Steam turbine trip signal when GCB opens

[npower0073]

Usually steam turbines control system implements trip signal from several reasons like HP steam temperature and steam, electrical protection of generator, generator slot temperatures, low condenser vacuum or level, lube oil pressure, control fluid temperature etc. The trip scenario is that when the breaker opens then then control flkuid )oil used for the control of steam supply valves) looses its presure and the steam valves closes. How can i implement a trip signal that will take in to accoubt the opening of the turbine generator circuti breaker in order to trip the ST?

 

[byrdj]

Interesting....

the steam turbine governor should be able to "control" a GCB opening ie prevent overspeed from exceeding emergancy governor setting and then let settling speed keep the STG spinning near synch speed so it can maintain house power.

I recall for the new units I help place into service, there was a boiler cross trip that would trip the turbine, and the boiler would trip on a GCB opening.

thus to "test" governor, the boiler cross trip had to be lifted

If you donot want to control overspeed on GCB, how about lowering the emergancy governor setting so it will trip, say at 102%. for most STGs, that will be about .2 seconds later

 

[npower0073]

The turbine finally trips from cross trip from DCS BUT this comes 2-3 sec after the breaker open, so gine with the breaker open and 3 sec to still feeed steam to the tubiwill finally overspeed. I was thinking maybe on control scheme to calculate accelearation by simulating a load loss and hence trip the turbine pwoer oil and thus close the steam supply valves, ESV and CV.

 

[byrdj]

If the CVs do not control, the overspeed trip should occur less than 1 second after breaker open.

Comparing real power (a stage pressure) to electrical power (current) could be used to calculate an expected acceleration if there is a mismatch (due to current going to zero). This has been used to provide CV closure prior to normal CV closure by pure regulation

why do you want to trip the turbine if it is still in control?

 

[npower0073]

Becasue when i have a load rejection with the turbine loaded to 110 MW until 140MW(140 MW is the base load) i get an overspeed. The control scheme and the vlaves (steam supply and bypass) cannot tkae the load rejection and also the turbine is uncapable of staying to FSNL after the syhcnronization. The breaker opens only form revese power and then the turbine trips.

 

[byrdj]

You said you "get an overspeed". Did you mean the speed goes to overspeed trip and you get that trip.

Because the speed will go faster than synch (and in normal control jargon, this is still called an overspeed) but this is expected. for the units I performed full load rejection testing on, the peak overspeed was near 109% with the overspeed trip at 110% and this was concidered excellent control response (supercritical with very low inertia rotors).

So if you want to trip on breaker open. just lower the overspeed trip speed. This would apply ONLY to an electrical OST AND you need to very the reseting dead band will allow on line testing if so equiped.

If the reason the turbine goes to overspeed trip is because of delays in the control computations (I am not familar with computer based controls, and still wonder how they can ever be fast enough to perform governor function) Then providing a circuit like the comparison between real power and electrical power still might not be fast enough. But if there is the capibility to add, it might be worth a try.

and last, why not put a trip contact (logic input) directly from the breaker status. this would be limited protection though since the load rejection might occur down stream of that breaker. thus the circuit that compares real to electrical is the most realiable

Now the correct thing to due would be make the turbine governor limit peak overspeed less than overspeed trip. the units I tested were able to do so was with the circuit that compars real/electrical power and additional hydrualic components to reduce the CV closure time to less than 0.1 second (and we still went to 109%)