I would like to know what is the expected speed follow up a trip in a SC5 steam turbine, where can i find out that information is there some GEK that talks about it? my unit reach 3717 RPM after a full load (125MW) trip is this OK? i am checking the logic in the MKVIe but i am stuck in the speed error block (i think this one can tell me something about) any comment is appreciated.

 

Are you suggesting that there should be a particular speed that is achieved after a load rejection?

I've never heard of such a concept, but that' doesn't mean it doesn't exist. But, reasoning through the situation, removing the load suddenly from a steam turbine necessitates closing the stop valve(s) very quickly, and opening any "intercept valve(s)" very quickly. A steam turbine without a load is like a pinwheel in a cyclone--there is nothing to slow it down and eventually it will just disintegrate.

Will there be some overspeed when the turbine control system senses a load rejection? Yes, indeed. But the idea is to limit the overspeed as much as possible. That's why steam turbine stop valves are usually placed very close to the steam turbine--to reduce the volume between the stop valve and the steam chest (where the steam is admitted to the turbine) and keep that volume to a minimum. Because, when the stop valves do slam shut there will be steam--under pressure--trapped between the stop valve and the steam chest and this steam will still flow into the steam chest/turbine. Without a load this will cause the steam turbine speed to increase; but by limiting the volume of piping between the stop valve(s) and the steam chest/turbine the amount of overspeed will be limited.

It's hard to understand why it would be expected that the turbine would reach a particular speed during a load rejection. Again, there will be a speed increase--simply due to the time between when the loss of load is sensed and the time required to close the steam turbine stop valve(s) and the volume of steam trapped between the stop valve(s) and the steam chest/turbine. But, the designers of the turbine and piping have (or should have) calculated that under a normal scenario the speed would not approach the normal turbine overspeed setpoint.

Every turbine basically has a maximum allowable speed--that is above the overspeed setpoint(s). And, again, steam turbine designers are cognizant of this and design systems and components to ensure that even if the turbine did reach the overspeed trip setpoint before the stop valves closed that when the stop valves did close on overspeed the turbine would (in theory) not reach the maximum allowable speed (above which centrifugal forces would begin to cause damage to the turbine).

Steam turbines, because they can overspeed SO quickly (they don't have a big "brake" like gas turbines do--the "brake" being the axial compressor which is consuming as much as two of every three horsepower produced by the turbine section) are provided with control systems which can have very sophisticated "trip anticipators," sometimes called Power-Load Unbalance modules or systems. These modules are looking at steam flow versus power output (watts) and when the two are beginning to get out of balance with each other (more steam is flowing than power is being produced) then the trip circuits are activated. These systems don't wait for reverse power--that would be too late for some larger steam turbines; they are looking for an early indicator of low generator power output versus steam flow input (which will occur during a load rejection long (milliseconds) before reverse power) to actuate the stop valve(s) to shut off the flow of steam to the turbine to limit the amount of the overspeed, and to prevent reaching the overspeed setpoint.

But, again--I've never heard of a "search" speed or something like that which a steam turbine will automatically reach during a load rejection. The ideal is to have as little overspeed as possible, but some overspeed is unavoidable just because of the piping, and control system and valve "lags". Again, all of this is supposed to be considered by the steam turbine designers in order not to reach the overspeed setpoint, and to never reach the maximum allowable speed on overspeed.

If you have a reference to some document that alludes to this "expected" speed after a load rejection, could you please post it here? (You'll have to copy it and paste the text; files can't be posted to control.com.) Again, the idea is to limit--as much as possible--the overspeed which will occur after a load rejection (or turbine trip), hopefully to keep it below the overspeed trip setpoint, but certainly below the turbine's maximum allowable speed (which may or may not be a published/known quantity).

So, I don't believe you're going to find this expected speed setpoint in the application code in the Mark VIe. If you're looking for the actual speed that was reached, I think that the <P> (PPRO IOPACKs) store that speed on a turbine trip, but once it's been re-set/re-started the value is lost). I believe it's also stored in the Trip History buffers, which should have been uploaded to the HMI.

 

CSA,

Can you expand on the trip anticipator logic?

My understanding... an over-speed can be detected in less than one complete revolution.

If so.. wouldn't there be a maximum attainable speed after the stop valves are slam shut based on design and piping.

 

CSA Thank you very much for the time and effort of your answer the explanation is more than clear.

Thanks a lot

 

MC,

'Trip anticipators' as they are sometimes called are done in different ways by different control system manufacturers. The one you are describing sounds like it may be done using a high-speed acceleration detector.

I would imagine that using enough maths and formulae that just about any speed could be calculated based on acceleration and rotor inertia and windage losses and on and on and on. How accurate those calculations might be is probably pretty questionable given such things as how the steam turbine stop valves can be "contaminated" with deposits which affect closing rates, and servo-valves can be contaminated with dirty oil, and on and on and on.

I imagine there are some control system manufacturers who could claim they could calculate the highest speed attained during a load rejection but I would presume there would be restrictions (new and clean condition of valves and servos and oil, for example) for such a calculation.

The majority of my personal turbine experience has been with heavy duty gas turbines; most of my experience with steam turbines, particularly large steam turbines, has been anecdotal and theoretical.

It's difficult for me to imagine how this number (maximum speed attained during a load rejection) could be used; would it be for historical data (trending maximum speeds over some period)? If so, that information would likely be most useful if every trip occurred at the same load when the turbine was at the same temperature and the vacuum was at the same level--which happens how often? Would it be used to predict some kind of refurbishment or maintenance on steam turbine stop valves? Or packing? Or blade tips? Would it be useful for understanding blade conditions? Would it be useful for determining if steam turbine stop valve closing rates/times are per specification?

