We have a Mark V on our turbine generator controls. It appears, however I am not sure, that the reference speed for the system is established when the sync switch is activated. Even though I do not normally work on the Mark V, I was tasked with finding out if there is a work around to our problem because we are not going to take the unit offline.

By the information given to me, the turbine's throttle valve was short stroked in error and after reading other discussions on Control.com, this is probably the cause - now we need to somehow get around it.

Here is the information given to me, and I will be investigating and reading more today (105 megs is max load and we can only get to 93):

In the past, during some turbine start ups, we have selected the synchronize switch to "On" before reaching 3600 rpm. This some how caused a different relationship than normal between the LOAD SETPOINT (The value the operator types in for valve position) and the V1 Position (Control valve position.) Normally at 105 megs, LOAD SETPOINT is about 86% and V1 Position is about 75%. One start up years ago, after changing some control constants to allow full load, LOAD SETPOINT had to be 120% to get a V1 Position of 75% for 105 megs.

In this case we have changed the control constants to the max allowed by the Mark Five, and still can not reach full load of 105 megs. The LOAD LIMIT REFERENCE is set at 128%, the LOAD SETPOINT is at 128%, the V1 Position is about 86%, and the max load we can reach is about 94 megs.

In the past we have changed the following control constants to allow reaching full load:

KLDR_H (Load Setpoint High Limit) from 110% to 125% and now at 128%
KLLR_H (Load Limit Setpoint High Limit) from 110% to 125% and now 128%

We have tried to change KV1POSR_H (V1 Position Setpoint High Limit) which is at 128% to a higher value, but the Mark Five will not allow any value greater than 128%."

Thank you for any assistance.

 

I'm not really clear here. You haven't told us how physically open the V1 rack is when the Mark V is at a reference of 128% and the LVDT indication is 86% (I'm presuming that the 86% you described is from the LVDT feedback). Does the valve rack have a position indicator on it? Is the rack at full open or only at 65% or 78% or 84% or ? Or is that the 86% that you listed?

I don't think reference speed has much to do with load on a steam turbine, like it does on a GE-design heavy duty gas turbine. I think the droop control on a steam turbine control is more of a load-based signal than a speed based signal, or some kind of load-biased speed reference.

It's been my experience that steam turbines have several modes of operation. But, when on "part load control" (i.e., not on Inlet Pressure Control or Extraction Pressure Control) that there is a relationship that has been established between valve position and load. So, when an operator says go to 100 MW there is a corresponding valve position that the Mark V positions the control valve to. I don't know if there's any feedback to the loop from a load (MW) transducer but I would suspect there is, but that's just a SWAG.

Either the valve rack is fully open and there's not enough steam going into the turbine, or the calibration of the LVDT feedback has been so mis-adjusted that the Mark V thinks the valve rack is more open than it really is and it doesn't think it can open it any more.

I have seen actuators put back together improperly, resulting in the actuator hitting a full open limit well before the valve is actually open, which would prevent getting enough steam into the unit.

One of the "limits" you seem to be hitting is that the maximum scale value for PCT is 128.00. You can't put in a value of 150% or 200%, because the way the Mark V is programmed the maximum value for PCT is 128.00.

To change LVDT calibration values, it would be necessary to download a new I/O Configuration to the control processors and then re-boot them (I'm presuming this is a TMR Mark V). If you're not willing to take the unit off-line to do the re-boot, that can be very risky. If you had some I/O Config printouts from when the LVDTs were properly adjusted, then you might be able to develop some "curves" but re-booting the processors would have to be done one at a time (which they always should be anyway) and that just raises the risk for problems. On the first re-boot, two of the three processors will be using the older values while one will have the newer values. On the second reboot, two of the processors will have the new values and one will have the old values. It's just pretty risky, and would require a lot of ... hair.

And I would say it would have to be done in steps, meaning several downloads. And each download would require three re-boots! Mucho hair required!

I believe there is an analog linear interpolation table in the CSP that sets up the relationship between load and valve position, because load is usually not directly proportional to stroke. For example, the V1 rack might be at 34% stroke for 50% load, and at 73% stroke for 75% load, and at 87% stroke for 95% load.

If you were really confident and knew exactly what you were doing, it might be possible to tweak some of the table values to try to fool the Mark V while the unit was running, but that would also be very risky. In my opinion, riskier than downloading new I/O Cfg. values for the LVDTs. And then you're going to have to decide what to do when you shut the unit down (if it doesn't trip), because there might be some spikes or dips as load is lowered that could also cause upsets or trips if you only change the mid- or upper range table values.

I just don't see how you're going to avoid shutting the unit down and performing a proper LVDT calibration. If that's the problem. I can't decide from your description if the LVDT calibration is bad, or if it's just not possible to get enough steam at the proper temperature and pressure into the unit?

We also don't know if there are any extractions on this unit and if they are working properly.

What kind of turbine is this? A utility turbine being driven by a fossil fuel-fired boiler? A steam turbine at a combined cycle power plant? An automatic extraction steam turbine? Single- or dual extractions? (Not that this makes a lot of difference; changing LVDT calibrations requires downloads and re-boots. Changing Control Constants on the fly is also risky. The Mark V isn't very easily "manipulated", not safely anyway!)

 

I sincerely thank you for your response. We have, within the last hour or two, finally gotten in touch with our turbine engineer contact who pretty much said exactly what you did - which is VERY good to know.

The avenue we are discussing is the LVDT curves with old data which we have.

There is a miscal in the LVDT which occurred during the outage and the valve is more closed than the indicating feedback. It developed due to a misunderstanding of the Mark V and maintenance was wanting to limit the last poppet from opening due to leaking hydraulic fluid - at least that is my understanding. The miscal came to light well after tie-in.

The only thing he said than you seemed to indicate we should avoid is that you suggested doing the change in steps and he said all at once and let the system slowly ramp to the new position.

Common sense makes me want to follow your suggestion, I will present it to the people involved.

Again, thank you for sharing your knowledge.