I hope to learn more on the temperature control reference BBL here. My machines are of GEfr9E dual fuel running on MarkV turbine control.
How is it possible to determine the temp curve for each machine?
Some details on my machines when both on base load are as follows:

(i) GT1
TTRX:568 degC
TTRXP:571 degC
TTRXS:590 degC
TTRXB:568 degC
TTKRXBP: 13.79 bar
FSR:62.81%
CPD:10.31 bar
Loading:106MW 4MVAr
Inlet temperature (CTIM):30 degC
Ambient temp bias minimum (TTKTCIMN): -29 degC
Temp bias offset constant (TTKTCIO): 1120 degC

(ii) GT2
TTRX:566 degC
TTRXP:566 degC
TTRXS:576 degC
TTRXB:566 degC
TTKRXBP: 13.79 bar
FSR:69.36%
CPD:10.7bar
MW:107MW 4MVAr
Inlet temperature (CTIM):31 degC
Ambient temp bias minimum (TTKTCIMN): -29 degC
Temp bias offset constant (TTKTCIO): 1120 degC

Lastly what does the constants TTKn_C, TTKn_S, TTKn_K, TTKn_M and TTKn_I represent (where n= 0,1,...7)?

 

This is a really big question, with a not-so-easy answer. The Control Specification, Section 3, Temperature Control, has most of the answers you are looking for. The descriptions of the Control Constants are mostly found in LONGNAME.DAT, and also in the Control Specification.

TTKn_x is an eight-"element" array, where 'n' is 0 through 7, and 'x' is I or C or K or S or M, etc.

The exhaust temperature control algorithm calculates a CPD-biased (or CPR, a fancy CPD calculation) exhaust temperature control reference using the Control Constants TTKn_x, where the value of 'n' is determined by which of L83JDn is a logic "1" (where, again, 'n' is 0 through 7). So, if L83JD1 is a logic "1", then the array TTK1_x is being used to calculate the exhaust temperature control reference. If L83JD3 is a logic "1", then the array TTK3_x is being used to calculate the exhaust temperature control reference.

Another factor which determines which array set is used is which fuel the unit is operating on, so for multiple fuel machines the fuel will also determine which L83JDn logic is selected.

The CPD-biased exhaust temperature control reference represents *constant firing temperature*. Since firing temperature is not measured, it is calculated using an empirically determined relationship between compressor discharge pressure (CPD) and exhaust temperature. It's known that for a given CPD and exhaust temperature the firing temperature is [some value]. The relationship for CPD and exhaust temperature can be determined for a *constant firing temperature.*

The exhaust temperature control reference has a negative slope, meaning that as CPD increases the exhaust temperature decreases *for a constant firing temperature.* That seems opposite of what would be expected, because as the unit is loaded and CPD increases, exhaust temperature increases, *but so is firing temperature.* So, the exhaust temperature control reference line (which represents *constant firing temperature*) is drawn on a graph where CPD is on the X-axis and exhaust temperature is on the Y-axis it slopes down and to the right (a negative slope).

Further, for many machines, the exhaust temperature control reference (*the constant firing temperature*) is not a straight line. It can "bend" at some point(s) looking more like a curve than a straight line, increasing power output or decreasing power output *while firing temperature remains constant.*

To approximate a "curve", the Speedtronic uses a series of intersecting straight lines, each one is represented by an element of the TTKn_x array.

Lastly, there is a "primary" exhaust temperature control reference (usually CPD-biased), TTRXP, and a "back-up" or secondary exhaust temperature control reference, TTRXS, which may be load-biased or FSR-biased depending on the age of the control system (newer units usually use load bias for back-up exhaust temperature control). Both references are calculated simultaneously, and the minimum value of the two is used as *the* exhaust temperature control reference, TTRX.

I believe that TTKn_I, TTKn_C, and TTKn_S are used for the primary, CPD-biased exhaust temperature control reference, and TTKn_I, TTKn_K, and TTKn_M are used for the back-up exhaust temperature control reference.

