GE Frame turbines Over Speed Testing

Hi All,
I Would like to know how individual’s carryout the Gas and Steam turbine overspeed test!

Here in Australia, We operate number of frame 9Es , 6Bs and a GE C7 steam turbine. The current practise is actual overspeed tests on Frame 9E units which was part of the proof testing requirement as per SIL document. For Steam turbine, there is no any SIL requirement and for 6Bs no testing has been done due to less utilization.

Recently GE issued a product service bulletin PSIB 20200911A saying units (both GT and ST) that does not have a mechanical overspeed test can conduct a shutdown overspeed test at 55% speed. This is given the fact that units are having Mark VI or Vie operating system and shut down overspeed test function installed in control system. Our units comply with the first two requirements, but not the shut down test function. OEM price for the test function implementation is huge; hence would like to find about alternatives. I have seen some old threads talking about what different users do (2of them) and would like to know how other users treat the PSIB and how they practise the test .

PSIB is attached!

Thanks
 

Attachments

Yoganak, I operate and maintain 2 Frame 7EA machines and 2 Frame 5 machines. The 7EA has no O/S bolt, the Frame 5 units DO have OS bolts.
We upgraded the control system for the 7EA machines in 2018 from MKVI to MKVIe and included with the hardware upgrade was the newer application code that includes logic and graphics for what GE calls the shutdown overspeed test. The Frame 5 units have a MKVI control system that does not include logic for a shutdown OS test.
I think you will find procedures from users that are all over the board for how they perform an OS tests. I think at the end of the day it all depends on what your insurance company will accept. Some people use a frequency generator to test, some change setpoints for the main controller, some for the protection controller, and some are very different for systems that have migrated away from a GE control system and are using something else.
Before upgrading the control system on the 7EA machines we would perform an OS test with the unit firing, but would reduce the setpoint of the main controller that would allow the machine to trip a 103% rather than 110%. We did not test the backup overspeed protection system setpoint. With the newer shutdown OS test logic for our 7EA we perform a SD OS test once a year that trips the machine as it is shutting down at 60% speed.
The Frame 5 units we operate are not currently tested for overspeed protection.
 
Yoganak, I operate and maintain 2 Frame 7EA machines and 2 Frame 5 machines. The 7EA has no O/S bolt, the Frame 5 units DO have OS bolts.
We upgraded the control system for the 7EA machines in 2018 from MKVI to MKVIe and included with the hardware upgrade was the newer application code that includes logic and graphics for what GE calls the shutdown overspeed test. The Frame 5 units have a MKVI control system that does not include logic for a shutdown OS test.
I think you will find procedures from users that are all over the board for how they perform an OS tests. I think at the end of the day it all depends on what your insurance company will accept. Some people use a frequency generator to test, some change setpoints for the main controller, some for the protection controller, and some are very different for systems that have migrated away from a GE control system and are using something else.
Before upgrading the control system on the 7EA machines we would perform an OS test with the unit firing, but would reduce the setpoint of the main controller that would allow the machine to trip a 103% rather than 110%. We did not test the backup overspeed protection system setpoint. With the newer shutdown OS test logic for our 7EA we perform a SD OS test once a year that trips the machine as it is shutting down at 60% speed.
The Frame 5 units we operate are not currently tested for overspeed protection.
Dear MIKEVI, Thanks and appreciate your response! Do you see any specific criteria that is getting tested by doing it in a shutdown and 60% speed?
 
