SRV Fluctuating

Dear
we have a problem with the SRV valve, which fluctuate for a load over then 6MW (GCV and load is stable), note that for lower loads the valve is stable.
The turbine is a 6B with an MK6e TMR controler; in a separate electrical grid.
The HP filters, the servovalve and its filter were changed without result.
Thanks in advance for all support.
 
luiss,

When did this problem start? After a trip from load? After a maintenance outage? ??? After someone "calibrated" the SRV LVDT(s)?

What process alarms are present for loads below 6 MW?

Which terminal board and associated I/O Pack is (are) the P2 pressure transducer(s) connected to? Have you checked the Diagnostic Alarms for the I/O Pack the P2 pressure transducer(s) is(are) connected to? If so, what are they?

Have you tried recording the SRV position just before it goes unstable, and then, with the gas fuel supply pressure isolated (shut off) have you tried opening the SRV using the Manual Positioning feature of the SRV LVDT Calibration function to move the SRV to that position while observing the valve (placing someone physically next to the SRV to watch the valve position) and then increasing the position reference slowly, all the while monitoring the SRV position? (SRV position should be recorded to a Trend during this operation, too--usually.)

It could be an actuator problem--particularly if the 6B is older and the Mark VIe is newer (an upgrade to an older control system).

It could be an LVDT problem; things like this have been known to happen if the movable core of the LVDT is not straight and rubs against the inside of the stationary armature of the LVDT. How many LVDTs does the SRV have? Have you tried monitoring the feedback from the(both) LVDT(s) and then observing what happens when the load is increased above 6 MW? If so, please describe what you have observed.

Was any mechanical work done on the SRV prior to this problem starting?

Is the SRV contained in the same assembly as the GCV (a large casting housing both valves)? Or, is the SRV separate from the GCV (and if so is it a rotary Fisher Cam Vee-ball valve)? If it's the rotary Fisher valve, does it use LVDT(s) or an RVDT assembly?

Do you know how to use the Verify Position and Verify Current features of the LVDT calibration function?

Has the gas fuel supply pressure changed recently (possibly just before this problem started)?

Does the unit normally burn gas fuel, or is it a multi-fuel unit, and if so, what are the other fuels the unit burns and what percentage of time is each fuel used (roughly--the estimation doesn't have to be exact)?
 
Dear CSA

This problem started recently, no intervention or calibration was done in beginning of this problem. But it is noted that the turbine has been operating with a reduced load of around 4 MW for a long period of time (more then one year).
P2 is connected to in PCAA TCAS Analog Input (R,S,T) and we don't have alarm or diagnostic alarm in them.
when we open the SRV using manual positioning, a small fluctuation of the two signals fsgr and fagr when the valve moves from one position to another (instantaneous fluctuation) .
The actuator servo valve already changed and still have the same problem. (do you have some recommendation for new values to upgrade setting of regulator).
We have two LVDTs on the SRV, some impurities was founded between the core and the axis of LVDT, but even after cleaning and restarting the unit the fluctuation is still persistent.
When the load is greater than 6 MW we notice that the position of LVDT also starts to fluctuate and which is remarkable that the value of the signal fagr (speed ratio valve current) is not stable even for low loads, and it becomes more and more fluctuating at intense loads!
No mechanical work was done on the SRV before this problem started and the gas fuel supply pressure never changed.
For the calibration i use the auto calibration function. and as I told you the values of fagr and fsgr are not stable when changing the valve from one position to another then stabilizes, just at the time of change we observe an instantaneous fluctuation (if you have a special procedure please send it to me).
The unit is dual fuel gas and liquid, but usually use the fuel gas.
Thank for reply.
 
luiss,

I'm going to presume the turbine-generator is older and previously had a earlier Mark* turbine control system and was upgraded to a Mark VIe.

I interpret your statement about "... operating with a reduced load of 4 MW for a long period of time (more then one year). ..." to mean the unit ran at about 4 MW only for more than a year.

You also stated the P2 pressure transducer (one, single P2 pressure tranduces) is connected to the PCAA TCAS card, which then "fans out" the one signal to all three control processors (I'm presuming the Mark VIe is a TMR (Tripe Module Redundant) system).

