Frame 7EA Post Outage Emissions Trouble

J

Thread Starter

Justin

We just came out of a month long outage to repair the dreaded Frame 7EA bucket rock. All three rows of turbine buckets were repaired/recoated. the 2nd and 3rd stage turbine nozzles were repaired/recoated and we put in a new/used (2055F rated) 1st stage nozzle. We also put in refurbished fuel nozzles, TP's and liners that were in the machine a few years ago. The turbine is early 90's vintage, steam injected, 2035F firing temperature with MARK IV controls in combined cycle. Since coming out of the outage we have been unable to meet full load NOx emissions and have been reduced to 90% load to maintain emissions compliance. In addition to this our exhaust temperature is 20F higher than design for this ambient temperature and our compressor discharge pressure has jumped up (pressure ratio went from 11.3 to 12.9). All of the wheelspace spreads are within limits and steam injection is running at pre-outage levels. Has anyone experienced this before? How can the machine appear to be over-firing when it should be running on the MARK IV exhaust curves?
 
Was an off-line compressor water wash performed immediately before or after the outage?

Were the turbine air inlet filters changed during the outage?

Was there any work done in the gas turbine exhaust during the outage?

Have you verified the compressor discharge pressure transmitter output to the Mark IV? If so, how? With what gauges?

Are they any Process- or Diagnostic Alarms active when the unit as at or near Base Load? If so, what are they?

You have quite a mix of parts in the unit, and those first-stage nozzles may be part of the problem (meaning that it might be prudent to have someone recalculate the exhaust temperature control curve).

By how much is the emissions out of compliance? Is it the NOx or the CO or both?

It's not unusual to have to make minor adjustments to the injection control curves after a major outage. And if the emissions monitors were "calibrated" or had some maintenance work, it's possible that might be part of the problem (meaning that it may have been slightly out of calibration before the outage started).

Have you made any adjustments to the injection control curve to try to see how much additional injection is required to meet emissions compliance? If so, how much?

Again, it would probably be a good idea to get someone knowledgeable with gas turbine innards to evaluate what kinds of changes the new nozzles might make to the compressor pressure ratios and flows and exhaust temperatures.

As for whether or not it's running on the exhaust temperature control curve, unless there's a problem with the CPD transducers or the unit is running on the back-up exhaust temperature control curve, which is another possibility) it most likely <b>*is*</b> running on the setpoints in the Mark IV. Whether that's the primary exhaust temperature control curve or the back-up exhaust temperature control curve. Digital Speedtronic panels are as good as the inputs, so as long as the inputs to the exhaust temperature control calculation are good, then the Mark IV is controlling to the setpoints programmed in the control system.

Remember: The actual firing temperature isn't being measured, it's being approximated by monitoring the compressor discharge pressure and the exhaust temperature, and controlling to the parameters programmed in the Mark IV which should represent a constant firing temperature when at Base Load. Presuming all the inputs are correct, and all the internal flows, pressures, and temperatures are correct.
 
We have not performed a compressor water wash since September due to our limited run time (typically run 1500 hours/year) and this current outage. We did replace the cap and cowl on the exhaust plenum due to our old one rotating out of place over time. I'm not sure which meter they used to test the CDP sender but it was tested and found to be dead on immediately after finding this problem. Our CO is fine but our NOx is anywhere from 1-2 ppm high (this is with increased ammonia flow). We did a test run today and slowly increased steam injection flow but even at 20 lbs/sec (up from our usual 16 lbs/sec) we could not get the machine over 87 MW (air temp 25F) without having the CEMS go into alarm for High NOx. Interestingly even with our CDP measuring 190 psig and our exhaust temperature seemingly higher than it should be we get no alarms and the turbine runs along fine. Any thoughts on the combustion liners causing this if they weren't coated properly and holes were plugged? I realize this machine is a mix match of parts which is an unfortunate legacy we have to live with for now since this machine was left for dead after it's PPA days ended.
 
Something seems to be amiss. I don't think I've ever seen 190 psig on a 7EA, even with compressor inlet temperatures below 30 deg F. Isn't there a CPD gauge on the Acc. Gage Cabinet? Has thought been given to installing a calibrated gauge to measure the CPD and compare it to the CPD feedback?

Again, is the unit running on Primary- or Back-up exhaust temperature control? The usual signal names are TTRXP (for primary) and TTRXS (for back-up, or, secondary). TTRX should be the lower of either TTRXP or TTRXS. Sometimes, there is a bias added to either, but not always. In general, TTRX should be equal to one or the other of either TTRXP and TTRXS. If it's equal to TTRXS, then the unit is operating on back-up, or secondary, exhaust temperature control. (If TTRX is equal to TTRXP, then it's operating on primary, CPD-biased, exhaust temperature control.)

