LVDT AUTO CALIBRATION

S

Thread Starter

SOHEIL

HI.
AROUND 2 MONTHES AGO OUR OPERATOR INFORMED ME ABOUT RECIEVING "SRV NOT FOLLOWING COMMAND" ALARM IN ONE OF OUR GAS TURBINES (GE-MS_5002D / MARK VI ).

WHEN I MONITORED THAT THE FEEDBACK WAS 3.2% LESS THAN THE COMMAND. FROM THE COMMISSIONING TO THAT DATE I DID NOT SEE THAT PROBLEM SO I ASKED THEM TO STOP THE TURBINE TO RECTIFY THE PROBLEM. AFTER STOP THE TURBINE I DID STROK CHEDK AND FOUND THAT THE FEEDBACK WAS ALWAYS AROUND 1.1% LESS THAN THE COMMAND. AFTER ADJUSTING THE LVDT'S (0.7 VOLT) I DID LVDT AUTO CALIBRATION AND CHECKED THE STROK. STILL THERE WAS AROUND 1% ERROR. WE STARTED THE GAS TURBINE AGAIN AND WHEN IT GOT LOADED THE ERROR REACHED TO 2.5% AGAIN. BY PASSING THE TIME AFTER AROUND 2 WEEKS AGAIN WE GOT THE ALARM. WE DID NOT STOP IT AGAIN AND NOW SOMETIMES IT IS IN ALARM AND SOME TIMES IS OK (ALARM SET POINT = 3% AND TRIP SET POINT IS 5%).

PLEASE INFORM ME WHAT IS THE PROBLEM AND WHAT SHOULD I DO ?!!! THANKS.
 
P

Process Value

With the given information it looks like a typical problem with the null bias. This is another topic which is beaten to death in control.com

look through the topic below and read everything in it. including the constructive criticism by CSA ;).

http://www.control.com/thread/1289667103

you will get to know the servo working , how to adjust the null bias. and it will also help you to understand why your srv is behaving in this manner. adjust the correct null bias value and the problem will be solved (hopefully).

you have said that this problem did not happen from the commissioning , so i am guessing that the servo gain , the polarity and all were right. with the given information , my bet would be on servo null bias correction.
 
I agree with everything said above. That thread is a very good explanation of the null bias. If your null bias has changed suddenly, it indicates a change in the balance between the strength of the spring pushing against the hydraulics. If your feedback is now lower than the reference, the hydraulic side has likely gotten weaker. Besides servo electrical and null bias checks, your checks should include verifying the hydraulic pressure and any last-chance filters on the valve and servo assemblies.
 
There are so many things wrong with this reply.

First, the originator did not say if the feedback matched the actual position. It appears he is just comparing feedback to reference without knowing if the feedback is even close to the actual physical position.

Second, it has been assumed that the current (existing) gain and null bias values are within specification. The originator should replay with the current null bias and regulator gain values to determine if they are within specification. Further, the originator should tell us what type of gas valve assembly is present on the unit--whether it's the combined SRV/GCV gas valve assembly (with both valves in a single casting, side-by-side) or some other configuration. If it's not the typical side-by-side combined gas valve assembly, then the originator should provide some information on the manufacturer and style of SRV in use (Fisher Cam Vee-ball, or Woodward right-angle plug, or ?).

Third, we don't know if the actual P2 pressure is equal to the P2 pressure reference, which is what the SRV does during normal operation--it controls the pressure upstream of the GCV, usually in proportion to HP shaft speed. The originator needs to tell us if the SRV is doing it's job--controlling actual P2 pressure to be equal to the P2 pressure reference. So, it really doesn't matter if the P2 is in the physical position (stroke) its being told to be in--as long as the actual P2 pressure is equal to the P2 pressure reference. (It should be close, but the real critical parameter is actual P2 pressure vs. P2 pressure reference.) The LVDTs are just there for stability and position indication. The Mark VI is going to put the valve at whatever position it needs to, regardless of the feedback, to make the actual P2 pressure equal to the P2 pressure reference.