The formulae are:

TTRXP = TTKn_I - ((CPD - TTKn_C) * TTKn_S)

TTRXS = TTKn_I - (([FSR or DWATT] - TTKn_K) * TTKn_M) (since the back-up is either FSR- or load-, DWATT, biased)

That's the very *basic* description of exhaust temperature control for a Speedtronic turbine control system used on a GE-design heavy duty gas turbine. Note that the actual firing temperature is never measured or expressed; it's represented by the exhaust temperature control "curve" and expressed as the relationship between CPD and exhaust temperature.

The back-up exhaust temperature control "curve" is supposed to be calculated so that it mirrors the primary exhaust temperature control curve, but is just slightly above the primary exhaust temperature control curve. Units are *NOT* designed to be operated for extended periods of time at the back-up exhaust temperature control curve; it's just there for "emergencies" when the CPD signal(s) are not available for a bias. If CPD signal(s) are lost, the unit should be shut down and the cause found and resolved so that when the unit is re-started it will be operating on the CPD-biased exhaust temperature control reference.

There are some other biases that are used on some machines; most do not use any kind of other bias. Newer machines use CPR bias, which is a kind of fancy CPD which is calculated using barometric pressure transducers (which must be very accurately calibrated for the optimum performance of the unit). Some also look at exhaust duct back-pressure.

Now, you're going to ask, "Why do I have two identical machines, built, installed, and commissioned at the same time, running side-by-side, in the same ambient temperature, in the same ambient (barometric) pressure, on the same fuel at Base Load, with the same exhaust temperature control constants, and the output of the two machines is different? Why aren't they both producing the same amount of power with the same exhaust temperature control reference and exhaust temperature?"

The answer is a question: Are you sure that the inlet air filters have exactly the same pressure drop, the axial compressors are identically clean (or dirty) with the exactly the same air flow characteristics, the IGVs are at identical positions (meaning the LVDTs of both machines are calibrated properly), the CPD transducers are identically calibrated, the hot gas path parts have the exact same amount of "wear" and internal leakage(s), and that exhaust duct back-pressures are identical for the two machines?

If one thinks long and hard about the above, one can see that the most important things that affect power output are: Inlet filter cleanliness; axial compressor cleanliness; IGV LVDT calibration; and CPD transducer calibration. (If the unit uses CPR bias, the calibration of the ambient pressure transducers is also very critical.)

All of these things are things that can be "adjusted"; the condition of the hot gas path parts and compressor blades (the internals of the machine) can't really be "adjusted"; they are what they are. But, inlet air filters can be cleaned or replaced; axial compressors can be washed (on-line or off-line); IGV LVDTs can be calibrated accurately; CPD transducers can be calibrated accurately.

Hope this gets you started on your understanding.

 

Thanks very much CSA. Now I have a clearer picture on the BBL. I have also found out from my control specs that L83JTn references to different control curves where 0=base load simple cycle, 1= peak load simple cycle, 2= combined cycle base load and 3= combined cycle peak load.

However I noticed that the isothermal limit setpoint is very much higher in base load open cycle (1100F) compared to other mode of operation where peak load simple cycle =1045F; base and peak combined cycle =1050F. As far as i know(pls correct me if not) the thermal efficiency for a combined cycle operation is higher when the calculated firing temp (which is is higher for a constant CPD.

Another thing is let say the CPD is at 10 bar and the exhaust temp is at 1050F. So when I increase the TTK2_I to let say 1060F, will the increase in the allowable exhaust temp will make the increase the firing temp or will it shift the entire -ve slope curve upwards or maybe change the slope of the curve to maintain the constant firing temp?

I'm studying the exhaust temp control-control setpoint diagram now from the gek106902. Will come back with more questions if you dont mind

 

Thanks for the feedback, and my apologies for the mistake on the logic signal name; I was doing that from memory as I didn't have any CSP to refer to at the time.

*PLEASE* *PLEASE* *PLEASE* do not be changing any exhaust temperature control constant, ***EVER***.

I had always argued that these Control Constants should be "locked" to prevent people from making changes to them since they will affect machine operation and changing them could result in a catastrophic failure. But, my pleas fell on deaf ears.

I'm surprised to see different Isothermal constants; and wonder if they're not related to the ability of the HRSG to withstand the higher temperatures.

