Gas Turbine Startup Problem

V

Thread Starter

Vinayak

We are having Frame 6B Gas Turbine. Since last 3 days we are trying to start the machine. However IGV is not at all opening during the startup and subsequently machine is tripping.

Conventional trouble shooting actions like calibration, moog valve replacement, Hydraulic cylinder replacement etc done. IGV operates very smoothly in calibration mode but hardly move during actual startup sequence.

We are facing such problem first time.

Has anyone ever faced similar problem and what are the possible reasons for this?
 
During start-up the IGVs should start opening at approximately 80% speed. The reference is (or should be) CSRGVPS which uses temperature-corrected speed (TNHCOR)--which is derived from the axial compressor inlet thermocouples (which is CTIM).

What Process- and Diagnostic Alarms are present and active PRIOR to the start, and what Process- and Diagnostic Alarms are annunciated just prior to tripping?

If the IGVs move during manual stroking there should be no reason for them not to move during starting. Unless TNHCOR (and/or the axial compressor inlet T/CS) are not correct. Or someone has changed the part-speed IGV Control Constants.

If the Speedtronic is TMR, then there should be no issues with the part-speed IGV control reference (CSRGVPS) that isn't accompanied by alarms (Process- and Diagnostic).

The only other issue I could think of would be a problem with the IGV Trip Oil dump valve--but that doesn't make sense because it should exhibit the same problems during manual stroking.

What signals are forced to get a "READY TO START" before initiating a START?

What kind of control system and is it TMR or SIMPLEX?

If you're really serious about getting help solving the problem you will need to answer <b>ALL</b> of the questions above. Please list all alarms in chronological order. Without <b>ALL</b> of the requested information it will <b>NOT</b> be possible to help.

Please write back with <b>ALL</b> of the requested information.

If you have the ability to trend data, please trend the following signals during a START attempt:

TNH
TNHCOR
CTIM
CSRGVPS
CSRGV
CSRGVTC (IGV temperature control reference)
CSRGVMAN (manual IGV temperature control reference)
TTXM
CAGV

We need to know the value of these signals at 70% TNH and at 75% TNH and just prior to the trip.

Also, list all signals which are forced (to "1" or "0").

Lastly, the IGVS position reference (CSRGV) is the high-selected value of CSRGVPS, CSRGVTC, and CSRGVMAN. Some Speedtronic panels may another input to the high-select block. Find the high-select block that chooses the value for CSRGV in the Speedtronic panel at your site and then trend all of the inputs to the block during starting to see what's happening.

And let us know what you find.
 
Thanks CSA for the reply. All the parameters were checked by us and were normal. The control actions were normal except IGV was not opening actually.

The IGV opened at last after various actions like replacing moog valve, replacing Hydraulic cylinder, checking hydraulic block, draining the lines, replacing Hydraulic filter etc.

There is no clear conclusion about why IGVs were not opening and also why it opened later.

There was similar thread in 2012 for 9FA machine, which was also similar to our problem and solved in similar way.

All we can say, that such problems give us opportunity to sit and go to the basics and learn from you.

Thanks for your prompt response and sharing
 
Happy to knew your problem is solved.

But if IGV was operating well during manual stroke and strange why didn't operate while starting the machine?

Just to learn, what alarm caused to trip the machine?

Take care
G.Rajesh
 
G.Rajesh,

You make an excellent point: If it was possible to stroke the IGVs manually, then there should have been no problem with the Speedtronic moving the IGVs during operation.

The Shotgun Approach to troubleshooting (as was done in this case) does not lend itself to understanding why the IGVs weren't moving. If the Speedtronic could manually move the IGVs (meaning that not only could the Speedtronic put out sufficient current to move the IGVs but that <b>the amount of current required was not excessive and the LVDT feedback from the IGVs during manual stroking was nearly equal to the position reference and did not differ from the position reference by more than approximately 0.5 DGA or so)</b> then it's just simply not conceivable how the IGVs could NOT move when the turbine was starting.

To open the IGVs, it's usually necessary to force L20TV1X (or sometimes L20TVX1) and sometimes L4X1 (for Mark IV Speedtronic turbine control systems) to energize 20TV-1 to get high-pressure hydraulic oil to the IGV actuator through the IGV hydraulic dump valve. And, to open the fuel stop valve it's also usually necessary to energize 20TV-1, so it would seem that 20TV-1 and the IGV hydraulic dump valve were both working fine.

That leaves the servo-valve--but it was (seemingly) working fine during manual stroking--<b>unless the current required was much more than normal to make the IGVs move, and that would mean that the LVDT position was much less than position reference during manual stroking--neither of which was reported.</b> So, if the servo was working properly (and that's not known) then the actuator shouldn't have been the culprit because it will move the IGVs when high-pressure hydraulic oil is ported to it via the servo-valve.