Fourth, it's not normal for the signal to "drift" over such a short period of time. It's likely something else is contributing to this problem. Oil temperature is one of the first things that comes to mind, and a leaky hydraulic actuator is the second, a servo problem would be the third, and an unstable hydraulic pressure or high hydraulic filter differential pressure would be the fourth. Usually, if there is any error (and there are literally thousands of machines that run with error between the feedback and reference, even if the feedback is equal to the actual physical position!) it stays pretty constant over time. So, something seems to be causing some drift, which isn't typical.

Another possible cause could be fluctuating or excessively low, or even high, gas fuel supply pressure. Or, even a varying gas fuel heat content (outside an acceptable range).

So, it's pretty important for the originator to tell us some critical information prior to telling him to "change the null bias value".

First, we need to know what the existing null bias values are.

Second, we need to know what type of SRV is in use.

Third, we need to know what the SRV regulator gain is.

Fourth, we need to know if the actual P2 Pressure is equal to and following the P2 Pressure reference, and because this is a two-shaft machine and HP speed is variable, it is important to know if there is any error between actual P2 pressure and the P2 pressure reference if it's constant over the operating speed range or if it's changing.

We should have more information and data before making any recommendations to change anything.
 
Hi my friend CSA,in advance thank you because of your kindly cooperation. here below are some information that you asked me:

null bias:2.67
regulator gain:1.8
Min LVDT:0.788646 0.789436 0.789859
Max LVDT:3.570506 3.572509 3.574672

The pressure between SRV & GCV IS ALWAYS 16 bar and it follows the reference by changing the LP Speed.

Our SRV and GCV are compact in an assembly:
NOUVO PIGNONE
CA216GRWCR BFM

NOTE:I monitored the P2 pressure is constant when speed is constant and it will change when we change the speed(very slightly and after stop the changing of speed the p2 pressure immediately will be constant in the reference value_16bar).
I hope my information is enough for you .

thanks
 
soheil,

No; this is not sufficient information. We asked if the P2 pressure was constant and if it was equal to the P2 pressure reference. The typical signal names for P2 pressure and P2 pressure reference are FPG2 and FPRG, though they may differ slightly depending on packager's preferences and current "standards". So, we don't know if the Mark VI is able to maintain the P2 pressure equal to the P2 pressure reference. Just telling us that it's "stable" and changes with load but then is "constant" doesn't tell us if P2 pressure is being controlled properly.

It's also not clear what you mean when you say P2 pressure "...IS ALWAYS 16 bar and changes when we change speed and it follows the reference when we change LP speed...." That just does not compute.

In my experience with two-shaft machines, which isn't a lot and it's been many years since I worked on them, when one changes the LP speed reference, the HP speed will usually change within a range, up to a point. It's probably likely that this unit is running at or near maximum power output and that may be the reason why the P2 pressure reference isn't changing, but that's just a guess on my part. I couldn't understand how LP speed could be used for a P2 pressure reference when the LP isn't turning during start-up of the unit. So, something's amiss here.

Again, I don't have a great deal of experience with two-shaft machines, but I suspect the SRV regulator gain constant is a little low. It's been many years, but if I remember correctly the gain for most of these units was approximately 2.8.

<b>TO BE CERTAIN</b> about the proper regulator gain value you should refer to 1) the Control Specification provided with the unit, or with the control system if this unit was retrofitted with a Mark VI to replace an older control system, or, 2) the packager (GE O&G, also known as Nuovo Pignone, or whomever you obtain OEM service/support from).

NOTE that I <b>am NOT</b> suggesting you just increase the regulator gain without investigating or enquiring as to what the specification is. Again, the value just seems a little low in my experience.

It's still not clear why the error is increasing with time ("drifting"). If the unit was retrofitted with a Mark VI control system, that means the unit is older, and there could certainly be problems with the actuator. Actuators which operate for long periods of time at a steady state condition usually do experience worn cylinders and this leads to leaks and error problems.

The null bias current value seems to be okay, at least it's not out of the normal range.

Can you have a look at the Prevote Data for the servo currents for the SRV, all three of them (<R>'s, <S>'s, and <T>'s) and tell us what they are, please?

Another thing which seems out of specification are the LVDT min and max values. When the zero stroke voltage is set for 0.700 VAC RMS, the maximum stroke feedback voltage should not exceed 3.500 VAC RMS. These two, 0.700- and 3.500 VAC RMS are approaching the endpoints of linear feedback versus position.