The Steam Injection flow-rate reference is based on the fuel flow-rate feedback. I've seen problems with the fuel flow-rate feedback cause problems like this, as well. Improper calibration of the fuel flow-rate transducers, or drift, is usually the culprit.

Also, there is usually a Humidity Monitor, and that can have an impact on the NOx Injection Flow-rate reference as it biases the flow-rate reference.

I'm concerned because you keep saying that the exhaust temperature is higher than normal with an increased CPD.
<pre>
|
|
|************
| *
TTRX | *
or | *
TTXM | *
| *
| *
| *
----------------------------------
CPD (increasing ->)
</pre>
A look at the exhaust temperature control curve shows that it has a negative slope. That means that as CPD increases, the exhaust temperature should decrease <b>*for the same firing temperature*!</b> (The sloped line represents a constant firing temperature, the 2035 deg F you mentioned, for varying conditions (ambient temperature; compressor cleanliness; etc.).

There are two exhaust temp references being calculated all the time: the primary and the secondary. The slopes should be similar, but offset slightly. And there is a MIN SELECT block that should be choosing the lesser of the two for the exhaust temp referencer, TTRX. (Again, there sometimes a bias added to the signal, but not always. The biased signal name is usually TTRXB.)

I'm not a hardware person. Coatings are usually there to protect the metals from higher temperatures thereby prolonging their life. I don't know how much, if any, they contribute to the pressure drops in the combustion system. Certainly plugged holes would be a problem.

But, I keep going back to the concept that as CPD increases exhaust temperature should be decreasing, per the negatively-sloped curve. You keep saying that the exhaust temperature is higher than before with an increased CPD, and that just doesn't seem to compute (for me, anyway).

Lastly, I would think that the fuel flow-rate for Base Load would be approximately 15 #/sec, or higher for a low ambient temp. Steam Inj flow-rates for 25 ppm would normally be about 1.3-1.5 times the fuel flow-rate (these are typical values, not exact values for any site). We don't know what fuel the unit is operating on, or what the emissions guarantee is.

But, something don't seem quite right.
 
For your Mark IV, I don't think that just by selecting peak temp (from 2020 to 2035 or 2055) will modify firing curve. This was the case with our old M4 control anyway. We had to manually go and change the Corner and constant values to modify the curve. But in our M6e once you select peak firing temp reference on HMI, program automatically redefine (by selecting 0,1, or 2) the curve. we were given constants from GE to change once we installed M6, I mean constants for same machine were different for M4 and M6 control.

Your wheelspace temps are good so there is less chances that cooling of first stage and second stage are affected. How about your fifth stage cooling / sealing air? Any pluggage there? If you go on back up control or assuming CPD transmitter fails, you will get alarm anyway. I have not seen 190 CPD on our F7E, again per our mechanical engineer if you go over 180 (in Alaska) there are chances for surge, and we have to shed the load. If engineering is true you are almost on surge line of compressor. If wheelspace temps are higher as compared to pre T/A, there may problem with brush seal or not enough air is bleeding of the 16th stage for cooling purposes of turbine nozzles and buckets. We have brush seals @ inner barrel and @ second stage bucket. Your holding pin for first stage shroud block is of which number? I think bigger the number more opening for air to pass thro. please post your findings, I am interested to know the remedy.

Thanks
 
Guys I can't stress it enough how much I appreciate the help. Let me clear a few items up with actual run data. During the test run on this machine this past Sunday we had the following two hour averages while at base load. Controls=Mark IV (no peak firing option), Design Firing Temp 2035F, Inlet Air Temp=31.18F, Humidity=55.1%, Baro Pressure=14.56 psia, IGV Angle=86.01 deg, Natural Gas Flow=12.65 #/sec, Steam Inj Flow=20 #/sec, Exhaust Temperature=975.18 F, CDP=185.47 psig, CDT=646.31 F, GT exhaust NOx=24.26 ppm and lastly GT Pwr Output=91.47 MW. When the machine is running we do not get any alarms regarding CDP which to me seems odd since as stated above it risks surge above 180 psig. Our NOx limit out the stack is 9ppm which we can maintain at loads around 85 MW but when we attempt to go to base load at 92 MW it spikes up to around 13-14.
 
And now for another overly long response....

NOx formation is a function of diffusion flame temperature, so if the NOx is high then the flame temperature (and likely the firing temperature--the temperature of the combustion gases exiting the first stage nozzles of a Frame 7E or -7EA) is higher than normal, or at least higher than that calculated to obtain the original design firing temperature of 2035 deg F.