Yoganak, I may not be fully understanding your question this time, but I will give it a go.
I understand with the older systems that used a mechanical overspeed bolt there was no way to test its function except to increase turbine rotor speed until the bolt actuated and released trip oil pressure causing the fuel or steam valves to close. With the introduction of digital control systems there is/are now a number of electrically operated trip oil control valves that dump trip pressure when deactivated and cause the fuel or steam valves to close.
Testing of these systems has gotten creative in my opinion, but again it is typically whatever satisfies the insurance company or the people above me making the decisions.
In my opinion a real test should include all the components that are part of the speed sensing system. The toothed wheel on the turbine rotor, speed pickups, wiring, speed cards, application code logic, electrically switched outputs, the electrically operated trip valves, and trip oil system and how it controls the fuel or steam valves. A complete test uses all these components to sense and act on an overspeed event and may include primary and backup overspeed devices as well.
Many users over the years have adapted their own testing procedures, typically to eliminate the need to actually operate the machine at speeds greater than nominal design, and also to eliminate the need to actually trip the machine while burning fuel or operating on steam. These procedures were adapted to reduce stress to the machine and reduce the chance for a mechanical failure that might cause injury to staff should a catastrophic failure occur.
I can only guess why GE chose the particular shutdown test parameters that they are noting in the PSIB. Letting the gas turbine or steam turbine shutdown normally and then trip at a value that is consistent reduces rotor and other thermal stresses. The logic and graphics make it possible for an operator to select the shutdown trip test, and then put the unit into a shutdown and watch it trip at 55%. No signal forcing or controls assistance needed to run the test. There is no chance for a mistake to be made when forcing logic etc. or forgetting to move a control constant back to its original value.
I think there are many other ways to test the system that may be better, or not. I am glad that GE provided guidance and proposed an option (even though it is expensive) that reduces machine stress and is a validated procedure.
I think if you have OS test procedures that work for you, are consistent, are approved by insurance and owners, then you don't necessarily have to go out and spend money to upgrade.
 
Yoganak, I may not be fully understanding your question this time, but I will give it a go.
I understand with the older systems that used a mechanical overspeed bolt there was no way to test its function except to increase turbine rotor speed until the bolt actuated and released trip oil pressure causing the fuel or steam valves to close. With the introduction of digital control systems there is/are now a number of electrically operated trip oil control valves that dump trip pressure when deactivated and cause the fuel or steam valves to close.
Testing of these systems has gotten creative in my opinion, but again it is typically whatever satisfies the insurance company or the people above me making the decisions.
In my opinion a real test should include all the components that are part of the speed sensing system. The toothed wheel on the turbine rotor, speed pickups, wiring, speed cards, application code logic, electrically switched outputs, the electrically operated trip valves, and trip oil system and how it controls the fuel or steam valves. A complete test uses all these components to sense and act on an overspeed event and may include primary and backup overspeed devices as well.
Many users over the years have adapted their own testing procedures, typically to eliminate the need to actually operate the machine at speeds greater than nominal design, and also to eliminate the need to actually trip the machine while burning fuel or operating on steam. These procedures were adapted to reduce stress to the machine and reduce the chance for a mechanical failure that might cause injury to staff should a catastrophic failure occur.
I can only guess why GE chose the particular shutdown test parameters that they are noting in the PSIB. Letting the gas turbine or steam turbine shutdown normally and then trip at a value that is consistent reduces rotor and other thermal stresses. The logic and graphics make it possible for an operator to select the shutdown trip test, and then put the unit into a shutdown and watch it trip at 55%. No signal forcing or controls assistance needed to run the test. There is no chance for a mistake to be made when forcing logic etc. or forgetting to move a control constant back to its original value.
I think there are many other ways to test the system that may be better, or not. I am glad that GE provided guidance and proposed an option (even though it is expensive) that reduces machine stress and is a validated procedure.
I think if you have OS test procedures that work for you, are consistent, are approved by insurance and owners, then you don't necessarily have to go out and spend money to upgrade.
Hi MIKEVI, thanks again for the great inside. Apologies for not specific on my query.
My question was more on comparison to a startup low setting test compared to shut down and on the selection of speed. for example, bringing the setting down to a low value like 55% speed is a doable in a startup, but doing it in a shut down is not possible with normal logic forcing. And again it raise the question on why 55% and why not a lower value like 30% provided that trip is happening as a fired trip where trip circuits/valves are tested for proper operation. is there any thermal stabilisation sort of criteria is considered that you are aware of for selecting shut down and 55% speed?
although does this trip add 1 to the counter of fired trip? as per below graph from GEK3620 it does for startup
Thanks!
1619055752034.png
 