GE's specifications for most LVDTs are that they be set to 0.700 VAC RMS at zero stroke, and that voltage is to be read using a TRUE AC RMS voltmeter. You had indicated previously that you had adjusted the zero stroke voltages to 0.7 VOLT, but did you use a TRUE AC RMS voltmeter to measure the voltages?

Finally, what is the position of the SRV under steady state operating conditions? And are both LVDTs indicating the same position or is one higher than the other? If there is a difference in position feedback of the two LVDTs of more than, say, 2%, how much is the difference? Remember the Mark VI is selecting the higher of the two LVDT feedbacks as the actual position, so this alarm is trying to tell you that the error between the two is becoming excessive and could lead to problems in the event one signal or the other fails.

So, to be clear, we need to know:

1) P2 pressure reference vs. actual P2 pressure

2) Is this a "new unit" Mark VI installation, or a retrofit of an older control system on an older unit?

3) Did you use a TRUE AC RMS voltmeter when measuring the LVDT zero stroke voltages? (Meters which are TRUE AC RMS are usually prominently marked as such on the faceplate, and they are more costly than non-RMS voltmeters hence the reason for prominently displaying the difference on the faceplate.)

4) What is the SRV position under normal operating conditions? While physically observing the SRV valve stem while the unit is operating, is it stable or does it oscillate ever so slightly? Could you balance a wide-edged coin on the LVDT bar that is clamped on the valve stem while the unit is running at a steady power output?

5) What is the gas fuel supply pressure, and is it relatively stable or does it vary during the day or under some other conditions (such as compressor change-over or ???)? The SRV has to change it's position to maintain P2 pressure if the gas fuel supply pressure is not stable even if the GCV position is stable.

6) What are the three processor servo output currents under steady-state operating conditions (from the Prevote tab/display)?

7) What Diagnostic Alarms are present on the VSVO card for the SRV and GCV?
 
Hello over there...
Refresher FR5002 basic control philosophy refer to: http://www.control.com/thread/1239086553#1239183223

Since you have intermittent problem with the feedback it appears that it's a mechanical problem such as sticky actuators, loose hydraulic lines on the actuators etc. (as mentioned by CSA).
Other contributor to this problem could be also due to continuity issues in the wiring of the LVDT's.

As regard the regulator configuration for BIAS, 2.67 is typical and GAIN 1.8.
Verify from the control spec. the mechanical travel of the valves. Nuovo Pignone valve for 5002x is typical 1.9” for SRV and 2.1” for GCV. Did you verify the mechanical stroke with the corresponding electrical travel?

Good Luck...
 
A. Oztas,

Are you saying that on these units the HP speed is constant over the operating range of the LP, and that the second stage nozzles can keep the HP speed constant over the operating range of the LP?

If so, that would mean that we really need to know if the actual P2 pressure is equal to the P2 pressure reference, and what the individual processor servo currents are. And it would be very helpful to know what the feedback is from each of the LVDTs.

AutoCalibrate has a feature called 'Verify Current' and if I remember correctly when selected (when the unit is off-line and the gas fuel supply pressure is isolated and at zero bar!) the Mark VI will put out whatever current is required to make the position change at a constant rate while opening and closing the valve and plots current vs. position on a graph which can be captured.

There's also a feature called 'Verify Position' and the Mark VI will put out out a constant current and plot the LVDT feedback and if the feedback (either of them is non-linear) then that would be obvious on the graph.

It will also be very interesting to see what Diagnostic Alarms are present when the unit is operating and the Process error alarm is present.

We are anxiously awaiting the data, soheil.

(But we're not holding our breath while waiting, either.)
 
HI MY FRIENDS:
I WANT TO INFORM YOU THAT FOR 5 DAYS I AM NOT IN OUR GAS PLANT,AFTER COMING BACK I WILL PREPARE PROPER DATA FOR YOU.
THANKS
 
Hi

we are using 6FA machine with MK-V TMR controller. My question is regarding MOOG valve operation. G771 series moog valve for GCV. In the Moog valve the flapper is fixed (fixed with bolt) at first stage. in such condition how the spool movement is taking place proportional to the command given.
 
Top