We have simple cycle frame 7EA's with MarkV and DLN1. We are considering eliminating the coast down from purge speed (~22% TNH) to firing speed (9.5%<x<10%).
If the unit is allowed to fire at 21%<x<21.5%, it would shorten the start time by ~4 minutes.
Has anyone done this? Were there issues? Lessons learned?
Please share your thoughts/experiences/insights.
>If the unit is allowed to fire at 21%<x<21.5%, it would
>shorten the start time by ~4 minutes.
Do you mind also shortening the hot gas path parts life probably by as much as 25%?
Sure, these unit can be fired at purge speed--but it will take a lot more fuel to establish flame, AND it will cause a large spike in combustor temperature and exhaust temperature.
What have you done to reduce starting time? There are other things which can be done to reduce starting times, but eliminating the coast down to firing speed is not one that is recommended.
Per NFPA there are ways to get "credit" for coast down after a shutdown or trip, as well as to get "credit" for time on cooldown (especially if the unit(s) has(have) slowroll instead of ratchet; some older 7EAs did, usually those with hydrogen-cooled generators, while newer 7EAs with air-cooled generators required a ratchet because of the weight of the air-cooled generator rotor). There are sites which have installed oxygen sensors in the exhaust to check for the presence of combustible gases/vapors and if below a certain level the purge time can be shortened. (Personally, I'm not a fan of this one because there can be so much stratification in flows in the exhaust duct transition area (which is where the O2 sensors are usually mounted), and if they fail to detect combustible gases (or a lack of oxygen), that could be not good....)
And, I believe if the unit experiences a failure to start for any reason, or a trip during starting/acceleration, while attempting a reduced-time start-up then it's back to the normal start-up purge and acceleration. And, since ambient temperature can vary greatly over a year--and this can affect the amount of fuel required to establish and maintain flame during starting--this can be a great risk if a site is getting paid for availability and misses a start time. And, there's no really good way to determine exactly how much fuel it would take to reliably establish flame and maintain flame during warm-up--it would be a trial-and-error process, with lots of starts required to try to reliably establish and maintain flame while at the same time trying to minimize thermal stresses caused at the higher firing speed.
But there are things which can be done which are much less harmful to the unit, though GER-3620 still lists fast starts (or reduced-time starts--by whatever method or for whatever reason) as having a large influence on Factored Fired Hours and parts life. (GER-3620 can be a very difficult read; it's got so much information for so many Frame sizes, but mostly F-class units.)
Myself, I wouldn't do this particular method of reducing start time without researching it very carefully with the OEM. Again, it's possible to fire at higher speeds--but it requires more fuel and results in higher thermal stresses to the hot gas path parts and the exhaust diffuser. And DLN combustors and hardware are not very forgiving of thermal stresses....
If the unit(s) is(are) operated in combined cycle mode and exhaust into an HRSG it might also cause unknown knock-on effects in the HRSG.
Hopefully one of our responders, mhwest, will chime in with his experiences with 7E/EA units and fast starts. He may have a different perspective or experience.
Hope this helps!
Here's a thought.... As a test, set up Trend Recorder to monitor TNH, L2TV, CSGV, TNHA, L14HM, L83GVMAX, and L20TV1X, and then start it monitor a normal START and stop the Trend when the unit gets back to 10% speed after coasting down to firing speed.
Then as a test when the unit is at rest and at the typical temperature for a normal START select CRANK Mode and initiate a START. As soon as the purge time is complete and the torque converter is de-pressurized force L83GVMAX to a logic "1" while watching TNH. The increased air flow will cause the speed to decrease at a faster rate. I would recommend unforcing L83GVMAX when TNH gets down to say 15% or 14% or so. And pay very close attention to when L14HM drops out (goes to a logic "0") and how long it takes for TNH to get back to 10% and L14HM to go to a logic "1". At this point (when L14HM goes back to logic "1") you can stop the Trend and initiate a STOP.
