the unit trip by IGV control failure and restarted again with no inspections except loose connections check due to load demand. what are the parameters that should be checked and what are the values of the feedback and excitation of LVDT. (the unit still running with normal condition).

 

Were there any alarms indicating a problem with IGV control prior to the trip? If this is a Mk IV, Mk V, or Mk VI, where there any Diagnostic Alarms related to IGV LVDT feedback or servo current prior to the alarm? Presuming there are two LVDTs on the IGV actuator, are the two voltage feedback values nearly equal (they don't have to be but usually are)? Are the scaled feedback positions nearly equal? Are the excitation voltages nearly equal?

(Remember, when working with LVDTs and measuring LVDT voltages--excitation AND feedback--one must use a True AC RMS (Root Mean Square) voltmeter due to the nature of the excitation output and its low magnitude which doesn't work well with non-True AC RMS voltmeters.)

There are several conditions which can cause this alarm (which does not appear to be a "normal" Process Alarm text message, but an abbreviated one or a "custom" one designed by one of the packagers of GE-design heavy-duty gas turbines). It would be MOST helpful if you would provide the Process Alarm Drop Number AND the exact text message which was displayed when the trip occurred (presuming the unit is equipped with a Mk IV, a Mk V, or a Mk VI SpeedTronic turbine control panel).

Several of the servo-related trips (servo-operated devices include the SRV-Stop/Ratio Valve, GCV-Gas Control Valve, GSV-Gas Splitter Valve, GTV-Gas Transfer Valve, IGV-Inlet Guide Vane, LFBV-Liquid Fuel Bypass Valve) are actually initiated by more than one condition--while not a great programming practice, it was necessary to conserve Process Alarm Drop numbers, and because that's the way it was done for several generations of SpeedTronic turbine control panels.

Each of the conditions is usually "latched" with a Master Reset which allows the operator or technician to check to see which of the possible trip conditions actually initiated the trip--AS LONG AS THE OPERATOR OR TECHNICIAN CHECKS THE LOGIC _BEFORE_ INITIATING A MASTER RESET!!!! (Unfortunately, one of the first things many operators do when ANY alarm condition occurs is to initiate a Master Reset, which AIN'T good operating practice, no matter how "convenient" or "easy" it might be to clear an alarm from the display. For many alarms, such as these, it's impossible to troubleshoot AFTER a Master Reset has been initiated!)

Among the conditions which can cause these alarms is a loss of feedback (actually, both of the LVDTs would have to fail low for this to happen, since most devices are equipped with dual redundant LVDTs to protect against this condition); excessive servo current; and a failure to "track" which means the actual feedback (from the higher of the two LVDT feedbacks) differs from the reference by more than a pre-determined amount. Use the Mk II or Mk IV SpeedTronic elementary, the Mk V CSP, or the Mk VI Toolbox to work "backwards" from the Process Alarm Drop Number to see the condition(s) which would cause this alarm.

As for LVDT excitation voltages, GE's practice was to use approximately 7.0 VAC RMS to excite LVDTs, and that is certainly the value you should be measuring (using a True AC RMS voltmeter!) if the SpeedTronic turbine control panel is a Mk IV, a Mk V, or a Mk VI.

The value of the LVDT feedback varies with the stroke of the device, the LVDT used, and the actual physical stroke of the actuator. It was GE's practice to set the zero-stroke (or, for the IGVs, the minimum mechanical stroke) voltage to approximately 0.700 VAC RMS, +/- 0.020 VAC RMS (this is for non-Intrinsically Safe Applications without Zener Barriers!), and for the stroke of the device not to cause the LVDT output to exceed approximately 3.500 VAC RMS. On some applications, the maximum stroke voltage would be very near 3.500 VAC RMS, and on some applications it would only be about 3.10 VAC RMS, and on others it would be less that 2.500 VAC RMS. (Again, these voltages are for non-intrinsically safe applications without zener barriers!)

It's always recommended that during calibration of LVDT feedback that voltages for zero-stroke and maximum stroke, as well as 25-, 50- 75- and 100% stroke readings be recorded for just such conditions as this.

