[Mark VI - Gas turbine] When the turbine is running with generator breaker opened, which control is responsible for integral action? I mean, if FSRN is proportional only control, how the governor is able to get zero speed error? (isochronous control is not enabled).

We had a situation here where the generator tripped due to 49 relay, but the turbine kept running. The machine was at base load when the generator circuit breaker opened.

As expected, the speed increased when the CB opened, and the controller was able to bring the turbine back to 100%. The TNR went from 105.3% (base load) to 100% automatically when the CB opened.

We noticed that the FSR in control was the FSRN all the time (FSR_CONTROL = 6 and FSR = FSRN). The question is, how the governor managed to control FSRN to control speed exactly in 100% if the FSRN is only proportional? Which part of the governor was responsible for the integral action?

This machine operates only in droop mode. We checked the FSRNI in this event to confirm, and it was 0% all the time, that is, isochronous integral control was not on. By looking at the FSRN block (FSRNV4, Mark VI) we cannot understand how it was able to reach zero speed error, because TNR was constant, and FSRN is only proportional control.

I could send you guys any data needed, just tell me to where I should send it and what variables do you need (we have all of them available in our historical server, with really good sampling).

What we can see is FSR (and FSRN) going to FSRMIN instantaneously when the gen breaker opens, and after that FSRN is controlled until TNH reaches 100% (like it had an integral action).

I'd like to post some pictures here to help, but I don't know how.

 

You don't need to post pictures (and we can't here at control.com); everything you describe is to be expected.

When the generator breaker is closed the actual speed is fixed (limited; controlled) by grid frequency. Droop speed control knows--and relies--on that fact. Increasing the speed reference increases the error between the reference and actual speeds and that's how fuel is increased.

When the generator breaker is open there is nothing limiting speed (except the load of the axial compressor). So, with nothing to limit speed there will be no error. Integral control is not required to make actual equal to reference when there is no error or nothing to introduce error.

 

CSA,

thank you for the quickly reply. However, I think I'm missing some point here. I mean, I agree that there is no grid reference now, but I do not understand how the control stabilizes exactly in FSRN needed to maintain FSNL, and this value was different from the FSKRN1 constant.

My FSNL configurable constant (FSKRN1) is 16.7% (we use standard droop control, FSRN = FSKRN2*(TNR-TNH)+FSKRN1), and the control took FSR to 19%. So that's why I'm confused. I'm my opinion, if it was only droop control in action, the FSRN would have to be 16.7% (and thus not necessarily 100% TNH, because FSKRN1 is not a exact perfect value for all turbine operational conditions).

The trend I collected shows the following:
- FSRN going to minimum (aprox. 11%) when the CB opens.
*and that's fine and totally expected;
- TNH reaches a peak of 102.57%, due to the load rejection.
*ok also;
- After some time, when TNH starts to reduce again due to the FSRN that went to minimum value, FSRN starts to increase again and control the speed. We can see an "overshoot" in FSRN and then it stabilizes slowly around 19%, and TNH around 100%.
*and that's what I do not understand. You saying that when we are not connected with the grid, we only need proportional control to stabilize in the TNR setpoint? It is like an "natural equilibrium point" for this condition? If whe change TNR for 101% or 102% for example, would the speed error be 0 also?

I posted two figures in google drive, if you want to check out, here are the links:
Figure 01 - http://goo.gl/2YxNc5
Figure 02 - http://goo.gl/ZYGMRQ

Again, thank you very much for the reply. If you want to, I can send you more data (trends, other variables, etc).

Best regards.

 

reboucasaraujo,

My best guess is that the gas control valve LVDT is not calibrated correctly, OR that the liquid fuel flow divider feedback scaling is not exactly what it should be or the liquid fuel supply pressure is not what it should be, OR that unit isn't using "straight" Droop Speed Control but is using "Constant Settable Droop Speed Control" (which is what most of the recently-built units are using, and which does have some integral action). OR, the most likely issue: The fuel has more heat content that was originally expected.

Actually, I think you are over-analyzing the "problem"--because you want the unit to go back to (approximately) 100% speed and stay there until you can decide to shut the unit down or re-synch the unit. That FSKRN1 being mismatched from actual FSNL FSR can really cause some problems when synchronizing because if there is a large mismatch between FSKRN1 and FSNL FSR the unit will NOT go to approximately 100% TNH. (And, actually, the unit should go to approximately 100.3% TNH, which is the usual, default value of off-line TNH in anticipation of synchronization which is usually performed when actual speed/frequency is a little higher bus frequency.)

How is it that you didn't notice this problem when synchronizing before? Because if it's a "problem" after a generator breaker opening event, it should be a "problem" during starting and synchronization.

