We are near the end of a major inspection / maintenance period for one of our two GE frame 6 turbines. During this outage, GE has upgraded the honeycomb seals at the third stage of the turbine.

This modification requires new temperature control curve setpoints. In the Mk IV, we have two curves: TTK0 and TTK1. TTK0 is the Base curve, and TTK1 was added with a 'PEAK' operation modification sometime after comissioning, but far before I arrived.

GE at first was unsure about what our TTK1 was for, I told them about the peak modification. I verified this and sent our complete 1D list to GE. Here's what were our control curve constants:

TTK0_C 135 psi
TTK0_I 1055 F
TTK0_K 72 %
TTK0_M 2.49 F/FSR
TTK0_S 2.02 F/psi

TTK1_C 139.1 psi
TTk1_I 1070 F
TTK1_K 72.1 %
TTK1_M 2.37 F/FSR
TTK1_S 1.62 F/psi

So, my understanding is that these 2 sets of 5 constants make up our Isothermal (I), CPD BIAS (C and S), and FSR BIAS (K and M) curves for both BASE and PEAK operation. Similar to what I have on Pages 1 and 6 of this pdf ( http://www.auto-ic.com/CURVES.pdf ).

I worked out the Y intercepts and came up with the following linear equations:

BASE

CPD BIAS -> Texh = -2.02 * CPD + 1327.7
FSR BIAS -> Texh = -2.49 * FSR + 1234.28

PEAK

CPD BIAS -> Texh = -1.62 * CPD + 1327.7
FSR BIAS -> Texh = -2.37 * FSR + 1234.28

This all seems pretty straight forward to me from a technician's perspective. But then I was handed GE's proposed new constants (they even added an equation at the top of each set of constants... presumably to help 'clarify things'):<pre>
CONSTANT NAME Description Value Units

BASE Texh = -27.5676 * Xc + 1355.0956

TTK0_C Exhaust Temp Control curve #1 CPD Ref Corner 134.2619 psi
TTK0_I Exhaust Temp Control curve #1 Tx Ref isothermal 1055 deg F
TTK0_K Exhaust Temp Control curve #1 OFFSET DW bias 65.8827 MW
TTK0_M Exhaust Temp Control curve #1 GAIN DW bias 2.6605 F/MW
TTK0_S Exhaust Temp Control curve #1 CPD Ref Slope 2.0270 F/psi


CONSTANT NAME Description Value Units

PEAK Texh = -22.1296 * Xc + 1317.2271

TTK0_C Exhaust Temp Control curve #1 CPD Ref Corner 138.151 psi
TTK0_I Exhaust Temp Control curve #1 Tx Ref isothermal 1070 deg F
TTK0_K Exhaust Temp Control curve #1 OFFSET DW bias 79.9542 MW
TTK0_M Exhaust Temp Control curve #1 GAIN DW bias 1.7624 F/MW
TTK0_S Exhaust Temp Control curve #1 CPD Ref Slope 1.6272 F/psi</pre>


First things first, I questioned GE's units on the K and M curve. That was apparently an oversight; they meant to say FSR, not MW or DW.

My first real question in all of this is, what is Xc in GE's equations above? I had to work out the CPD BIAS and FSR BIAS curves again with these new numbers because their slope was so steep... and their Y-intercepts were so close...

Here's what I got:

BASE

CPD BIAS -> Texh = -2.027 * CPD + 1327.15
FSR BIAS -> Texh = -2.6605 * FSR + 1230.28

PEAK

CPD BIAS -> Texh = -1.6272 * CPD + 1294.80
FSR BIAS -> Texh = -1.7624 * FSR + 1205.62

So the curves I work out from their points dont correspond to the equations that they have given with either y intercept or slope. I just can't wrap my head around what Xc represents. The intercept is *close* but not exact, and the slopes are nowhere close.

The other thing that I can't work out is this. While the CPD curves do not change much with this modification, the FSR curves do. The base load curves move in a more conservative direction. As FSR increases, we are more limited in temperature. That makes sense with better 3rd stage seals; more work is being done --- for a constant exhaust temperature, combustion temperature would have to be higher. Since we haven't done anything to allow higher combustion temps, the exhaust limit would have to be lower to take this into account. SO...

Why would the slope be smaller than before, allowing for higher temps for a given FSR? (see pdf http://www.auto-ic.com/CURVES.pdf page 4)

I'm looking for a general concurrence that this shift in curves makes sense or doesn't. And Xc.

 

There are several things that need correcting here.

First, the exhaust temperature control "curves" represents <b>constant FIRING temperature</b>, <b>**NOT**</b> constant exhaust temperature. The temperature of the combustion gases leaving the first stage nozzles is the firing temperature, not the exhaust temperature. The negative slopes of the curves mean that as CPD (or FSR) increases that <b>for the same, constant, firing temperature</b> the exhaust temperature will decrease. This is counter-intuitive, but it is the way it is.