Loose connections could certainly cause problems--they have thousands of time over the years. It's a good practice that during any major outage (Combustion Inspection, Hot Gas Path Inspection, or, particularly during a Major Inpsection) to check all terminal board connections in every junction box as well as in the SpeedTronic turbine control panel for tightness. It's _NOT_ necessary to use excessive force and destroy the terminal screw heads to ensure good tight connections, just reasonable force to ensure tighness. Many "electricans" think they MUST feel some rotation of the screw head to ensure tighness; think of what can happen if this is done EVERY time a screw is checked for tighness! Just using a reasonable amount of force to ensure tightness is all that's required; if some rotation is experienced, then the connection was probably loose to begin with and should be re-tightened. If no rotation is experienced, then the connection is probably tight enough. (There are several manufacturers that sell fixed- or adjustable torque screwdrivers, but except for the US nuclear industry that's just a little overkill for most applications!)

Again, use your sequencing/logic elementary/printout to work backward from the Alarm Drop Number, and you should be able to determine what could have caused the alarm. And then investigate those causes--if you're not convinced that loose terminations could have caused the problem. Or, if there were no related Diagnostic Alarms present before--or after--the trip!

markvguy

 

the unit is frame 5 with TMR MK5 and HMI monitoring system. According to trip history page, the tripping was caused by L4IGVT signal which is (as you know) activated when the value of CSGV drops below 50 degree. We didn't recive any other alarms except what is mentioned above.Regarding the checks of IGV, could you please give me the name of command that is used to open IGV at crank speed? Another thing, in our unit the IGV is variable (i.e. from zero speed to approx. 85%=44deg, from 85% to F.S.N.L=57deg. and then vary to max. 85 deg.). My question, where is the min. stroke and max. stroke position?

 

Modulated, or Variable, Inlet Guide Vanes typically operate as follows: From initiation of START to approximately 75-85% TNH (depending on the Frame size and the vintage of the control scheme) the IGVs are held near the minimum mechanical stop which is usually 34 DGA (Degrees Angle). As you noted they then open slowly to the Minimum Operating Position of 57 DGA, and they are usually at 57 DGA at FSNL.

There have been several iterations of IGV Control Schemes, but one of them had the IGVs stay at 57 DGA until the Exhaust Temperature reached approximately 900 deg F, at which time they would open slowly to maintain 900 deg F as the unit was loaded until the IGVs were full open (CSKGVMAX), usually 84 DGA for Frame 5s, and they would remain at 84 DGA all the way to Base Load. This was sometimes known as "Simple Cycle IGV Control." Sometimes when the IGVs were being modulated open during this period a "IGV Temp Control" message is displayed.

If the unit's exhaust is directed to a Heat Recovery Steam Generator (HRSG, or "boiler"), there was usually an option to select "Combined Cycle Exhaust Temperature Control," which usually kept the IGVs at 57 DGA until the unit reached the isothermal temperature (approximately 1040 deg F to 1100 deg F for Frame 5s depending on vintage), at which time they would then open to maintain the isotherm temperature or a value just slightly less than TTRX (sometimes TTRXGV, IGV Exhaust Temperature Control Reference) until they were fully open and then the unit should go smoothly on CPD-biased Exhaust Temperature Control.

The IGV Reference is usually CSRGV (Control Stroke Reference-Guide Vane). That value is sometimes written to another signal, CSRGVOUT, which is the signal assigned to the servo-valve output to which the IGV servo-valve is connected (which should usually be SVO5, Servo-valve Output #5).

If you want to stroke the IGVs at zero speed (assuming you have an AC motor-driven Aux. Hyd. Pump), you need to establish L.O. Pressure, Hyd. Pressure, you usually need to force L20TV1X (check you I/O Report and CSP for the EXACT name used on your application) to allow high-pressure Hydraulic fluid to get through the IGV Dump Valve to the IGV Actuator. Then, use the 'Manual Setpoint' feature of AutoCalibrate to position the IGVs at the desired angle and observe the LVDT feedback. Or, some units weren't properly configured to use AutoCalibrate and have a User-Defined, or Demand, Display for stroking the IGVs. (The sequencing on some units required the Master Select to be in CRANK Mode to get a permissive to use AutoCalibrate...)