You say you have changed the SRV hydraulic actuator's electro-hydraulic servo valve and still have the same problem.

My recommendations for replacing an electro-hydraulic servo valve are to check the servo regulators' Null Bias Current settings and make sure they are set to 2.67, then immediately after replacing the servo-valve and establishing hydraulic pressure and SRV Trip Oil pressure to perform a servo valve current polarity test--as has been outlined in MANY previous posts on control.com. This is important to ensure the polarity of the servo current being applied to each of the three coils of the electro-hydraulic servo valve are correct (or to the two coils of an electro-hydraulic servo valve being used in a SIMPLEX or DUAL REDUNDANT application). (The electro-hydraulic servo valves used on most GE-design heavy duty gas turbines are bi-polar devices--meaning the current applied can be positive or negative, and the polarity of the current determines if the device will open or close, so the polarity needs to be correct for everything to work properly and can easily be verified at the time of change-out/installation of a servo-valve.)

After the servo current polarity has been verified (that is done under the control of each individual processor) and the wiring has been restored after correcting any polarity issues if found, then that is ALL that should be required when changing an electro-hydraulic servo valve. IT IS NOT NECESSARY to calibrate/re-calibrate LVDT feedback if the only thing that was done to the device was to change the electro-hydraulic servo valve. At this point, one should use AutoCalibrate to check that the actual position and the calibrated feedback position are nearly equal--which means one uses the manual positioning feature of AutoCalibrate to move the device to one or two or sometime more positions and compare the device's ACTUAL, PHYSICAL position to the indicated position on the screen (from the scaled LVDT feedback). If the actual, physical position and the indicated position are significantly different, then some further action should be taken to reconcile the two. This may require changing the servo regulator Null Bias Current, or it may require another "calibration" of the LVDT feedback, depending on the analysis of the differences between the LVDT feedback and the actual, physical position.

But, there is nothing else which should be necessary. It should NEVER be necessary to change the servo regulator current gain or any other parameters--without consulting the OEM/packager or a knowledgeable person. The as-found parameters (as left after commissioning) should ALWAYS be the proper values, as determined by GE or the packager for the actuators and electro-hydraulic servo valves installed on that particular unit. (That's actually one of the GREAT things about GE-design heavy duty gas turbines--those values (current gain, in particular) are usually well-known and easily calculated and determined by "the factory" based on actuators and servo valves which have been used for decades. Many other OEMs tell technicians they have to calibrate the "loop" for stability--and one person's interpretation of stability is NOT EVER the same as another person's interpretation of stability, which leads to a lot of differences in opinion and settings. Not so on GE-design heavy duty gas turbines--the values in the Control Specification, which is supplied with every Mark* turbine control system--are almost ALWAYS the proper values, except when an experienced and knowledgeable commissioning person detects an issue and confirms it with factory engineers.)

Do you, or does anyone at your site, know how to use ToolboxST Trend Recorder to capture data to analyze when you are experiencing this problem? If so, it would be very helpful--almost critical.

A lot of older machines with older hydraulic actuators can have very worn actuator cylinders and actuator piston seals which can cause problems like the one you are trying to describe. Oil quality plays a big part in this as well. I have seen non-working electro-hydraulic servo valves replaced with new electro-hydraulic servo valves which failed, in some cases almost immediately, because of poor oil quality (high total dissolved solids; unclean replacment conditions/practices; high levels of varnish in the oil; etc.). Usually, there is a "last-chance" filter very near to the hydraulic actuator and electro-hydraulic servo valve arrangment and because they rarely get checked and almost never have any kind of alarm connected to the control system to warn of high differential pressure these filters actually eventually rupture releasing LOTS of dirt and contaminants right in to the servo and actuator they are supposed to replace. Last-chance filter monitoring and maintenance is VERY important. If the those filters haven't been changed in a couple of years, it's an excellent idea to do so as soon as possible, and to institute an periodic maintenance routine to at least check them or just replace them on a periodic basis. (Unfortunately, most packagers of GE-design heavy duty gas turbines do no install differential pressure gauges across these last-chance filters. Some, but not all, have a little "button" that will pop out to indicate a high differential pressure, and many people think that just pressing the button back in does something like cleaning the filter (which it doesn't!). So, in a lot of cases, simply replacing the filters on a regular basis is an excellent idea.)

