output 5) screen and noting the servo null bias current value. (Just exit the I/O Configurator without saving any changes; and all will be good.) And, again, how fast were the IGVs moving when you were trying to calibrate the LVDT feedback?

Are you certain that the servo-valve you used to replace the as-found one is the correct servo-valve for the IGV actuator? Servo-valves all are the same size from the outside, but the flow-rates through them can be very different. If one for another hydraulically-operated device (say the SRV or GCV) were used and it had a lower flow-rate (the IGV actuator is usually much larger than most other hydraulic actuators--plus it is a double-acting piston/cylinder while most others are single-acting) then the IGVs would move slower than they should.

There is also a filter (or should be) directly upstream of the IGV servo-valve manifold/actuator, called the "last-chance" filter. Have you changed the filter element of that "last-chance" filter? Many times, because there is no alarm switch to warn of high filter differential pressure, the filter dp gets so high the filter ruptures--which makes the little pop-up visual indicator worthless and people looking at it always think the filter is good--when it's worse than good. A ruptured filter will release a lot of dirt and debris into the system downstream of it, as well as allowing dirt and debris through the filter.

Changing the servo-valve "pencil" filter (so-called because it has a long and narrow cylindrical shape) requires no special tools--but it does require a clean place to do the work. However, I will say that there is a GREAT temptation to tweak other screws and lock-nuts on the outside of the servo-valve when replacing the pencil filter--and that can result in problems that can make the servo-valve useless. So, if you're going to change the pencil filter, make sure the person changing the pencil filter does so in a clean environment with clean hands and clean tools, and only replaces the pencil filter--and doesn't turn/tighten any other screw or locknut on the servo-valve. (The fail-safe spring tension is set by adjusting a screw with a lock-nut on the outside of the servo-valve; but without proper testing equipment adjusting this screw can result in making the servo-valve useless.)

What kind of operator interface does the Mark V use? An <I>, or a GE Mark V HMI (which runs MS-Windows and CIMPLICITY and TCI)? Have you ever used the VIEW tools before to capture high-speed data in a graph? You need to find out what is causing the "position diff high alarm" and to do so you need to monitor the LVDT feedback when the problem is occurring. You might be able to see a problem if you note the IGV reference (CSRGV) when the Diagnostic Alarm is annunciated and then manually position the IGVs to the same reference and note the LVDT feedback from the two devices. You could try increasing the manual position reference by one degree at a time for a few degrees, then decreasing the manual position reference by one degree at a time for a few degrees and monitoring the LVDT voltages from the two LVDTs on the AutoCalibrate display. (I presume you are using AutoCalibrate for manually stroking the IGVs; or are you using a Demand Display?)

Dear CSA,

Thanks a lot for your help. I apologize for replying late. We did replace the Gas Fuel Servo Valve after that and the turbine is working fine since then.

I think oil varnish in the Fuel Gas Servo was the main cause for fluctuations in the fuel gas pressure. It made the IGV servo to fluctuate in response. At one point, as you mentioned, they reached a resonant frequency where the difference between the required IGV and actual was large enough to generate an alarm and trip the turbine after that. That can also be anticipated by looking at the time trends of the required and actual IGV signals. We could notice that the 'required IGV' signal was preceding the actual IGV signal, i.e. it was changing first then the 'actual IGV' signal will follow it. If the problem was in the IGV system, let it be the servo, LVDT, or actuator, then the 'actual IGV' signal will change first, but the 'required IGV' signal will not change, or maybe change slightly. Am I right ?

I think by increasing the load, we just 'solved' the oil varnish problem in the gas fuel servo for a period of time by fully opening it.

Thanks a lot CSA. Your replies were of great benefit to us.

Mohamedibr752,

Thanks very much for the feedback! Glad to hear things worked out!

Dear Mohamed,

Here you said alarms appear at above 10 MW. Please could you tell us at which position of IGV, you get those alarms?

I think if the same position every time you get trip exp:50deg, knowing you've changed the servo, I suggest you verify physically the IGV stage in the bellmouth, and verify the actuator again.