If your unit was not supplied with a Control Specification-System Settings document (which would have many signal names and descriptions of control schemes, the IGVs are detailed in Sect. 06), then your OEM Instruction Manual should have something which describes the operation in some detail (perhaps not as much as you'd like, but better than nothing).

One way to understand (the BEST way!) functions is to work "backwards" from an input or output, in this case the IGV Servo-valve Output. Look in the I/O Report or IO.ASG to find the signal name of the output which drives the IGV servo. Then look up that name in the CSP Cross Reference to find where it's written to (the segment and rung number will have a "-" in front of it), and then go to the CSP and find the signal and start following it backwards. Most of the BBLs are not that difficult to understand--if you have questions, ask them here--and you can learn a great deal about the control schemes and how they work and what Control Constants and signal names are used to control operation, and at what points/speeds/times events are supposed to occur.

It is most helpful to also have the "Piping Schematic" drawings, as GE calls them, or "P&IDs" as the rest of the world calls them (Piping and Instrumentation Diagrams). By looking at the drawing/diagram for the IGVs, you should see the IGV Dump Valve Solenoid, usually 20TV-1. This solenoid must be energized to allow the servo-valve to use high-pressure hydraulic fluid to position the IGVs so that the IGV LVDT position feedback signal, CSGV (Control Stroke-Guide Vane), is equal to the IGV position reference, CSRGV.

In your case, if L4IGVT is actuated when the higher of the two LVDT feedbacks drops below 50 DGA, that means that the feedback from BOTH LVDTs was less than 50 DGA. If the unit was properly configured and wired, this would mean that the LVDT excitation signal from two of the control processors (<R>, <S>, and/or <T>) failed (not likely), or there might have been a problem with the LVDT feedback terminal board, usually TBQC, which receives the LVDT feedback signals and "fans them out" to all three control processors (see the Mk V App. Manual, Sect. 7.6 'Exmples of Regulator Applications,' for more details and drawings of typical applications). BOTH LVDT feedback signals are wired to ALL THREE control processors, where they are converted to digital signals and then compared, in software, to select the higher of the two feedback signals, which is used to compare against the reference. Again, this is done in all three control processors, and the servo-valve output current of each control processor is adjusted to try to make the feedback equal to the reference.

It's NOT likely that the output of both LVDTs failed, and since each LVDT _SHOULD_ be excited by a separate control processor (in other words, both IGV LVDTs should NOT be excited by the SAME control processor--same for all redundant LVDTs) it would require two control processors' LVDT excitation outputs to fail, which is not likely either.

Another thing which might be the cause of the problem is a failure of the IGV Dump Valve and/or it's solenoid, 20TV-1, or the solenoid output from the Mk V. If 20TV-1 is not energized, then high-pressure hydraulic fluid will CLOSE the IGVs. If there is a loose connection or faulty solenoid output/relay, this could cause the IGVs to actually drop below 50 DGA. If this happened, there would have been a power output decrease BEFORE the trip as the IGVs were closing. If the IGVs went closed fast enough, the exhaust temperature would have been rising quickly, approaching an exhaust overtemperature alarm/trip. You did not say what load the unit was at when the IGV Trip occurred.?.?.?

So there's several potential problems: Improperly configured/wired LVDT excitation (BOTH LVDTs excited from the same control processor would allow a problem with that excitation to cause the LVDT output from both LVDTs to go low); a problem with the TBQC terminal board and/or it's cables/connectors; a problem with the IGV Dump Valve Solenoid, the dump valve, the solenoid output, or a loose wire/terminal in the solenoid circuit.

As an aside, this author has seen several HMIs not properly configured to display Diagnostic Alarms, and found that the Process Alarm which sounds to warn of a Diagnostic Alarm condition was locked out because of nuisance Diag. Alarms. This prevents any Diag. Alarm from ever being "annunciated" or displayed or looked at.

It's been said before, but Diagnostic Alarms are your friend, and no unit should have nuisance Diagnostic (or, "Diagnasty") Alarms. Every Diagnostic Alarm can be resolved, some GE even permits to be permanently "disabled" because they are aware of hardware/software problems which cannot be easily resolved and the conditions are NOT critical.

