Single Core Drop During Gas Valve Manual Movement

J

Thread Starter

JMarkV

Hello All,

We are having an issue on one of our machines that is intermittent, and I am having difficulty narrowing down the root cause, I'm hoping you can help.

Background:

2000-2001 vintage GE 7EA combustion turbines
Mark V controls with HMI
Dual fuel machines
Simple Cycle
Midwestern US

Due to oil quality we have experienced servo failures on our gas valves, typically during cold months. We have been diligently working over the course of a few years to correct this issue and have found the most success using a varnish removal unit that is equipped with a chiller that cools the oil before filtering. With this filtering and regular testing we feel that we are correcting the oil quality issues. But, due to past failures, management requires us to test operate our gas control valves daily.

Issue:

Recently when manually testing a valves movement using the auto calibrate feature, one core will drop out and register "Speed increased more than 5%" for that cores status at the bottom of the window. This condition will happen intermittently, always on the same core, the valve can be placed back into "idle" and then back to "manual" the problem may re-appear, or it may not. The valve will move normally without the issue. But when it is in the core dropped state, the valve will move but will be off by about 15%.

The issue will appear on any of the unit's gas valves, it is not just isolated to one. When in the dropped state, there are no additional diagnostic alarms that appear, the only alarms that are in are the normal ones associated with manually moving gas valves.

Typical servo readings at 50% with no issue:<pre>
<R> <S> <T>
Required 50 50 50
Actual 49.91 49.83 49.9
Servo Current -4.66 -4.68 -4.05</pre>
Same valve with the core dropped<pre>
<R> <S> <T>
Required 50 -25 50
Actual 32.52 32.37 32.5
Servo Current -35.4 56.83 -34.77</pre>
Troubleshooting efforts so far:

As I mentioned, this problem is intermittent so trouble shooting is slow going.

Checked shield connections at Mark V - No change
Rebooted <S> core - No change
Re-seated ribbon cables in <S> core (applied nyogel)- No change
Replaced TCQA board in <S> core - No change

My servo connections in <S> core terminate on my QTBA terminal board, travel to the TCQC board, and then on to the TCQA. If I unplug just the JGG cable running from QTBA to TCQC, then the valves will move just fine and <S> core will not have any information but will not drop out.

My question to the group is, has anyone experienced this before? If so, what did you find? If anyone has any insight that might point me in the right direction it would be much appreciated.

Hopefully I have provided enough information, but if not I'll get whatever info the group deems necessary as long as it is within my capabilities to get.

Thanks in advance.
 
JMarkV,

I believe every AutoCalibrate display shows TNH towards the bottom of the display. What does TNH do when this happens?

Newer Mark V PROMsets used TNH < 27% (I think that's what it was) as a permissive for allowing AutoCalibrate to be enabled. And, there was a logic signal, L97HP0TBYP (that's a zero not a capital O) as another of the permissives for AutoCalibrate enabling.

The turbine shaft speed pick-ups (the "primary" shaft speed pick-ups, 77HC-1, -2 & -3) are connected to the QTBA of <R>, <S> & <T> respectively. Are you cranking the unit during these checks?

If you're not cranking the unit during these checks, is the unit on cooldown (hydraulic ratchet, or slow-roll--a cooldown method only typically used on 7EAs with hydrogen-cooled generators)? If it's happening while the unit is on ratchet, can you observe to see if it's happening when the unit is ratcheting on a forward stroke, or at the very end of a retraction stroke when the forward stroke usually starts for just a few seconds?

If it's happening while the unit is on slow-roll cooldown, it's likely either a problem with <S>'s speed pick-up or speed pick-up wiring, or the speed pick-up is loose or the gap isn't correct. (This could also be a possibility even if the unit has ratchet or some other method of cooldown--a bad speed pick-up or speed pick-up wiring or speed pick-up gap.) You could try swapping speed pick-ups between <R> and <S>, or <S> and <T> (if the wires will reach) and see if the problem follows the speed pick-up.

If not, I'd try swapping the QTBA card with another core, and observing if the problem follows the QTBA or not. If it doesn't, then I'd try replacing the TCQA card (or maybe swapping it first with another processor's to see if the problem follows the card). Finally, you say you've unplugged the JGG cable and not had the problem; try replacing the JGG cable, or swapping it with another processor's JGG cable. If it's still in <S>, then all that's left is the TCQC.

Even the ribbon cables with pull tabs can suffer problems if the pull tabs are not pulled gently and evenly. And if it takes a lot of force to unseat the cable using the pull tab it's very easy to damage the connections between the ribbon cable and the cable end--they're just press fits and there's not usually a great deal of contact between the ribbon cable conductors and the sharp metal edges of the cable receptacles.