!!!Those Control Constants should *NEVER* be changed without consulting GE or the machine's packager!!!

*PLEASE* don't make any changes to those Control Constants without consulting GE or the machine's packager. There are too many other factors to consider when changing these particular Control Constants, and they can cause serious damage to the machine.

Those intersecting straight lines are all calculated to have smooth transitions from one curve with one slope to another curve with another slope. A change to one parameter can have knock-on effects for other curves.

 

But CSA you havent answer my question yet on how the TTRX curve would behave "should" the constants are changed just for my own knowledge.

 

It's better to do this for one's self, and it's difficult to do it without being able to do so with a graph.

The formulae are there.

The formulae are of the form: f(x) = b - (m * x) (because the slope is negative).

One can create an MS-Excel spreadsheet to do all the calculations if one doesn't want to to them manually, even plotting the results.

I question the use of "simple cycle" and "combined cycle" descriptions for the Control Constants. Do not *EVER* trust the descriptions in the Control Specification or LONGNAME.DAT as *the* correct description for what the Constants actually do or how they're actually used. What they actually do and how they are actually used in the CSP are the only real "descriptions" that count, not some text message.

I believe you told us the units were dual fuel (i presume that means gas-distillate (liquid); not dual gas). Usually, -0 and -1 are for gas fuel, and -2 & -3 are for liquid fuel. It's not common for the same constants to be used for both gas and liquid fuels (and, again, I'm presuming dual fuel means gas-distillate (liquid)).

Also, you didn't tell us if the units have a bypass stack and diverter that can be used to operate them in simple-cycle mode. If there's no provision for bypassing the HRSG, it wouldn't make sense to have control constants for simple-cycle mode, and there aren't a lot of combined-cycle plants with Frame 9Es with diverters and bypass stacks, because they are very large and costly and consume a lot of space.

As for thermal efficiencies, the over-all thermal efficiency of a combined cycle plant is higher than that of a simple cycle plant. The individual component efficiencies don't always follow directly, so be careful when comparing apples and oranges.

But, I cannot stress enough: Do not use your turbine(s) as a lab for testing what will happen. Explaining how exhaust temperature control works was not intended as a how-to for changing exhaust temperature control settings. Even though these Constants can be changed in the field, does *NOT* mean they should be changed.

 

Dear CSA

I would like to understand how this Constants are selected for particular Gas Turbine. We have 02 Nos of Frame-VI Running machine on GAS with MARK-IV.

How GE desides value of TTKn_C(bar) & _K(%FSR) for Base Load.
My question in detail is:

During Base Load,CPD is higer than TTK0_C(i.e. TTRX has been reduced by some extent from TTK0_I) now if the CPD transmitter fails, the new TTRX,based on %FSR,will come in picture. To have bumpless transfer of TTRX one must know that what should be the value of FSR at that reduced TTRX(which is based on TTK0_C & TTK0_S).

Thanks in Advance.

 

GE uses a very powerful program running on a very large mainframe computer to run a very long (sometimes more than 24 hours) series of "performance runs" based on expected fuel characteristics, type of inlet filter, length and type of inlet duct, type of exhaust, expected ambient temperatures, expected ambient pressures, expected ambient humidities, and such.

They have extensive databases from thousands of performance tests of turbines from all over the world. They have extensive databases from their combustion laboratory experiments. They have extensive experience with turbine materials and coatings and cooling and various auxiliaries and various fuels.

The program factors all of this information into its "performance runs" when calculating the exhaust temperature control constants. It's not something one can do on their PC in their office.

The back-up FSR calculations depend on accurate expected fuel characteristics and proper calibration of the Gas Control Valve LVDTs. Again, the back-up exhaust temperature control curve is *NOT* intended for long term operation; it's just for emergency situations when the unit has to be operated with failed CPD transducer(s).

If there is a loss of CPD feedback during exhaust temperature control, the "switch" to the back-up curve is "bumpless" because there is a programmed ramp rate for FSR changes to prevent sudden "bumps" from causing problems.

 

i will try 2 plot the curves and get back 2u..

as for your info, my units do have bypass dampers but its permanently locked in closed position as there were causing alot of nuisance trip in the past.