But, you haven't told us what the GT exhaust NOx is (exhaust/stack "outlet"), before the SCR where ammonia is being injected. If you don't have a GT exhaust NOx analyzer (exhaust/stack "inlet") then it would seem you don't really know if the GT NOx is higher than normal or not.

And, I'm not any kind of SCR/catalyst expert, but I think they are somewhat temperature sensitive, and you have said that the GT exhaust is higher than before the outage for the same ambient conditions.

Something sure seems not quite right. You have said your stack outlet NOx limit is 9 PPM, but we don't know what the GT exhaust guarantee is. In my experience it would usually take somewhere around a 1.2-1.3 steam inj-to-fuel ratio to meet an exhaust NOx guarantee of 25 PPM, sometimes a little higher.

The high CPD sure seems unusual, and I don't recall any Frame 7EAs with Mark IVs that had a max IGV angle of 86.0 DGA as the original setting. What is the value of CSKGVMAX in the Control Specification document? Do you know if an uprate was purchased sometime during the life of the machine that would have included an increase in IGV angle if the Cont. Spec. value is less than 86.0 DGA?

Back in the mid-1980s there were a lot of failures of 17th stage exit guide vanes during cold ambients when CPDs would be very high. I don't recall any surge incidents, though; just EGV failures (which were pretty catastrophic).

But I still keep going back to the negative slope of the CPD-biased exhaust temperature curve. A common complaint from new combined-cycle power plants being operated at Base Load is that during the winter the GT exhaust temperature drops, and during summer the GT exhaust temperature increases, which is what happens when ambient temperature decreases (which increases CPD) and when ambient temperature decreases which decreases CPD)--per the negatively-sloped exhaust temp control curve. They think the GT exhaust temp should increase in the winter and decrease in the summer.... (It's counter-intuitive, but that's the facts.)

And, you still haven't told us if the unit is operating on the primary exhaust temperature curve or the back-up exhaust temperature curve.

As for CPD alarms, I don't believe many of the Mark IV turbine control systems had any compressor margin protection algorithms. A lot of that was implemented beginning with the Mark V which had more compute horsepower and memory. Some Mark IVs were modified for the 17th Stage Compressor Protection to actually limit output if ambient conditions exceeded capability.

Most of the GE-design non-F class fleet are pretty robust machines (not that the F-class machines aren't robust; they are just operating on the limits of materials and design whereas the other machines had much more margin and could take more "abuse").

Lastly, there's an awful lot we don't know, and can't know, without looking at the sequencing and Control Constants for your machine. And, there's that hardware "issue" (I'm just bringing it up because I don't have any experience with using 2055 deg F nozzles in a 2035 deg F machine, but it would seem to raise a flag--at least for me). It would really seem prudent to get some help from an engineering company familiar with GE machines and the combustion hardware you have installed in the machine. And to know what the GT exhaust NOx is (not the stack outlet NOx).
 
I have seen exhaust NOX almost 40 to 50 PPM on our GTG (7-E), but we control stack emission with NH3 injection @ SCR catalyst. Our stack Nox runs about 2 to 3 PPM. IGV angle at our machine is about 83 to 84 degrees. At higher IGV angle it is inducing vibration at inlet (Cantilever fashion), We had M4 before we updated to M6e, and I am sure that one has to go M4 (sheet# 28C) and change the slope and corner. TTKn_K and TTkn_M to get TTRXS (if on back up) or TTRXP (CPD TEMP REF) high so firing temp is referred to 2035 or 2055 unless i am referring to some old M4. Make sure that all the constants are same on your sick GTG and other good GTG. just compare the constants.
Thanks
 
Justin,

My bad for not seeing the GT exhaust NOx reading in the reply. 24.6 ppm would be less than 25.0 ppm which is likely the NOx guarantee for the machine when firing at 2035 deg F. And if I re-read the post again, correctly this time, the 24.6 ppm NOx reading was at 91.47 MW, which is approximately the 92 MW Base Load rating.

So, it would seem that the GT exhaust NOx is within guarantee, achieved with a steam-to-fuel injection ratio of (20 #/sec / 12.65 #/sec =) 1.58, which is kinda high. Was this high steam inj flow-rate achieved by biasing the steam inj reference higher than the calculated value? If so, it should be noted that high NOx steam inj flows can excessive pressure pulsations in the combustors, leading to flame instability (and even loss of flame trip). Continued operation with excessive pressure pulsations will eventually cause premature failure of hot gas path parts (liners, hula skirt seals, transition pieces).

To sardar9, the original poster hasn't mentioned Peak firing, only that 1st stage turbine nozzles for a machine with a 2055 deg F firing temp were installed in a machine with a 2035 deg F firing temp. And, I can't find a reference to a second "healthy" machine in the post.
 
Top