Too much grease is worse than no grease; a LOT of sites make the mistake of just slathering grease on the male pins and plugging in the ribbon cable. Usually with not very good results.

That's about all I can think of. A methodical and logical method of troubleshooting. If you have a lot of spares, you can try just replacing printed circuit cards one at a time. But sometimes, cards can get inadvertently damaged because of problems with another card/cable--so that's not my preferred method of troubleshooting. I like moving the card to a different core and seeing if the problem follows the card.

Look forward to hearing about your cooldown method, and what TNH is doing on <S> when this happens. TNH is common to all three processors, so it's kind of odd that this is only happening on one processor/core. But, stranger things have happened.

Hope this helps! Please write back to let us know how you progress in resolving this issue!
 
>I believe every AutoCalibrate display shows TNH towards the
>bottom of the display. What does TNH do when this happens?

I went back and reviewed the data. I have and see nothing other than 0 for TNH during the issue.

>The turbine shaft speed pick-ups (the "primary" shaft speed
>pick-ups, 77HC-1, -2 & -3) are connected to the QTBA of <R>,
><S> & <T> respectively. Are you cranking the unit during
>these checks?

We are not cranking the unit during these checks

>If you're not cranking the unit during these checks, is the
>unit on cooldown (hydraulic ratchet, or slow-roll--a
>cooldown method only typically used on 7EAs with
>hydrogen-cooled generators)? If it's happening while the
>unit is on ratchet, can you observe to see if it's happening
>when the unit is ratcheting on a forward stroke, or at the
>very end of a retraction stroke when the forward stroke
>usually starts for just a few seconds?

This is a good point... I could not say for sure that the unit was not on a ratchet sequence while performing these checks so I forced the unit on cooldown while holding one of the valves at 50%. The core did drop once, but then I could not get it to repeat, the valve moved without issue and held fine with the unit on cool down and ratcheting.

We have an air cooled 7A6 generator.

>If not, I'd try swapping the QTBA card with another core,
>and observing if the problem follows the QTBA or not. If it
>doesn't, then I'd try replacing the TCQA card (or maybe
>swapping it first with another processor's to see if the
>problem follows the card). Finally, you say you've unplugged
>the JGG cable and not had the problem; try replacing the JGG
>cable, or swapping it with another processor's JGG cable. If
>it's still in <S>, then all that's left is the TCQC.
>
>Even the ribbon cables with pull tabs can suffer problems if
>the pull tabs are not pulled gently and evenly. And if it
>takes a lot of force to unseat the cable using the pull tab
>it's very easy to damage the connections between the ribbon
>cable and the cable end--they're just press fits and there's
>not usually a great deal of contact between the ribbon cable
>conductors and the sharp metal edges of the cable
>receptacles.
>
>Too much grease is worse than no grease; a LOT of sites make
>the mistake of just slathering grease on the male pins and
>plugging in the ribbon cable. Usually with not very good
>results.

I understand the issues with to much grease and ribbon cables, I was very careful not to over apply, and handled the cables delicately. I did swap the TCQA card with no change in the issue, I believe I will work down the path of QTBA next to see if that makes any changes to my symptoms. I will report back with what I find.

Thanks again.
 
JmarkV,

Thanks for the feedback!

>This is a good point... I could not say for sure that the
>unit was not on a ratchet sequence while performing these
>checks so I forced the unit on cooldown while holding one of
>the valves at 50%. The core did drop once, but then I could
>not get it to repeat, the valve moved without issue and held
>fine with the unit on cool down and ratcheting.

I'd say try swapping speed pick-ups at the QTBAs (move <S>'s to <R>, and vice versa, or whatever works easiest from a wire length standpoint). You could also "swap" them at the JB closest to the #1 bearing, or a main JB close by (I always forget if JB1 or JB2 is high level or low level!).

Please continue to keep us updated!
 
>I'd say try swapping speed pick-ups at the QTBAs (move <S>'s
>to <R>, and vice versa, or whatever works easiest from a
>wire length standpoint). You could also "swap" them at the
>JB closest to the #1 bearing, or a main JB close by (I
>always forget if JB1 or JB2 is high level or low level!).

I had the chance to get back on this today.

I moved the speed pickup between cores by swapping the wires at the closest JB to the #1 bearing. For me that was JB 20BA located on the air inlet plenum wall. This swap made no change in my conditions.

Next I swapped TCQC boards with a new board from our inventory. Also made no change in conditions.

My next move will be to change the JGG cable if we have one, and then the TBQA terminal board. I'll report back with my results.

Thanks.
 
> and then the TBQA terminal board. I'll report back with my results.

I just noticed I made a mistake here, I'll be looking to replace the QTBA terminal board... not the TBQA. Sorry, I should have paid more attention when I made my post. How embarrassing.
 