And, there are two LVDTs, and the Speedtronic takes the higher of the two values at all times (during manual stroking and during normal operation and starting) so they shouldn't have been the problem, either, as they were reported to have been working fine during manual stroking.

My guess is at this point that someone had used the Manual IGV Control function that some operator interfaces have and left the reference at some very low position. AutoCalibrate would bypass that during manual stroking, but would not during starting/normal operation.

But, we'll never know. Sometimes, we just have to live with the information presented. If the Speedtronic processors were re-booted at some point during the Shotgun Troubleshooting approach (which was not reported) then it's likely that the Manual IGV reference was returned to the normal default value--which is what caused the IGVs to being working again.

But, again--we'll never know. It's running now--and that's all that matters at most sites.
 
Dear CSA,

Since machine started running, we safely assume that the unknown problem is solved.

But a technical person is never happy even if problem is solved, if root cause is not identified. This is true for us too.

IGV was moving in calibration mode and not opening during starting. The points made by CSA were already checked and IGV was not left in manual mode.

The assumption by most of our team members were that the hydraulic system was capable of moving IGV in no flow and was unable to exert enough pressure against compressor air flow.

So one or more action among various action like moog valve, cylinder, filters replacement and line flushing, may have clicked.

But believe, nobody is comfortable, since we don't know why IGV was not opening in so many trials.

Thanks once again.
 
Vinayak,

Have you disassembled the hydraulic cylinder to inspect it? Some actuators had a problem with the piston ring being much harder than the cylinder bore, and gouging the cylinder wall. The IGVs would work normally during calibration, but did not open under the IGV aerodynamic loading during operation. (The air loads try to drive the actuator in the closed direction during operation.)

> But a technical person is never happy even if problem is solved, if root cause
> is not identified. This is true for us too.
 
A

Andrew Turner

Vinayak,

It is very important now to check that the IGV position is stable during running. You can check this by using a high speed log to monitor its operating position. Frame 6 have a common problem where the hydraulic cylinder shears off the dowels which keep it located. Were these intact when you replaced the cylinder? It is very important to check the stops are correctly positioned. That is, from memory, software stops of 36, 86. Cylinder stops of 34, 88 and IGV ring stops of 32, 90. This means that the cylinder should drive to its full extent in either direction before the mechanical stop on the IGV ring comes into effect. To test this drive the cylinder fully one way or the other then you should be able to remove a sheet of paper placed in the ring stop. It is quite common that this ring stop comes into effect before the cylinder stops driving and this means the full force of the cylinder is driving against the holding down bolts which were never designed to take it, hence the locating dowels can shear off. Please check this and if you need further help get back to me.
 
The position was checked physically and there was physical gap. The cylinder was not opened to check. It was not opened, because problem was continued with new cylinder also.
 
Vinayak,

Contrary to popular belief the IGV actuator is actually trying to keep the IGVs from opening during normal operation (during starting and loading). Air flow through the IGVs of <i>MOST</i> GE-design heavy duty gas turbines is always trying to open the IGVs--and the IGV actuator is trying to maintain the desired position by trying to keep them closed.

So, the assumption that the hydraulic actuator was unable to open the IGVs against air flow is simply not based in fact, and was, in fact, just a (bad) guess.

And, again, it is so unlikely that the IGVs could open further than the mechanical stop--unless the mechanical stop was broken (as suggested). Further, air flow through IGVs that were open more than 90 DGA would be putting a huge force in the opposite direction than normally applied--which could have very serious knock-on effects on IGV bushing, gears and blade integrity. If the problem was caught early enough, maybe no damage was done. But, sucking IGVs into a running turbine (at approx. 5100 RPM) causes "corn cobbing"--which is very ugly, and very costly. (When a lot of axial compressor blades are broken off the compressor rotor, when the rotor is hanging from the crane during removal from the unit it looks like an eaten ear of corn--a corn cob. It's gnawed down to an uneven surface and tapered like an eaten corn cob.) This is one reason there is supposed to be mechanical stops to prevent the IGVs from opening past approximately 86-88 DGA (depending on machine configuration) and thereby preventing damage from reverse forces being applied to the IGV blades, bushings, and gears.
 
V

vinayak_paranjape

We have closed this problem with following learning:
There was some kind of jamming in rack and pinion. In calibration mode it was operating nicely, but due to some force in running condition, rack or pinion must be getting pressed against each other hence requiring more force to operate.

Whenever we share such problems, due to rich discussions on the topic we get deep knowledge. Thanks to CSA especially.
 
Top