Just because a unit continues to run with Diagnostic Alarms present and "coming in and out" (i.e., nuisance, dithering alarms) DOESN'T mean everything is okay.

A couple of the sites this author visited with improperly configured HMIs had been experiencing mysterious and intermittent trips and other problems. When the Diagnostic Alarm Display functionality was fixed there were several pages of Diag. Alarms, several of which, once resolved, cured the mysterious operation and nuisance trips.

Does your HMI display Diagnostic Alarms? There are Diag. Alarms even on solenoid/contact outputs.

markvguy

 

That was the most enlightening piece on IGV control trouble trip troubleshooting that I have read. We have a 20 MW,Frame V, Mark V unit which tripped today on IGV control trouble. This unit has been running well beyond the MI schedule. The machine was running with 16 MW load and CSGV at 57 degrees. The trip log also shows the value to be at 57 degrees moments before and after trip. It implies that the IGV had closed beyond 50 degrees momentarily ( within 500 milliseconds).The tripping is instantaneous and without any alarms. Why does not GE put at least a minimal time delay of say 0.5 to 1 second for this trip. The instantaneous flag can be used for alarm purpose.
Varnish in hydraulic oil due to prolonged use and oxidation causes stickiness of servovalves. This can be a root cause of trips of IGV control trouble. The last chance filter in IGV servovalve may be choked and it can also cause the same trip.Testing of the oil for it's hydraulic properties can give you an insight.
However, I still feel a time delay ought to be given to take care of unplanned outages.
Pls reply if you know of units employing time delay in IGV control trouble logic with L14IGVT

Regards,

K. Jayapal

 

Some units do alarm before the trip... Depends on the vintage of the control system as well as who the OEM/packager was and what their control philosophy was at the time the unit was built/configured.

Do you know exactly which IGV Fault condition the unit tripped on? (There is usually more than one which feeds the single trip and alarm message; most are latched with a Master Reset--which is why it's SO important to check the trip log and sequencing BEFORE issuing a Master Reset when the unit trips.)

How many IGV LVDTs does your unit have? Some OEM/packagers only provided one.... It would not be good to operate the unit with the IGVs at less than 57 DGA while loaded for any period of time. Do you think an operator could respond to a problem in half a second? Are you experiencing problems with the LVDT(s)?

markvguy

 

I operate a GE Frame 6B Turbine with a mark v control System. My unit tripped on high exhaust temperature due to the failure of our exhaust frame blower.
After the trip, the following diagnostic alarms showed D 1240 TCQA LVDT9 out of limits, D 1316 TCQA Servo Current #5 disagrees w/ref, D 3339 Voter mismatch, <R> CSGV, D3580 Voter mismatch <S> CSGV, D 1340 TCQA LVDT Position diff high REG#5 and process alarm P109, Inlet Guide Vane Position Servo Trouble.

The following permissive Inlet Guide Vane Position Trouble L3IGVFLT is preventing me from starting up and the IGV angle is 39 degrees as against 36degrees which shows the closed position.

Please any suggestions on what I should do?

Thank you.

 

Upongabasi Akpabio,

The Diagnostic Alarms seem to be saying the LVDT feedback from the two LVDTs (which should be nearly identical) is not similar, and the typical difference between the two feedbacks that causes this alarm is 5 DGA (DeGrees Angle). You should be able to use the Prevote Data Display to check the two values of LVDT feedback and see what they are indicating--though there is a little "feature" of Mark V software that can make that difficult.... Unfortunately.

If the unit has an Auxiliary Hydraulic Pump (AC motor-driven), have you tried starting the Aux. L.O. Pump and then starting the Aux. Hyd. Pump and watching the IGV angle (CSGV) to see what happens? (It's possible the IGVs might move to the fully closed position.)

Another possible way to check the two LVDT feedback values is to open AutoCalibrate, and go to the IGV Regulator screen (SVO5) and check the two feedback values--LVDT1 (actually LVDT input 9) and LVDT2 (actually LVDT input 10).

Once you know what the difference between the two LVDT feedback values are, you can then decide how to proceed. But, I would first suggest trying to use hydraulic pressure to get the IGV to move--if it will and if the unit has an Aux. Hyd. Pump--and hopefully this will resolve the problem easily (it's possible, so it should be attempted).