 

Dear CSA,

I have gone through the this thread, and also seen your posts in other threads on temperature control.

I have understand that after going through thread that in Part load, exhaust temperature decreases with decrease in CPD and Increases with increase in CPD.

While in temperature control mode, exhaust temperature decreases with Increase in CPD and Increases with decrease in CPD, So am i right till now???

Now (As i saw in Mark VIe toolbox)that formula for TTRXP and TTRXS is as shown below (Taken from post of Mr CSA):

The formulae are:

TTRXP = TTKn_I - ((CPD - TTKn_C) * TTKn_S)

TTRXS = TTKn_I - (([FSR or DWATT] - TTKn_K) * TTKn_M) (since the back-up is either FSR- or load-, DWATT, biased)

My question is during Base load (Temperature control)how can TTRXP (Equivalent to TTXM, as TTXM follows the TTRXP) can decrease while if CPD decreases.

By formula it seems almost proportional (as TTKn_I, TTKn_C and TTKn_S are constant).

Please correct me if i am wrong.

 

Aliya,

First you state:

>While in temperature control mode, exhaust
>temperature [exhaust temperature is TTXM!]
>decreases with Increase in CPD and Increases
>with decrease in CPD, So am i right till
>now???

Yes; you are 100% correct.

Then you write:

>My question is during Base load
>(Temperature control)how can TTRXP
>(Equivalent to TTXM, as TTXM follows the
>TTRXP) can decrease while if CPD
>decreases.

What part of yourself don't you understand?

You properly stated that exhaust temperature (TTXM) will follow TTRXP while at Base Load, which means that if TTXM decreases as CPD increases then TTRXP will decrease as CPD increases. And you said that TTXM will decrease when CPD increases when the unit is at Base Load.

You are contradicting yourself!

<b>DO THE MATHS.</b> If you agree that TTXM will follow TTRXP when the unit is at Base Load, and that TTXM will decrease as CPD increases when the unit is at Base Load, the rest should be self-explanatory.

The maths will tell you that TTRXP will decrease as CPD increases. All you need to understand (which you should have determined by now) is that you need to choose a value of CPD that is greater than TTKC_[0] (since until CPD is greater than TTKC_[0] TTRXP will be equal to TTKI_[0]. And, yes it is proportional. y=mx+b, only in this case 'm' is negative, so the equation becomes y=b-mx, which is the same thing as y=(-mx)+b.

Correct?

<b>HAVE YOU WATCHED TTRXP WHILE AT PART LOAD?</b> You will see that once CPD is greater than TTKC_[0] that TTRXP will <b>DEcrease as CPD INcreases</b>, and that TTXM will <b>INcrease as CPD INcreases.</b>

Once TTRXP and TTXM are equal, the unit will be at Base Load, on Exhaust Temperature Control! And then, TTXM will follow TTRXP, and that means (drum roll, please!): <b>TTRXP and TTXM will DECREASE AS CPD INCREASES, and TTRXP and TTXM will INCREASE AS CPD DECREASES</b>.

It is counter-intuitive, <b>BUT</b> the exhaust temperature "curve" has a <b>NEGATIVE</b> slope, which means that TTRXP, and TTXM when the unit is operating at Base Load on Exhaust Temperature Control, decreases as CPD increases

The only other piece of information that you will need to determine from the application code running in the Mark VIe at your site is how the array values are switched. That means you need to determine how L83JTn is changed as load is changed. The _[0] array values are used when L83JT0 is a logic "1". When L83JT1 becomes a logic "1", L83JT0 goes to a logic "0", and then the _[1] values are used. When L83JT2 goes to a logic "1", then L83JT1 goes to a logic "0" and the _[2] array values are used. Usually, there are comparators that monitor CPD as as CPD changes, they change the value of L83JTn.

<b>YOU HAVE TREND RECORDER ON YOUR MARK VIe, FOR HEAVEN'S SAKE! USE IT!!!</b> If you won't do the maths for yourself, let the Mark VIe do it for you, and record the information for later analysis.

If you would just do the maths, or, if you are really against doing the maths let the Mark VIe do it for you, you are quickly going to find out that you are making this much more difficult than it is.