Mark VI gas control fault

W

Thread Starter

Wahidul Islam

pg-6541 gas turbine (mark-4 speedtronic).

''gas control fault'' =L3GFLT shows at the alarm section.
L3GFLT (software relay) has to be energized only when L3GCVFLT or L3GRVFLT or L3IGVFLT any one of these three is energized. but the problem is none of these three relays is energized. but still L3GFLT/ gas control fault is showing '1' instead of '0'.

it is to say that at logic forcing section we see 'R' core is '1', 'S' core is '1', 'T' core is '0'. when i disable the R of S core the fault is gone. i am thinking that the problem is in R and S core control cards.

can anyone answer me that am i right or wrong? if right then can you tell me which card is responsible for this false fault? if wrong then what is the fault do you think?

my email address([email protected]), if possible send me the documents including which control cards control which thing..

thank you
 
Without being able to examine the sequencing running in the Mark <b>IV</b> at your site it's impossible to say for sure what's happening, but I'm pretty certain that L3IGVFLT ('IGV Fault') shouldn't be one of the conditions which would cause a 'Gas Control Fault' alarm, unless someone has been mucking with sequencing for some very odd reason and without understanding how turbines work.

<b>WHEN did this problem start?</b> After a Maintenance Outage, when someone was trying to calibrate gas valve LVDTs?

After a trip (emergency shutdown) from load?

After some maintenance on a servo-valve or gas valve? Which valve?

After a card was replaced in one or more processors? Which card and which processor(s)?

What Diagnostic Alarms are active on the Mark IV?

What other Process Alarms are active on the Mark IV?

Usually, a Gas Control Valve Fault (L3GCVFLT) or Stop-Ratio Valve Fault (L3GRVFLT) is annunciated when one of three conditions for either valve is detected:

1) Servo-valve current during shutdown is more than a setpoint

2) Valve position is more negative than it should be during shutdown

3) I'm sorry, but I forget the third (I don't have access to a Mark IV Speedtronic elementary at this writing), but I think it might be that the valve is more open than it should be when it's closed--but I may be wrong about that without being able to see an elementary

You should be able to work backwards from L3GRVFLT and L3GCVFLT to find which one of the three conditions for each of the valves is causing the respective alarm to be annunciated.

The Mark IV uses one LVDT for one of the three processors, the second LVDT for the remaining two processors, and the third processor compares the two LVDT feedbacks and uses the higher of the two as its value. This is all set by the positions of Berg jumpers on the HSAA cards. Unfortunately, quite often the Berg jumper positions on the Speedtronic elementary sheets are NOT shown in the proper positions as specified on the "B" sheets of the Mark IV Speedtronic elementary.

I suspect that for some reason, one of the LVDT feedback signals from either the SRV or GCV is significantly different from the other LVDT feedback value from the same valve. This can also lead to improper servo current. But that's just a guess based on the information provided.

I would suggest the problem is NOT a Speedtronic card--UNLESS one or more HSAA cards was recently changed in the panel. And, then I would suggest that the Berg jumper positions (all of them) on the new card be compared to those on the card which was removed to make sure the Berg jumper positions on the new card are EXACTLY like the positions on the card which was removed (and was working fine before it was removed!).

Most problems like this are NOT Speedtronic card problems, unless there are Diagnostic Alarms which exist that are related to the circuits involved and which we haven't been told about. Diagnostic Alarms are indicators of Speedtronic card and -firmware problems and are not just nuisance alarms, contrary to popular lore and belief.

Please write back to let us know how you fare.
 
Dear Wahidul,

this is a very interesting problem you are encountering. I have reviewed some old MKIV prints I have and am able to "see" the logic you speak of pertaining to the alarm L3GFLT.

This alarm should be annunciated if the MKIV identifies a problem with the gas control valve, gas ratio valve or interstage pressure transmitter. But one of the other signals you spoke of would need to be true, L3GCVFLT or L3GRVFLT or L3IGVFLT. So I like you am at a loss.
I would be curious to know when this problem began. Is it possible that there were some changes to either of the gas valves that may have changed the "calibration" of the LVDT's? That is where I would be focusing in this case. The HSAA card should be the card that is handling LVDT excitation and servo valve outputs, and the HAIC card for the interstage pressure transducer.

Please write back and let us know how you proceed.
 
Dear Wahidul,

After reading CSA's post I reviewed my MKIV elementaries and I think the last signal you are looking at is L3GIVFLT, not L3IGVFLT. I apologize for not catching this earlier, I just did a copy/paste of your post for the alarms without fully reviewing my prints. If I understand correctly this is for the GasInterstageValveFault, not to be confused with the IGV, InletGuideVaneFault.

Either way one of these 3 signals needs to be true for the alarm L3GFLT to be true.
 
W

Wahidul Islam

thanks...here more.
After a trip (emergency shutdown) from load this signal occurs. but i'm telling you that none of the 3 signal are active (at '1') except the L3GFLT. there is no other alarm except wheelspace temp. high.
 