Yoganak, sorry for the delay, our outage is keeping me busy.
I can't directly answer your question as to why the 55% was chosen, it would be interesting if GE would provide details that answer these questions. I like you want to understand the logic, behind the logic. I want to ask the engineer "Why this way?"
Does the shutdown test configured for our site add to the trip counter-yes.
In my opinion if the owners insurance company would accept it I would write logic that allows for a selectable startup trip test. Selecting the startup trip test would automatically change the OS trip setpoint to 15% speed, or some speed that could be attained by the starting device of the unit if it were a CTG, and the unit would crank up and then trip at the low trip setpoint. Once the startup trip test is de-selected the OS trip setpoint would return to the normal value. For testing of an STG it would be again nice to write logic for a startup trip test that would execute and trip at turning gear speed if possible, or a slightly higher value that would be reachable with a small amount of steam. All of the above tests would not increment the trip counter.
As discussed earlier I have seen lots of different procedures in place at other sites, each procedure chosen for a specific site's needs. One thing I used to like about GE gas turbine logic when it was being written from the USA was it was very consistent from site to site. The core logic was nearly the same which made diagnosing and assisting with issues much simpler. Sadly it seems these days with logic being written from other areas of the world, logic is not so standard, or easy to uderstand.
 
MIKEVI,

You have touched on the primary reason for tripping gas turbines during shutdown: It fully tests the ability of the control system (and fuel delivery system) to shut off the flow of fuel when an overspeed condition is detected. Which is what the overspeed test is really designed to do--make sure no energy is delivered to the equipment when an overspeed condition is detected. Electronic testing of the overspeed protection doesn't do this--shut off the flow of fuel. It annunciates the alarm--and if the trip logic is written properly, it will trip the turbine by shutting off the fuel flow. But, those pesky insurance companies want more assurance.

Further, insurance companies have now traced some generator failures to forces experienced during (improper) overspeed testing at overspeed speeds--meaning, for example, 110% or greater while running (burning fuel). (A LOT of GE GT overspeed testing was manually initiated and speed was manually controlled. Instead of increasing speed quickly to just below the anticipated trip speed and then slowly increasing speed, most operators would take 10-20 minutes or longer at speeds above 100% to reach the overspeed trip setpoint. This apparently causes wedge problems and other issues in some generator rotor design/construction.) So, tripping at speeds less than 100% are even better for the generator.

So, tripping during shutdown does several things: Tests the ability of the control system (and fuel system) to trip the unit on overspeed (even if it's not the actual speed setting that would normally trip the unit). It does so when the unit is warm. It does so without inflicting unnecessary forces on the generator rotor. It even does so without the operator having to change speed manually. And, tripping a gas turbine during the shutdown sequence also reduces the thermal stresses on the hot gas path parts and exhaust. All good things.

Lastly, I believe the choice of the tripping speed is most probably the result of some calculation of critical speed operation for different units--hence why it might be different for different gas turbines, and even for different units in the same class which might have different rotors than other machines of the same class.

Hope this helps!
 
Thanks MIKEVI and CSA.
As an practical,doable and appropriate approach, we have selected 100% speed shutdown over speed test function (not yet implemented though). This is given the fact that no extra startup is required, turbine is fully heat soaked and trip can verify the function of all the systems that comes in to play during an actual overspeed event (sensors, input cards, hydraulic dump valves and SRV actuation, etc..). We believe Insurers would be satisfied anyway Since OEM is recommendation is to test at 55% speed only. Same strategy has been proposed for Steam Turbine.
 
Top