<i>In theory,</i> you are correct (if the unit is using straight Droop Speed Control): the unit should not go to exactly 100.0% TNH when the breaker is open. But, again, this is proportional speed control--without any integral action (if it's straight droop speed control), so it could have just been a fluke that it did what it did and there actually <b>IS</b> some proportional action at work.

To answer the other question, yes. When you performing overspeed testing with the unit running, it's still in Droop speed control and as you click on RAISE SPEED/LOAD the speed does increase as TNR increases.

Again, I think the off-line speed setpoint (TNR) should be (typically is) 100.3%, and if there were ZERO error then the unit should have gone to 100.3% speed. (Unfortunately I can't recall the name of the off-line speed reference setpoint Control Constant.) So, there may, in fact, have been some error.

I really do think you're over-analyzing something that's not really a problem. Further, I'm interested to know why Base Load TNR is 105.3%.?.?.? Because, most GE-design heavy duty gas turbines have 4% droop, and that would mean that FSR would be (for a machine in new and clean condition and at ISO conditions) approximately 104.0%--and you listed 105.3%. Unless it was MUCH colder than rated (ISO condition) the TNR shouldn't have been that high--or the unit doesn't have 4% droop, or there's a lot we don't know about the machine and it's condition and the ambient conditions at the time of the event.

I can't really say much more than that. You would need to tell us about the fuel the unit was running on, whether or not it is as expected (as stated in Sect. 05.0n.nn, 'Expected Fuel Characteristics' for the fuel the unit was running on (gas or liquid)), if the LVDTs were calibrated correctly (if we're talking about gas fuel) by providing calibration data sheets that were done using a dial indicator or calipers as well as the section from the Control Specification which defines 100% stroke for the GCV, and the actual speed (TNH) and speed reference (TNR) to two decimal places (at least one) for the incident. Also, I don't have access to a Mark VI to look at FSRNV4 block, so you would need to post a screen shot of that block, and provide the value of the off-line speed setpoint Control Constant (I think it's something like FSKRN4, or FSKRNn). We'd need to know about how long since the last compressor wash, if the IGV LVDTs were calibrated correctly, and so much more.

There's just a lot we don't know--and again, the fact that the machine went to 100% and stayed there is not bad.

It's good.

 

CSA, thank you for the explanation.

Well, you are right. It is not a problem that the speed stabilizes around 100%, actually the turbine is doing exactly what we want to.

To clarify the purpose of my question, what I want is to understand better the control. In fact, we are simulating the turbine control, and the real reason I want to know if there is a "hidden" integral action somewhere is because I'm trying to match a full load rejection field data with my simulator data, and make sure that my model is representing well the actual turbine.

I already had a problem before with "hidden" things in Mark VI, for example, once we looked the FSRACC control and realized that the function block diagram shown in the Mark VI was not representing the actual equations/gains used. But there is subject for another post maybe, so let's go back.

So, I'll try to organize some points:

1) Please see FSRNV4 block in the link: http://goo.gl/1PVhSc

2) You are right again, TNR is 100.3% to help synchronization. It's the constant TNKR2 in Mark VI (Breaker open setpoint).

3) About the TNR 105.3% in base load, our unit does have a 5% droop curve, and not a 4%. In this event, the compressor was really clean (washed maybe two or three weeks before), and just went through a major inspection six months before. However, the external temperature here is much greater than project temperature (30oC). This turbine exhaust goes to a HRSG, so maybe that's why the droop is smoother (my guess). Do you have any guess about this?

I have to apologize about one thing. I kind of was surprised by the TNH being so close 100% like it was zero speed error, and because the recent "hidden" problem we had, then I just assumed that some "hidden" integral action was doing that. Actually, the TNR was 100.3% and TNH stabilized in 99.98% (there is no integral action, and we do have speed error). So, I think I can conclude that's only standard droop control.

I'll run the test again in the simulator, and make sure TNR goes to 100.3% when the breaker opens, and I'll check de results (luckily and if my model is correct, the FSRN will stabilize around 20% like the real event ).

Quick question: should I adjust the FSKRN1 constant to make it match the actual FSNL? How frequent we should do this (or maybe we should not)?

Well, thanks again for you help! I'm kind of a control enthusiastic, so I could talk about this for hours, and your posts here in the forum are always really interesting and didactic.

 

RRA,

> Quick question: should I adjust the FSKRN1 constant to make it match the actual FSNL?
> How frequent we should do this (or maybe we should not)?

Actually, unless you are having trouble synchronizing or you know the gas fuel make-up is significantly different from expected I don't normally recommend changing FSKRN1.

My posts have been called a lot of things over the years, but never didactic. Thanks! (I think...)