Wahidul,

We are telling you there are three conditions which can cause EITHER a SRV fault, or a GCV fault, and then there's the GIV fault. Any one of these seven (7) conditions will cause L3GFLT to go to a logic "1". You are ONLY looking at the general, aggregate alarm (L3GFLT) for any of as few as seven (7) conditions. You need to review the Speedtronic Elementary for the names of the seven signals (minimum) that could cause the Gas Valve Fault alarm to determine which one of them is causing the alarm.

NOTE: Almost all of these conditions are reset (when resolved or cleared) by a Master Reset, so when the alarm first occurs you must NOT immediately initiate a Master Reset in order to keep the alarm condition "latched in" so you can determine which one of the several alarms is causing the one Gas Control Fault alarm.

You have NOT told us what cause the emergency shutdown (trip), and whether this Gas Control Fault comes BEFORE or AFTER the trip.

So, while you can keep focusing on L3GFLT being a one in two of three processors, you will not likely find the problem until you widen your focus and find out which one of at least seven conditions is causing the problem. You have to work "backwards" through the sequencing to get to the root cause of the problem which is what it seems you have not yet done.
 
Dear Wahidul,

It is difficult to assist you since I cannot view the your logic exactly as it is running, so my suggestions are based on assumptions of your logic. The MKIV elementary drawings I am reviewing are "typical" for GE design heavy turbines, but may differ from your exact logic.

Are you absolutely positive that there are no other alarms active prior or immediately after the trip?

Is your alarm dot matrix printer working and noting all alarms in correct order?

Is someone issuing a master reset immediately after the trip?

As CSA mentioned, upstream of the signal L3GFLT are 3 signals that can trigger the alarm. Following further upstream are 7 set/reset blocks that can set this alarm. All of these blocks are latched and should stay true until after a master reset is issued. Again this all assumes that the logic I am referencing is the same as yours, and that just may not be the case.

But without being at your site this is all the help I can offer. My guess is this alarm you are seeing may be a "result" of the trip, not the cause. But we need more information to be able to help much more.
 
Wahidul,

I found an electronic copy of a Mark IV Speedtronic Elementary on an old hard disk; most Mark IV Elementaries were very, very similar, but some were slightly different. This should help you in your search.

On Sh. 30J there is a block called MK4_L3GFLTV1_00XX, 'Gas Fuel Control Fault'.

The block has eight (8) logic outputs based on some comparisons done using Control Constants as setpoints/parameters against five (5) input signals: FSGR (Stop-Ratio Valve LVDT feedback); FAGR (Stop-Ratio Valve servo current); FSG (Gas Control Valve LVDT feedback); FAG (Gas Control Valve servo current); and FPG1 (sometimes it's FPG2 on some Elementaries; sometimes it's just FPG on others... GE is wonderfully consistently inconsistent an things like this).

The eight logic outputs (on this elementary) are:

L60FSGH - Start-up Fuel Flow Excessive
L3GFLT1 - SRV LVDT Feedback less than LK3FGLT1
L3GFLT2 - SRV LVDT Feedback greater than LK3FGLT2
L3GFLT3 - FAGR greater than LK3FGLT3
L3GFLT4 - SRV LVDT Feedback less than LK3FGLT4
L3GFLT5 - SRV LVDT Feedback less than LK3FGLT5
L3GFLT6 - FAGR greater than LK3FGLT6
L3GFLT8 - P2 (Gas Intervalve) Pressure less than LK3GFLT8

The rungs below show how the logics drive L3GFLT from Sh. 30K:<pre>
L3GFLT1 L3GRVFT
----| |---------------------------------( )
|
L3GFLT2 |
----| |-----
|
L3GFLT3 |
----| |-----

L3GFLT4 L3GCVFT
----| |---------------------------------( )
|
L3GFLT5 |
----| |-----
|
L3GFLT6 |
----| |-----

L3GFLT8 L3GIVFLT
----| |---------------------------------( )

L3GRVFLT L3GFLT
----| |---------------------------------( )
|
L3GCVFLT |
----| |-----
|
L3GIVFLT |
----| |-----</pre>
What you need to do, Wahidul, is look at the three signals that "drive" L3GFLT and determine which one of them is a logic "1". Let's say it's L3GRVFLT. Then you need to look at the three signals that drive L3GRVFLT and see which one of them is a logic "1". Let's say it's L3GFLT1; that would mean that the SRV LVDT feedback is less than LK3GFLT1, which is usually a number around -5.0%. So, if you looked at FSGR and it was -5.19%, then the problem would be the LVDT feedfback from one of the SRV LVDTs was more negative than it should be and requires some investigation and resolution.

So, just because L3GFLT is a logic "1", that's not the end of the story. You need to determine <b>WHY</b> L3GFLT is a one, and then get to the bottom of that, which requires a little more "working backwards" to determine.
 
Top