Hi All,

Thought I would stop back in and update everyone with the status of this issue. None of my troubleshooting efforts yielded any results. However, when the ambient temperatures in the area began to increase the core drop events frequency began to drop. Now we have went weeks without a single event. This leads me to believe that the condition was most likely related to the oil quality issue, since the oil quality problem presents itself during cold weather.

Like I mentioned in my first post we have been diligently working towards correcting or oil cleanliness issue and I think we are on the right path.
 
Do you have working oil reservoir heaters in your control oil system?

What is the operating temperature and what is the heater temperature switch set to?

There is a minimum operating oil temperature so that cold weather should not be a factor and water in oil is minimized by eliminating reservoir breathing and condensation.

good luck
 
JMarkV,

Thanks very much for the feedback!

So, I have re-read and re-re-read and re-re-re-read your original post and the replies and I am having a VERY difficult time relating oil temperature and speed. Your original post stated that when the core "drops" it reports "Speed increased more than 5%" as the reason for "dropping" (a term I really am not all that familiar with, but kind of have a sense of what it is you're trying to describe).

Can you understand my dilemma? Could it be that the temperature in the area where the Mark V is located isn't being kept at a constant temperature during cold weather months for some reason? A LOT of people mistakenly believe the air conditioners in the PEECC are for human comfort--and that's NOT true. They're primarily for humidity control, and secondarily for temperature control.

I can understand that there seems to be a correlation between ambient and L.O. Tank temperature and cores "dropping" due a turbine shaft speed change, but I just find it very difficult to find anything which would relate turbine shaft speed and ambient or L.O. Tank Temperature.

Now, having said that, it's entirely possible that the TCQA PROMs in use on the units at your site have a bug in the software. Usually, when LVDT feedback doesn't change at the expected rate during an AutoCalibration procedure, it reports that with a message something to the effect that the LVDT feedback isn't changing at the expected rate (I don't recall the exact error message). It could be that cold oil is causing the actuator to be sluggish which is causing LVDT feedback to be sluggish and instead of the TCQA PROM software reporting it as a slow LVDT it's reporting it as a change in speed. (BOTH conditions would result it a core "dropping" in my recollection.)

That would "explain" why the Mark V is reporting an error (the WRONG error in this case if the real problem is cold oil and sluggish actuators/LVDT feedback) because of slow actuator response and LVDT feedback because of cold oil. Remember, the passages in the servo valve are pretty small and it's probably partially due to the cold oil causing the servo spool piece to be sluggish.

Have you tried insulating the servo valves? They don't have to be heat-traced, but maybe just putting a small blanket of insulation on them would help--of course you won't know if that works until the weather/L.O. temp gets cold again.

I would also second dwpatterson's recommendation to ensure the L.O. tank temperature is kept up. When the tank heaters are energized the Aux. L.O. Pump is started and that should circulate oil to all the off-base skids and to the IGVs. And, the temperature switch, 26QL-1, that starts/stops the Tank Heaters usually has about a 10 deg F deadband. In my personal opinion the GE-recommended setting for that switch is way too low. I always recommended that sites increase the setting to turn on the heaters when the tank temperature dropped below 80 deg F, and the 10 deg F deadband would shut of the heaters when the temperature got above approximately 90 deg F. Some people insulated the L.O. Tank, and that would be GREAT for keeping the tank temperature more constant, BUT that doesn't keep the oil in the lines to/from the off-base skids and the IGVs warm (unless they are insulated, and heat-traced where exposed to ambient temperatures).

Finally, I have written extensively on control.com about alleged servo problems attributed to L.O. varnishing. GE and one of their Customers in the UK in the early 2000's worked together with BP-Castrol to determine what was causing a high number of servo problems. It was determined that around 2000 that oil refiners made a significant change to turbine lube oil formulations that resulted in better lubricity and longer lube oil life. BUT, that change wasn't so good for turbines that used lube oil for hydraulic fluid. BP-Castrol changed their oil formulation and that solved the site's servo failure problems right away.

You can see that GE doesn't really want to tell people this--they have been using lube oil for hydraulic fluid on their heavy duty gas turbines for decades with great success. That is, until the oil refiners changed their formulations. GE doesn't want to tell people you should strongly use this or that manufacturer's lube oil for any specific reason--especially because of problems with the hydraulic system (the servos). Other gas turbine OEMs have criticized GE for decades for not having a separate hydraulic fluid and system. And, so, admitting this is a problem would be problematic for GE. (As it turns out, they are now supplying electrically-actuated fuel control valves and even IGV actuators now.... which conveniently negates the problem created by the new lube oil formulation!)