Look at the two Trend events and compare the times it took to go from when L2TV picked up (went to a logic "1") and when L14HM went back to a logic "1", AND the times L14HM dropped out until it went back to logic "1" for the two events in the Trend. Also monitor the changes in TNH and TNHA when L14HM dropped out until it went back to logic "1".
If opening the IGVs during the coast down improved the time it to to get to firing speed AND the time it took for the starting motor to get the speed back to firing speed after L14HM dropped out until it picked up again didn't increase drastically then this would be something to think about for decreasing the time to coast down to firing speed. If the time required for the unit to recover from L14HM dropout to L14HM pickup was "excessive" you could try unforcing L14HM a little sooner, say at 16% or 17% so that the rate of deceleration might be a little slower when L20TV1X picks up to start trying to accelerate the unit back to L14HM pickup.
If this is an acceptable decrease in the time to coast down to firing speed it would be a simple application code change to make and it even temporarily increases purge air flow and volume during a a START with NO thermal stresses on the unit. A win-win for everything!
I don't know how much this will shorten the time to coast down to firing speed; it may be negligible. But by tinkering with the point at which the IGVs are closed in order not to make the re-acceleration back to firing speed not take too long this could be a simple way to improve overall start time.
Please write back to let us know how this works for you! (Before a year passes, hopefully! ;-) )
Couldn't you open the IGV's at startup to increase airflow and reduce the purge speed? This would reduce ramp up and ramp down time, and maybe even the purge time itself. Someone would have to do the calculation to see how low you could go and still move the same amount of air.
When purging at ~20-22% TNH the electric cranking motor on a GE-design Frame 7EA heavy duty gas turbine is (should be) drawing about 150%-160% of rated current. If the IGVs are open at starting and during purging the starting motor will not be able to spin the unit up to normal purging speed (though it will be drawing the same amount of current--until someone figures out they can increase the purge speed at which point the starting means will be drawing even MORE current!).
Yes, someone would have to "run the numbers" to see how much time it would take to achieve the same purging action. (Which means someone needs to know how much air is flowing at different speeds with the IGVs at different angles, and how much volume the exhaust (and HRSG and stack) have. The air flows for various IGV angles at reduced speeds would probably be the hardest to obtain from the OEM.... They don't normally give out that information very readily.
Just opening the IGVs during coastdown involves the least amount of "work" and calculations. I don't even know if it would reduce the total starting time by very much.
A LOT of Frame 7EAs (and even 6Bs) have trouble with off-line water washing and rinsing when the IGVs are at maximum angle and the starting means is at minimum torque. Many of them drop below 14HM drop-out and that trips the off-line water wash sequence.... It's even worse when water/detergent is flowing into the compressor with the IGVs wide open; the extra torque required to move the air and water/detergent with the IGVs wide open at reduced torque just can't be supplied with the normal torque converter settings (on a 7EA). Makes for some very frustrated operators and plant managers who simply cut the rinse (and/or wash) short and don't get the full effect of a proper wash and rinse.
If one had access to all the data your idea might possibly be a solution--but I think the OEM has considered your proposal and has not considered it a good option (for whatever reason). The original poster asked about eliminating the time for coasting down to firing speed, and this was the best option I could think of without causing serious damage to combustion hardware (especially DLN combustion hardware). I'm hoping the original poster will write back with some times and speeds (acceleration (or deceleration, rather) rates).
>I'm hoping the original poster will
>write back with some times and speeds (acceleration (or
>deceleration, rather) rates).
PI data shows that our unit is at 22% TNH when the purge timer ends. L14HM is a "1" the whole time.
It takes 171 seconds to coast down to 9.5%, at which time L14HM goes to "0" and the torque convertor re-engages.
It takes 26 seconds to accelerate to 10%, whereupon L14HM goes to "1", the ignition exciter send sparks, and fuel is admitted.
It takes 10 seconds for L28FD (flame detected.)
