Hi every body

We have GE gas turbine frame 9 dual fuel controlled by Mark IV system.

The following alarm is coming frequently during normal running since long time and without relation with any particular situation:

<R> voting mismatch (CSRGV)
Also some other alarms are coming but not too much like the above mentioned one:
<R> voting mismatch (FSRS1)
<R> voting mismatch (FSRN)

any way the first alarm is the most repeated one, 90% of time it is coming for <R> , but some time also it is coming with <T> as per our operation staff.

My information that: CSRGV is the VIGV reference angel and as per rung diagram it is the output of the minimum select gate of different inputs (CSRGV manual, part load and temp control).

What I want to say that the inputs to the minimum gate are same in the 3 controllers, why the outputs are not identical? And what are the possibilities of that alarm if any one faced some thing like that before? Also what I can do with this?

One more note which may help: the unit was overhauled before 2 weeks only, but actually that alarm was coming also before the overhaul.

Thanks and regards.

 

Hello... i also have gas turbine power plant with mark IV controller. if there's alarm on CSRGV, actually it will affect on FSR calculation. so we should not be suprised with the next alarm.
CSRGV trouble can be caused by the LVDT feed back voltage (CSRGV_V) that indicate the position of IGV. this voltage is calculated on HSAA card to produce servo current to actuate the servo valve. my suggest :

1. check the LVDT feed back voltage. and re-adjust the voltage
2. check the servo current. if the current is too high ( more than 2.5 mA) it's mean the actuator cylinder is stick and you must replace it with new one
3. check the wiring from site to mark IV to make sure there's no short circuit
4. check the hidraulic pressure
5. check the filter hidraulic, replace with a new one
5. re- calibrate the IGV


hope this help
< nanang >

 

Dear nanang

thanks for your kind response. i agree with you that all the possibilities you mentioned can affect the value of CSRGV. but if it is affected by any one, the change will happen in the 3 controllers?
why it is coming only in <R>? why miss-match is happening?

 

Hello......

there are only 2 LVDT installed for IGV on frame 9 gas turbine which is transmitted to <R> and <S>. and <T> is high select function that select the higher value between <R> and <S>.

the value of <R> <S> and <T> then communicated to <C>. the value on <C> is the average value of <R><S> and <T>. voting mismatch appear if the value of <R> <S> or <T> is so different with <C>. therefore, if appear voting mismatch only on <R>, you should check the LVDT feed back voltage on site which is transmitted to <R>

so for the first step, i suggest you to check the LVDT feed back voltage.

for minimum stroke, the excitation voltage must be 0.7 V AC rms. and re-calibrate the IGV characteristic
i'm looking forward for your reply ......

 

The originator mentions there are other Diagnostic Alarms that seem to come and go, but that this one is the most prevalent and most persistent.

I've been looking at a couple of Mark IV Speedtronic elementaries examining the CSRGV algorithms. I don't see the IGV LVDT feedback affecting any of the references which feed the MIN SELect block that determines which signal drives CSRGV. The IGV LVDT feedback (usually CSGV) is only used in the servo-output on the HSAA card, to be compared against the reference, CSRGV, to adjust the servo-output current to try to make the feedback equal to the reference. Other uses of CSGV (which is derived from CSGV_V) are for alarm and display. So if the Voting Mismatch is on CSRGV <b>in one processor</b>, as described by the originator, then it's not clear how the feedback could affect the reference in any processor.

The originator needs to tell us what the values of CSRGV are for <R>, <S>, and <T>. Also, what are the values of the signals feeding the MIN SEL block in the algorithm for <R>, <S>, and <T>: TTRXGV, CSRGVPS (are the most common)?

I would also like to know what the values of TNH are for <R>, <S>, and <T>, and what the values of CPD are for <R>, <S>, and <T> when these Diagnostic Alarms are being annunciated. I would also like to know what the values of L4R, L4S, and L4T are when these Diag. Alarms are being annunciated.

TNH because it feeds CSRGVPS and CPD because it can feed TTRXGV, depending on the vintage of the Mark IV. If these values are not similar in <R>, <S>, and <T> they can drive other signals to be out of line with each other as well. There are usually Diagnostic Alarms for these signals as well, so that's why it would be helpful to know what all the Diagnostic Alarms are.

We also don't know if this occurring at Base Load or Part Load (if so, at what load(s)). We don't know what the servo-valve output currents are for each of the processors when this alarm is occurring.

The suggestion to adjust LVDT voltage and recalibrate IGVs (one doesn't calibrate the IGVs; one calibrates the IGV LVDT feedback) <b>without first verifying that a difference between the actual physical position and the reference angle exists that warrants adjustment or re-calibration</b> gets so many people in trouble. And, there are also people who believe that when a servo-valve is replaced the LVDTs must be recalibrated, which is completely unnecessary and can also cause a great deal of trouble for the same reasons.

Before any temperature switch or pressure switch or pressure transducer (transmitter) or limit switch or any kind of mechanical device is "re-"calibrated, the first thing that's always done is to check the current, or 'as found', calibration and then, make adjustments, or, recalibrate, <i>only if necessary</i>. And yet people are continually simply recalibrating LVDT feedback (which means the unit must be taken out of service; which means lost production) <b><i>when there's been no proven reason to recalibrate LVDT feedback</i></b>. (I think it just makes it look like the technicians are doing something. I've been to sites where whenever some similar Diagnostic Alarms are annunciated, the operators immediately start planning for a shutdown for the technicians to recalibrate the LVDTs and/or replace the servo (and recalibrate the LVDTs), it's so engrained in some sites' mentalities. When asked if this resolves the Diagnostic Alarms, the usual response is, "Sometimes." When asked for any other granularity in their response, one usually gets none.)

 

Dear CSA:
thank u for your detailed reply, actually i have one problem. I am maintenance engineer and the reports which are coming to me from operation staff do not contains all these details you want to know. just they tell me there is one alarm like that. But they do no not tell at which load this is happening and what the value of TNH is at this time and....

Anyway, I clearly understand your point, and these data are important for troubleshooting. I will try during the next 2 days to collect more data about the problem.

Some points I can mention now:

The values of the signals feeding the MIN SEL block:

CSRGVMAX max
CSRGVMAN manual
CSRGVPS part speed
CSRGVTC temp control

Values of L4R, S, T, and TNH. CPD I don’t know actually because at the time of the alarm usually I am not attending, but I will try to find that .

I agree with you also know that the LVDT feedback is not used in the calculation of CSRGV value after I gone to the MARK IV elementary drawing.

Thanks again for your help.