The rating of the gas turbine is based on the <b>firing temperature</b>, <b>**NOT**</b> the exhaust temperature.

'Exhaust temperature control' doesn't mean they are trying to maintain a constant exhaust temperature. It means they are measuring CPD (or FSR) and exhaust temperature, and they know that there is a relationship between these parameters and <b>firing temperature</b> and that firing temperature can be held constant by varying exhaust temperature in response to changes in CPD (or FSR).

Second, the unit is only supposed to operate on <b>EITHER</b> primary (CPD-biased) <b>OR</b> secondary (FSR, in your case) exhaust temperature control. A primary- and secondary exhaust temperature reference is being calculated at all times, but only one exhaust temperature control reference (Primary or Secondary) should be active at any time.

Without being able to examine the sequencing in the Speedtronic at your site it very difficult to say when secondary exhaust temperature control becomes active. Some older Mark IV Speedtronic panels would not shift to secondary exhaust temperature control unless there was a failure of the CPD transducer(s). Most did a low-select of either the primary- or the secondary exhaust temperature control reference (usually signal names TTRXP and TTRXS, respectively) to decide which value was going to be TTRX (the exhaust temperature control reference).

The intent of the back-up, or secondary, exhaust temperature control is that it would never come into control unless there was a failure of the CPD transducers. AND, that if there was a failure of the CPD transducers that the unit could continue to run (while the CPD transducer problem was being quickly resolved!) until control could be reverted to CPD-biased (Primary) exhaust temperature control.

Third, for the primary exhaust temperature control curve (CPD-biased), the x-axis is CPD. For the secondary exhaust temperature control curve, the x-axis is FSR. It's very difficult to super-impose the two with any degree of accuracy, so it's very difficult to compare the two unless one knows precisely the correlation between CPD and FSR.

The intent is that the back-up exhaust temperature control curve should "mirror" the primary exhaust temperature control curve and be slightly offset from it (to the right of the primary exhaust temperature control curve) by a small amount, so that if the CPD transducer(s) failed causing the primary exhaust temperature control reference to be greater than the secondary exhaust temperature control reference that the unit would be able to continue to operate, albeit with a small increase in power output (which also slightly increases firing temperature, which negatively impacts hot gas path parts life--slightly, but it's implied that the CPD transducer problem will be quickly resolved).

The basic equation for exhaust temperature control is:

TTRXx = ISOTHERMAL -((BIAS - CORNER) * SLOPE)

where the lower-case 'x' is either P or S (Primary or Secondary).

TTKn_C = CORNER (PRIMARY). in PSIG
TTKn_I = ISOTHERMAL, in DEG F
TTKn_K = CORNER (SECONDARY), in % FSR
TTKN_M = SLOPE (SECONDARY), in DEG F/% FSR
TTKn_S = SLOPE (PRIMARY), in DEG F/PSIG

So, at all times, there are two calculations being made:

TTRXP = TTK0_I * ((CPD - TTK0_C) * TTK0_S))
TTRXS = TTK0_I * ((FSR - TTK0_K) * TTK0_M))

(The above is true when Base Load, the "0" array, is active at your site. When the Peak Load, the "1" array, is active at your site.)

I would say that the change in slope of the FSR curves represents a more modest attempt to prevent "overfiring" if the unit ever operated on secondary exhaust temperature control (which it never should!). Overfiring occurs because the FSR-biased, secondary exhaust temperature control curve is supposed to be to the right of the CPD-biased, primary exhaust temperature control curve. As the sloped line is shifted to the right, the <b>firing temperature</b> increases; as it is shifted to the left, the firing temperature decreases.

I hope it's clear that it's very difficult to have primary and secondary exhaust temperature control curves based on two different variables (CPD and FSR) be exactly equal to each other at all times. So, that's why the two are slightly offset from each other, with the secondary being slightly higher (to the right) of the primary. Again, the unit should never operate on the secondary exhaust temperature control curve, and if it does, it should only do so for a short period of time (while the CPD transducer problem is being resolved).

As for the Xc value, I've only recently seen this formula and notation, and I've not been curious enough to investigate. I'm not even going to hazard a guess, because I have no idea what it means, nor how these values are determined, nor how they are used to derive the TTKn_x array values. I would suggest you ask your GE field service person to ask someone in GE Engineering what Xc means, and how the values in the formulae are determined. (My guess is that this formula isn't supposed to be provided "to the field" and that it is considered proprietary and so it will be difficult to get information about its terms and usage. But, that's just my personal guess; I hope I'm wrong, and that you can find out what Xc means, and then explain it to all of us!)