I'm having a difficult time trying to understand how achieving a particular speed during a load rejection event would be considered "good" and desirable. Or, even how it could be useful for monitoring turbine health and condition. I guess if the turbine exceeded a particular speed enough times in a certain period that the condition of the turbine or blade tips or packing or valves could be questionable, but, that is, again, using more information than just the value of the speed attained--it's the number of times AND the speed attained over a period of time.

So, would it be possible to calculate a particular speed which might be attained during a load rejection? I'm sure, with enough information--and the manufacturer certainly has the theoretical values for such a calculation--that such speeds ARE calculated, again, in helping to determine piping arrangements and valve actuators. But, is that "value" coded into the control system for all conditions? And, what value does the actual attained speed provide?

What would the tolerance of that calculation be: plus-or-minus 10 RPM, 20 RPM, 50 RPM, 100 RPM? How accurate would the calculation be? What if the speed were much higher or much lower than calculated? What does that indicate?

Seemingly the calculation and any real-world value would lead to more questions and the need for more data and analysis. But, as I read the original poster's question it was asking if there was a particular speed which <i>should be</i> attained during a load rejection. Maybe I misunderstood the question, but in thinking through why such a value might be useful I'm having a very difficult time finding one or two compelling reasons why such a specific value would be useful or meaningful.

 

first
CSA explanation is extremely detailed and I will not be nearly as eloquent in my posting

I do not have knowledge of GE digital steam turbine controls (left field after MarkIIa)

I can not find the configuration of a SC5. I did find a C5 is a non reheat double flow LP in your MW range.

My comments and thoughts

on a "turbine trip" i would not expect any rise above rated. A turbine trip implies the SVs (and CVs) closed prior to the generator breaker opened. for a trip, the generator breaker should open once the logic senses valves closed to isolate steam to the turbine and the reverse power relay has been ergived for a few seconds.

as CSA mentioned, steam turbines have a very great acceleration rate for a load rejection, a normal expectations is 10% speed per SECOND. thus the logic of only opening the breaker once turbine power is less than 0 MWs.

A "load rejection" is completely different than a "turbine trip". the evolution of turbine controls I worked were designed not to trip on a load rejection, however most plants were designed so they would not survive a load rejection and some other plant condition initiate a cross trip of the turbine.

the simple axioms of control preached to us back when were
1) the speed governor will be fast enough to control peak speed below over seed trip on a load rejection.
2) if there is a complete failure of the governor, the overspeed trip will activate soon enough to prevent peak speed from exceed destructive limit

Now for a load rejection, the peak speed the unit would accelerate to and then return to settled speed near rated will be proportional to the % of rated MWs prior AND the level of governor protection available. given the nominal over speed trip of a steam turbine is 110% of rated, with a simple governor response, the peak speed following a load rejection will be proportionally from 100% speed at 0% load to 109% speed at 100% load.

that single line expectation would only be applicable for the simplest of governors and an expected acceleration rate on load reject much less than 10% second. For example most units with reheat would accelerate past overspeed trip with a load rejection of only 20% rated load, so additional governor protection of IV fast closure.

so the expected peak speed would now be a saw tooth line with 100% speed at 0% load, 109% speed at 20%, but step down to 102% at 21%.

even with IVFC, the utility units I worked would exceed overspeed trip at 40% load so an addition circuit call Power Load Unbalance was added, thus putting a second step in the expected peak speed.

Now you stated you went to 3717. If the rated speed was 3600, then that was only 103% peak, BUT if the rated was 3000, then that is 123% over speed which very near destructive and indicates a load rejection with 100% governor failure.

here is a write up for a customer showing the expected peak speeds following load rejection (and discussion relate to plant changes)
http://www.slideshare.net/JosephFByrdJr/first-line-of-over-speed-protection1

I would be interested to learn more of the facts of your event once you apply the information received.

 

I'm not familiar with your turbines, but we have large GE steam turbines.

For a non-electrical fault type trip - when you can afford to wait a couple seconds before de-energizing the generator, a reverse power relay is normally used to ensure that the rotor will immediately slow down once the breaker is opened. So the logic is normally to verify that the turbine valves are closed, that the turbine is tripped, and that you have reverse power - before you separate from the grid. It normally takes a half second or so for all the stored energy contained in the extraction piping, turbine casings, etc. to dissipate via blown down to the condenser.

If you do not have reverse power relays, you will have some overspeed following each unit trip, due to the stored energy in the shells and piping. If you have reverse power relays, this will only happen during electrical fault situations.

During a true load rejection, where there is no turbine trip along with the loss of load, the turbine will accelerate before a trip occurs. GE has tried to design the control system to theoretically survive such an event without tripping - and theoretically be capable of supplying its own aux load with no offsite power, and to be capable of resynchronizing to a dead grid minutes later. Easier said than done.

3717 RPM following a full load trip doesn't sound high to me if it is a 3600 rpm unit, - if reverse power isn't used as a breaker opening permissive, there is extraction piping, etc. But it might be high for your unit.

There could be a problem with leaky feedwater heater extraction check valves, turbine valves closed limit switches, valve closing times, etc. You need to investigate.

For a steam turbine, there is no better friend than a redundant reverse power detection scheme that must be satisfied (for non electrical fault type trips) prior to opening the breaker. If the turbine is sucking power in from the grid prior to opening the breaker, there will be no overspeed.