It's difficult to understand why the two LVDT feedback values might drift apart from each other or why a trip from load would cause them to drift apart, but then we don't know what they were before the trip, and we don't know the condition of the IGV actuator (which is very difficult to get to in order to visually examine). It's possible that one of the jam nuts on one of the LVDT cores has vibrated loose and is causing the error in output. We don't know if one or both of the LVDTs was replaced "recently." There's a LOT we don't know about the unit at your site, how it's been maintained, and what your familiarity with the Mark V and other systems on the unit are.

But we would really need to know a lot more about your machine--and most importantly, what the two values of LVDT feedback from 96TV-1 and 96TV-2 are to be of much more help. AND, it would be a good idea if you or someone at your site started trying to get to the IGV actuator which includes the two LVDTs (usually located directly underneath the IGV inlet plenum base, and make a visual inspection of it, it's mounting bolts, and the LVDT cores/jam nuts. Pictures or videos would be good for others at the site to be able to see the as-found condition(s). Good light is ESSENTIAL to a visual inspection and photos/videos!!! It's VERY dark down there and a torch sometimes won't provide enough light. Usually there is an access plate in the bottom of the inlet plenum which can be removed to at least partially see the IGV actuator.

If it's been a LONG time since any maintenance was done on the IGV actuator, be aware that that particular actuator assembly is known to take a lot of abuse even under normal conditions, and the mounting bolts which hold the actuator in place can very often come loose and allow for a lot of play in the system--which can cause lots of problem with calibrating the LVDTs. The Heim joint can also wear out, and the actuator's double-acting piston can also have internal wear and seal leakage (though this won't usually cause the condition you are describing). I'm just trying to let you know that because this actuator is SO difficult to access, it's usually very often overlooked in normal maintenance.

Best of luck! This is not an easy problem to try to help with via a World Wide Web forum. If the problem isn't easily resolved with starting/running the Aux. Hyd. Pump, I would strongly recommend getting a knowledgeable person to site to help with the problem. You can check all the wiring between the LVDTs (which are, again, attached to the IGV actuator which is usually in a very difficult position to get to (here again, we don't know how old the unit is or where the IGV actuator on your unit is actually located!)) to be sure there's not a loose connection somewhere (embarassing and expensive to have to pay someone to come to site to find/fix this when it is a relatively simple thing to check).

 

I started the Aux Hydraulic pump, but what i noticed on the screen was the IGV angle increased to 41, instead of reducing to 36. Now resistance test on the coils of the servo valve shows 1.92M ohms, 1.91M ohms and 0.6M ohms. The resistance of the LVDT is 32.6 ohms.
From visual inspection the IGV angle is in close position 36degrees.

 

In my experience the servo valve coil resistance can be derived from Ohm’s Law. The maximum current passing through a coil should be plus-or-minus 10 mA (between positive and negative 10 mA). This should yield voltages between +10 VDC and-10 VDC. This means the coil resistance should be approximately 1000 ohms. Each.

If you’re measuring megohms, something is wrong—most likely your measurement method.

There should (normally) be two LVDTs—96TV-1 & 96TV-2. The Mark* is supposed to be configured to use the higher of the two feedback signals. In the Mark V the higher of the two feedback signals is written to CSGV. This is where the “feature” appears. Usually 96TV-1 is assigned to be CSGV. BUT if the feedback signal from 96TV-2 is higher than the feedback signal from 96TV-1 then the feedback signal from 96TV-2 is ALSO written to CSGV. (Isn’t this fun??!?!?!!?!!)

So, if the feedback signal from BOTH LVDTs is the same value, then the feedback signal from 96TV-2 is higher than the feedback signal from 96TV-1–and it’s very difficult to then see the actual feedback signal from 96TV-1…. NOT FUN!!!!
I strongly recommend you have a knowledgeable person travel to site to help with the problem. That person can gather more information than you are sharing now and probably has a good idea of how to troubleshoot the problem, especially given that the Mark V has a special little feature for LVDT feedback.

it may even be that the unit at your site only has one IGV LVDT—that did happen on older, early Frame 6B GE-design heavy duty gas turbines.

Best of luck!