Hope this helps! Again, the unit should only operate on Primary (CPD-biased) exhaust temperature control. Secondary (FSR-biased) exhaust temperature control should only be active in the event that there is some problem with the CPD transducers, and even then it should only be operated on secondary exhaust temperature control while the CPD problem is resolved.

And, finally, it's very difficult (at least for those of us without access to the data and computing horsepower that GE has!) to closely correlate (cross-reference; super-impose) the two graphs (CPD-biased and FSR-biased) on top of each other to see how they might interact. The best thing I can say about this is to monitor TTRXP and TTRXS when the unit is operating. And that TTRXS should be slightly higher than TTRXP at all times when operating on Base Load or Peak Load, so that the unit only operates on Primary exhaust temperature control.

Which really translates to firing temperature control!

Isn't this stuff fun?

 

Hi Chris,

I think Xc means compressor pressure ratio (Xc=Pout/Pin). The Pin, if you reverse calcutae from the coefficient of Xc, will arrive to 13.6 psi, which I think has taken into consideration of the total pressure drop due to intake and bell mouth.

The pressure ratio formula has been adopted when GE introduces MarkVI.

Regards,
Kevin
ITS-Industrial Turbine Services GmbH
info[at]turbineservices.asia

 

This doesn't precisely relate to the issue of primary vs. secondary exhaust temperature control, but I have a few questions on the general topic.

Background:
- 1x1 combined cycle with a Frame 5 with a 1991 vintage date.

- Control system is Mark IV and is original

- The unit has always operated at base load.

- The unit has always operated with a 100F increase in firing temperature. This is supported by an average GT exhaust temperature of 1000F while the GT nameplate is 900F.

- Over the past few years, the unit has averaged 40hrs/start.

A few questions:
- Does an average GT exhaust temp of 1000F corrolate to a 100F increased GT firing temperature? (I'm asking becuase I'm trying to determine the impact to the hours-based maintenance factor)

- How easy is it to change the firing temperature? Would it require heat balance engineering analysis? Better phrased: What would it require?

thanks in advance for your input.

 

MS5001fan,

I don't understand this statement:

> - The unit has always operated with a 100F increase in firing temperature.
> This is supported by an average GT exhaust temperature of 1000F while the
> GT nameplate is 900F.

I don't know the correlation between exhaust temperature and firing temperature (the temperature of the combustion gases leaving the first stage nozzle of a typical GE-design Frame 5 heavy duty gas turbine). I would tend not to think it is 1:1, but that's just a guess; I may be completely wrong. At any rate, I don't believe it's the same ratio for all GE-design heavy duty gas turbines based on some conversations I've had with Performance Test engineers over time--but I've never directly asked or read anywhere what the relationship is between exhaust temperature and firing temperature.

Every Speedtronic turbine control panel is provided with a Control Specification document which lists the parameters and setpoints used for controlling and protecting the unit. If you have a unit packaged and provided by GE of Mark IV vintage (presuming the unit was installed new with the Mark IV) you should find the Control Specification in Vol. I or II of the Service Manuals, usually in the back of the manual in a tab labeled 'Control Specification.'

Depending on the vintage of the Mark IV, there may be one or three documents comprising the Control Specification. If there are three, you would be looking for the one called 'Control Specification - System Settings.' In Sect. 03 you will find a listing of the exhaust temperature Control Constants (setpoint values) which were the original ones provided with the unit. My guess is someone changed one or more of the Control Constant values to achieve a higher firing temperature--and exhaust temperature. By comparing the original values to those running in the Mark IV you should be able to see what was changed, and what it should be.

The Mark IV Maintenance Manual provides instructions for how to change Control Constants, and how to make the change permanent (the latter depends on the vintage of the Mark IV). The Control Constants you are looking for are contained in an array of Control Constants, and they usually are named TTKn_x, where 'n' is an integer from 0 to 7, and 'x' is a letter like C, I, S, M, K, etc.

Hope this helps.

If the Control Constants in the Mark IV match those in the Control Specification, then you're going to need to contact GE or some other firm for more help with determining what to change and how to change it. If all the Control Constants are the same as those in the Control Specification then one of two things has happened: GE issues an as-built Control Specification drawing, in which case they should have issued a new nameplate, or, the logic/sequencing in the Mark IV was modified in a very unusual manner, and that will require someone knowledgeable in reading relay ladder diagram logic such as is used in the Mark IV to review and understand in order to make any changes.

Please write back to let us know what you find. Hopefully some other contributor will write back with some idea about the relationship between firing temperature and exhaust temperature. I've just always presumed it was not 1:1, but I've never investigated it more fully.

 

Xc is the pressure ratio.