AutoCalibrate also has two "verify" functions: Verify Current and Verify Position. These are useful when troubleshooting servo and/or LVDT problems, but they will only work properly AFTER an AutoCalibration has been performed on the device being troubleshot. I ALWAYS recommend recording the LVDT calibration paramaters in the Mark VIe PRIOR to performing an AutoCalibration of LVDTs, just to have something to compare against AFTER the procedure is completed (and for back-up purposes). Then you can run either verify function, each of which will produce a trend of the results which can then be saved and used for analysis. These will usually indicate either a sticky servo spool piece (which requires excessive current to keep the device moving at a constant rate to keep the LVDT feedback changing at the proper rate), or a problem with LVDT feedback (which you described you may have found, and possibly didn't fully correct that is erratic or doesn't change properly when the device is being moved with a constant current). But, again, these only work AFTER an AutoCalibrate of the device being troubleshot has been performed. And, when the troubleshooting is finished, the as-found (recorded) LVDT calibration parameters from BEFORE the AutoCalilbration should be compared to newly-determined LVDT calibration parameters AFTER the AutoCalibration, and if the two are significantly different then something should be done to reconcile them.

The good news is: SRV LVDT calibration is NOT critical to turbine operation. By that I mean: if the calibrated feedback says the SRV is at 23.74% stroke when it's actually at 26.59% stroke the SRV is still going to work just fine. That's because, the SRV loop is a PRESSURE CONTROL loop--not a position control loop. The Mark* will move the SRV to whatever position is necessary to make the actual P2 pressure equal to the P2 pressure reference--regardless of what that position is on the HMI or at the device. It's going to move to a position that makes--and keeps--the actual P2 pressure equal to P2 pressure reference. So, SRV LVDT calibration accuracy isn't really very important. IT IS IMPORTANT that the LVDT feedback be linear and proportional to stroke--that IS critical; but it's not important that the feedback on the HMI be exactly equal the actual SRV position at any or all positions of the SRV. As long as the LVDT feedback is linear (doesn't "jump" as the SRV moves) and it's "close" to a proper indication of position--it's going to work just fine. Perfection is NOT necessary for SRV LVDT feedback calibration.

Hope this helps! One ALWAYS MUST DO a servo-valve polarity check when replacing an electro-hydraulic servo valve. And that should be done with the servo regulator Null Bias current values at the nominal value of 2.67%/%.
 
luiss,

To Continue (due to the 10000 character limit imposed by control.com)...

You say the SRV is fluctuating. What is the GCV doing when the SRV is fluctuating? Is it also moving? Do you always used Pre-Selected Load Control to operate the turbine to maintain a particular load?

There is a method for "gagging" the GCV which can be useful for troubleshooting problems like this. If you can make the GCV stable but the SRV keeps fluctuating, then the problem is probably related to something to do with the SRV. BUT, if when you make the GCV stable and the SRV stops fluctuating then the problem is someting to do with the GCV. A lot of times SRV instability gets incorrectly blamed on the SRV, when it's either gas fuel supply pressure/flow, or something amiss with the GCV and associated devices and controls.

AS A TEST, if you normally use Pre-Selected Load Control to operate the unit when producing power, use the RAISE SPEED/LOAD and LOWER SPEED/LOAD buttons to change load and monitor what happens and report back. (I PROMISE the unit isn't going to blow up or trip if you disable/de-select Pre-Selected Load Control!!!)

Finally, what other Process Alarms (ALL of them please) are active when the unit is running and you are trying to increase load. It would be helpful if you could also supply the Diagnostic Alarms, as well (easier said than done, but not impossible).

Hope this helps!
 