It then takes 63 seconds to re-accelerate back to the 22% where the "coast down to fire" event began.
I like your idea of increased "braking" during coast down with L83GVMAX. The 171 seconds from 22% to 9.5% should shorten. I will try to convince my Manager to allow this experiment.
My thoughts are that GE is being conservative and applying purge time (atmosphere exchanges) that are required by NFPA for boilers and HRSG's to all units.
Turns out that since 2011, NFPA 85 allows for "Purge Credit". The purge credit, simply defined, is when, during a normal CT shutdown, there is sufficient CT post-purge airflow to remove any residual combustible materials from the CT and HRSG system. If gases cannot seep into the ductwork, then the post-purge meets the NFPA requirements and a redundant start purge is not required.
JUST get to 10% and fire!!!
They have 3 allowable gas supply piping configurations for natural gas turbines.
System 1 requires 3 valves; There must be vent valve and pressure transmitter between each pair of valves. All of which must be monitored by your control system. This allows start with no purge for an 8 day time period, which you can re-start by performing an unfired purge.
Systems 2 & 3 provide for an unlimited time interval for zero startup purge.
System 2 uses the same valves but does a constant inter-stage purge with compressed air or inert gas.
System 3 has the same valves but after gas is vented, they close and the inter-stages areas are pressurized to >3 psi with compressed air or inert gas.
Since we do not have compressed air available to our units, I find system 1 to be the most appropriate for our units. The units have the 3 valves but only the SRV to GCV inter-stage area is vented and pressure monitored. I would need to add another solenoid vent valve and pressure transmitter, not to mention cut into piping for manual vent so the solenoid vent will go to the proper elevation discharge; conduit & wire runs to MarkV; signal names, inputs, and logic for automating HMI monitoring, triggering 8 day timer, a select block to swap purge time required, and some status graphics.
This may be ~$10K per unit, as long as we do all work except the vent piping mod ourselves. While I want to do, I'm not sure if my owners will want to spend the money.
The experiment can be tried an hour or two before the unit is to be put on line, and it doesn't involve firing the unit. Just put it CRANK mode, and perform the test(s). You may want to play around a little with the speed at which you close the IGVs. If the unit is slowing down too fast when it goes through L14HM dropout (9.5% speed) it may take longer to "catch it" and accelerate back through L4HM pick-up to fire (of course in CRANK mode it won't actually fire).
And, before you can just go to "10% and FIRE!" you must have the double-block-and-bleed arrangement, or the inert gas injection. As you rightly noted, there has to be NO chance for fuel or combustible gases to make their way into the unit to 'legally' just go to "10% and FIRE!"
Just going to 10% and firing isn't going to be possible on every start, and that is going to take some more sequencing mod's. If you want to make it automatic (meaning no operator intervention to make it fire at 10% without any purge) you will need more application code (logic). And, if you make it require an operator to initiate the "10% and FIRE!" logic then you will need to make buttons on the HMI and have more logic--and INVARIABLY an operator will forget to initiate the logic, and then the unit will miss its window and it will be the damn Mark*'s fault! (It's ALWAYS the damn Mark*'s fault anyway...!) And, if you have Mark VIs, you will have a lot of puts and gets for the GTs and the exciters and the HMIs--not fun stuff in my personal opinion (too much "opportunity" to make a mistake or three).
Please write back if you can convince the ownership to perform the test; yes, it will consume MWHs for the cranking motor, but it could result in increased revenues in the longer run. And, after all, it's all about the beans!!! (At least that's what the bean counters and the owners believe.)
I finally got opportunity to experiment the effect of opening IGVs during coast down from purge to firing speed.
I forced L83GVMAX to "1" as soon as the torque convertor disengaged. I unforced the logic at 9.6% TNH.
The time to coast down was reduced from 171 seconds to 160 seconds.
Future start time reduction plans will be focused on fuel valve/inter-stage pressure monitoring to legally claim purge credit as allowed by NFPA 85.