Anyway, hope this helps in some small way. I just find it very difficult to relate ambient/hydraulic fluid (lube oil) temperature and turbine speed. So, that's why I wonder if that TCQA PROMset doesn't have a bug.... And, I would expect that GE doesn't even know about that little bug, because most people do not pay attention to Diagnostic Alarms or alarms which are annunciated by AutoCalibrate. Further, GE doesn't support Mark V any longer (shame on them) and so even if they were made aware of the problem and could verify it they probably wouldn't issue an updated TCQA PROMset.

Please do continue to keep us informed! And, thanks for doing so!
 
Years ago, I had a recommendation to short the UNUSED (LP) Speed inputs during autocal if this happens.
 
> Do you have working oil reservoir heaters in your control oil system?

Yes we do, 23QT 1&2.

> What is the operating temperature and what is the heater
> temperature switch set to?

70 deg. F decreasing, 84 deg. F increasing

> There is a minimum operating oil temperature so that cold
> weather should not be a factor and water in oil is minimized
> by eliminating reservoir breathing and condensation.

This is true, the low lube oil temp alarm is set to 60 deg. F. However, the servos that experience the varnishing issues are in the gas control valve compartment, which is a separate compartment from the accessory compartment which houses the lube oil tank reservoir. I have heat traced the lines that feed trip oil and hydraulic oil to the gas compartment to keep them at 70 deg. F. But when the units are operating the gas skid vent fans keep the compartment under negative pressure which pulls cool air into the area.

> good luck

Thanks, I think we are on the right track with getting the oil filtered out. We've even purchased a second filtering skid with a cooler to increase the amount of oil we can filter. Servos need super clean oil!
 
anon,

>Years ago, I had a recommendation to short the UNUSED (LP)
>Speed inputs during autocal if this happens.

GE did recommend that if there was no LP shaft on a turbine that the <P> core LP shaft speed pick-ups be jumpered to prevent nuisance Diagnostic Alarms. In fact, towards the end of the Mark V production the factory was automatically installing hardwire jumpers in the LP shaft input terminals (PTBA-3 to PTBA-4; PTBA-7 to PTBA-8; PTBA-11 to PTBA-12).

I am so accustomed to seeing jumpers at the terminals I just completely forgot to ask if they are installed when a question like this comes up. I just assumed they were there; this could, indeed, be related to the problem--though it wouldn't explain why when the ambient and lube oil temperatures warm up the problem isn't so prevalent.

Anyway, thanks for the reminder!
 
Hi again,

Thought I would reply to CSA here and provide an update. We have been running constantly and I have not had the chance to sit down and type out a response until now.

>I can understand that there seems to be a correlation
>between ambient and L.O. Tank temperature and cores
>"dropping" due a turbine shaft speed change, but I just find
>it very difficult to find anything which would relate
>turbine shaft speed and ambient or L.O. Tank Temperature.
>
>Now, having said that, it's entirely possible that the TCQA
>PROMs in use on the units at your site have a bug in the
>software. Usually, when LVDT feedback doesn't change at the
>expected rate during an AutoCalibration procedure, it
>reports that with a message something to the effect that the
>LVDT feedback isn't changing at the expected rate (I don't
>recall the exact error message). It could be that cold oil
>is causing the actuator to be sluggish which is causing LVDT
>feedback to be sluggish and instead of the TCQA PROM
>software reporting it as a slow LVDT it's reporting it as a
>change in speed. (BOTH conditions would result it a core
>"dropping" in my recollection.)
>
>That would "explain" why the Mark V is reporting an error
>(the WRONG error in this case if the real problem is cold
>oil and sluggish actuators/LVDT feedback) because of slow
>actuator response and LVDT feedback because of cold oil.
>Remember, the passages in the servo valve are pretty small
>and it's probably partially due to the cold oil causing the
>servo spool piece to be sluggish.

The above explanation is what I am leaning towards as an answer to my condition, though I lacked the words to describe it as well as CSA. I am thinking that the varnish-like substance that drops out of our oil in cold weather is partially clogging the pencil filter inside the servo reducing flow and causing sluggish movement of the valve. As I began to review my notes, I noticed that the only unit that did not exhibit the core dropping issue was the unit that we had cleaned the oil up on the most with our polishing filter skid. When I combine this observation with the fact that the core drop events have stopped with warmer ambient weather, I'm inclined to believe I have solved the cause of the issue.

>Have you tried insulating the servo valves? They don't have
>to be heat-traced, but maybe just putting a small blanket of
>insulation on them would help--of course you won't know if
>that works until the weather/L.O. temp gets cold again.

I have made considerable efforts to keep the area heated when the units are offline (We are a peaking station) as well as heat tracing and insulating the oil lines that feed the gas skid. I will also investigate the possibility of making some blankets for the servos as you suggest, I think that is a good idea.

>I would also second dwpatterson's recommendation to ensure
>the L.O. tank temperature is kept up.

I definitely do this!

Thanks for all the help.
 
Top