Dear CSA

I tried with the null bias current, it is good value, I also try to check the polarity of the coils so they are well connected.
what is remarkable that when one changes the position of the valve from 0% to another position 25% by expemble one notices that the current rises more than necessary and the valve make great oxcillation then stabilizes. and this at each change of position 75% ...
If the two coils of the servovalve are disconnected and a single one is left, this phenomenon would no longer be and when changing the valve from one position to another it would be stable.

it is also remarkable that the value of fagr is not stable for a setpoint given for example when we send the consinge 25% the value fsgr osxcillates then stabilizes, on the other hand the value of fagr remains varies between 1.9 and 3.2%.

Note also that I discovered that when we change manuel position for GCV valve we observe that it does not stabilize is it remains oscillates physically, yet it is stable when the turbine running unlike the SRV.

thank you for reply
 
luiss,

People always expect servo current values to be very stable, but in fact they are changing at the rate of 100 or 128 times per second (those are the most common rates of execution of the servo regulator functions--comparing the indicated position to the reference position and adjusting the current to try to keep the indicated position as close as possible to the reference position).

There is a very large spring which is always trying to close the SRV, and so the regulator has to try to overcome that--continuously.

I don't know how you test for polarity, but I still have a suspicion that one of the servo outputs is not connected properly. I have found servo coil wiring to be opposite of what it should be from the factory (meaning the factory misconnected the wires and the color codes are not consistent). I heard a report once of the coil wire colors being misconnected so that two coils ended up being in series with each other instead of being individual; I believe this happened on a refurbished servo not on a new one from the manufacturer. You can test for this very simply (the unit has to be shut down, of course) by lifting all three pairs of servo wires and measuring each pair with an ohmeter (or a VOM set to measure resistance). Each pair of coil wires should measure approximately 1000 ohms.

Finally, usually when the servo current has to be increased and increased and increased to begin to see movement and then there is usually some over-shoot and some instability as it moves back to being steady at nearly the reference position it indicates a problem with the servo itself. (I'm presuming the servo polarities were correct, and the wiring was correct.) Dirt or debris or a damaged servo spool piece or o-ring causes the servo to be mechanically bound requiring more current than should be necessary to result in a change of position, an overshoot and instability. The same thing, though, can also happen if the device being positioned by the servo has a mechanical binding issue--preventing it from moving. I have encountered several SRVs (usually the vertical, linear SRVS--not the rotary SRVs) which had bent shafts which acted like this when trying to be moved.

I have also observed some pretty bad conditions inside the L.O. reservoir tank where the SRVs (the vertical, combined SRV/GCV assemblies are mounted above) and the people changing the servo didn't use the best care when removing the old servo and installing the new servo resulting in dirt getting into the servo's very tiny passages and causing problems almost immediately after being replaced. Cleanliness (of the work area and of the oil) is very critical to servo reliability and operation.

I have also had many issues with refurbished servos--at one site we had to swap six (6) servos before we finally got one that would work properly! (That site had four GE-design heavy duty gas turbines, and had a good stores warehouse.) After that experience, the company stopped buying refurbished servos. It took about 30 hours to get one working servo installed and tested and that site was severely penalized (financially) for not producing enough steam in that period.

I can't offer much more than that. From the description, it sounds like something is mechanically bound--either the servo spool piece or the hydraulic actuator of the SRV or the SRV valve stem or seals. It is indeed odd that the instability doesn't occur when the servo is under the control of one servo. (Have you tried testing with only one servo disconnected--leaving two connected--and recording the results?)

Have you used the Verify Position or Verify Current functions of the LVDT Calibration function? And, then saved the resulting trends for further analysis?

Please write back to let us know what you find!
 
Hi,

We have faced similar issue in our site and we have followed many steps to identify the problem and fix it.

The problem was due to reading missmatch between 3 P2 pressure transmitters. We have re-calibrated all 3 transmitters to read exactly same and then the SRV fluctuation has stopped.
 
This is a real possibility, but it should be accompanied by Diagnostic Alarms (usually is accompanied by Diagnostic Alarms). Recently, though, I saw a site which had turned off the 4-20 mA Diagnostic Alarms thinking that it would fix the root problem.... Of course it didn't; it just prevented the alarm from being annunciated/displayed.

I have only seen this particular problem as described by ssunist on Mark V turbine control systems, but I presume it could happen on Mark VIe if the problem were bad enough (mismatched 4-20 mA outputs).
 
Top