Exhaust Thermocouples Flagging Red and Yellow


Thread Starter


Please, I need help on this problem. we are using GE frame VI, with speedtronics Mark Vi turbine controller. The problem with about two of the machine is that whenever it's on base load one of the exhaust thermocouples will be showing red, and one or two more will be showing yellow although, the effect did not TRIP the unit. during system outage, the affected thermocouples were changed. when the unit was later run up, we found out that the problem is still there.

Then for second, we use multimeter to test the D.C. resistance of the thermocouple. we got 6.1 ohms and 7.5 ohms for the ones that are showing yellow and red, while the other ones that are not showing give us value of 8.7 ohms, 8.5 ohms, 10.3 ohms etc., and the resistance of the new ones lie within the range 5.6 ohm to 8.1 ohm.

Please i need your assistance on next step to take to solve this problem.
Basically, you're saying you believe that the Mark VI should be tripping the GE-design Frame 6 heavy duty gas turbine because the individual exhaust T/C bargraphs are turning red and yellow when the unit is at Base Load.

GE, in their (not so) infinite wisdom (mistakenly) thought it would be a good idea to change individual exhaust T/C bargraphs whenever they exceeded the generic alarm and trip values for "median" turbine exhaust overtemperature protection of the turbine. Let me explain why that was a complete bozo decision.

Exhaust overtemperature protection for the turbine uses the "median" exhaust temperature (and it's actually the average exhaust temperature) calculation and compares that to the lesser of two alarm and trip values. The "median" exhaust temperature, TTXM, is calculated by ignoring any individual exhaust T/C values below approximately 500 deg F AND also ignoring the highest and lowest remaining individual exhaust T/C values, then calculating the average of the remaining individual exhaust T/C values.

So, let's say one exhaust T/C was failed and reading a negative value. Then the highest individual exhaust T/C value was 1141 deg F and the lowest individual exhaust T/C value was 1023 deg F. If the unit had 18 exhaust T/Cs then only 15 would be used to calculate the"median" exhaust temperature.

Now, there is an absolute maximum allowable turbine "median" exhaust temperature and it's usually 1140 deg F. AND there's a maximum allowable turbine "median" exhaust temperature alarm limit and it's usually 1125 deg F. BUT, there's also a possibly lower maximum allowable turbine "median" exhaust temperature alarm and trip limit that's a function of the exhaust temperature control reference, TTRX. (Isn't this fun?)

GE turns any individual exhaust temperature bargraph yellow if that individual exhaust T/C value exceeds the fixed alarm turbine "median" exhaust temperature value. And, they turn any individual exhaust T/C value red if it exceeds the fixed trip turbine exhaust temperature value.

The "median" calculated exhaust temperature value, TTXM, has to exceed the lesser of the turbine exhaust trip values. If the calculated exhaust temperature reference, TTRX, was 1087 degree F in the example above only one individual exhaust T/C value exceeded the fixed "median" exhaust temperature trip value so its bargraph turns red. And, if in our example two of the individual exhaust T/C values were above 1125 deg F then their individual exhaust T/C bargraphs would turn yellow. But if the calculated "median" exhaust temperature, TTXM, was 1087 deg F (equal to the calculated exhaust temperature reference, TTRX) then the turbine "median" exhaust temperature is neither in alarm nor trio--even though one bargraph is red and two are yellow. And the red bargraph's value was thrown out by the "median" exhaust temperature calculation.

Pretty stupid of them to use the "median" exhaust temperature alarm and trip values to change the individual exhaust T/C bargraph values yellow and red, wasn't it? Made for a potentially colorful display, but it only causes a LOT of grief and consternation.

Lastly, when exhaust temperature spreads are being calculated the highest- and lowest individual exhaust T/C values greater than 500 deg F are NOT thrown out. That would greatly skew the spread calculation! The highest- and lowest individual exhaust T/C values greater than 500 deg F are only thrown out when calculating the "median" turbine exhaust temperature.

So, if you're worried the Mark VI is not protecting your turbine because some idjit used the wrong values to change the individual exhaust T/C bargraphs red and yellow: Don't be. But DO take the time to learn more about how the "median" turbine exhaust temperature and exhaust temperature spreads are being calculated. And, at the first opportunity remove the individual exhaust T/C bargraph animations from the CIMPLICITY display. (It's not called STUPIDITY for no reason (as CIMPLICITY is commonly referred to by field service people).)

Hope this helps to reduce your fear and anxiety!!! (And doesn't increase your blood pressure and aggravation too much.)
But after removing those thermocouples and install new ones. The problem is still there. What might be the problem?
A LOT of sites never check their T/C wiring all the way from the exhaust JB to the Mark VI terminals. There are VERY specific ways T/Cs have to be connected to T/C extension wire, and if the unit weas provided by GE then special T/Bs are used and there are VERY specific ways T/C and T/C extension wires have to be terminated on those terminal boards. Incorrect terminations anywhere along the length of a T/C circuit can cause unintended cold junctions, and can affect the readings. ALL exhaust T/C circuits should be checked, not just the ones you are experiencing problems with.

A LOT of sites also neglect the radiation shields in the exhaust and they can be bent or have broken welds and heat-caused damage that can also affect exhaust T/C readings.

It's also VERY possible that there is a combustion problem with one of the fuel nozzles. We don't know how the affected T/Cs are located with respect to each other.

There are several old threads with some "drawings" of how T/Cs and T/C extension wiring has to be terminated, with some "drawings" of the typical T/Bs provided by GE. Also, check the exhaust T/C radiation shields in the exhaust. The problem is likely some combination of variables, including a fuel nozzle problem.
Thanks for your contribution

Although, one of my colleague was saying something about the cleaning of the fuel nozzle. Please can you tell me how that might be connected to the problem.

So, most everyone assumes when an exhaust temperature spread is increasing or develops the problem simply MUST be the turbine control system! After all, just look at all those wires and flashing LEDs and cables--the problem surely must be in the turbine control system.

Well, when they get done trying to prove the problem is in the control system (which it usually isn't!) the next step in any logical process of troubleshooting is to move one to the next possible cause.

Exhaust temperature spreads are normally caused by combustion system problems. Plugged fuel nozzle orifices can restrict the flow of fuel out of the fuel nozzle--however, this will usually cause a "cold" spot and your original post describes a HOT spot. You haven't told what kind of combustion system is in use on the GE-design heavy duty gas turbine(s) at your site, and if it's multi-fuel or single fuel. If it's multi-fuel, what are the fuels that are burned--and how often do you transfer fuels?

I have seen, only twice in more than three decades in the profession, problems with fuel quality affect nearly all the fuel nozzles except for one or two, and those one or two weren't plugged and resulted in a hot spot.

The topic of exhaust temperature spreads has been covered MANY times in previous threads on control.com--all accessible by using the cleverly hidden 'Search' field at the far right of every Menu Bar at the top of every control.com webpage. (It's suggested you use the Search 'Help' feature the first couple of times you search for something; it's not like most World Wide Web search engines--but it is fast.)

The thing to know about exhaust temperature spreads on a turbine with multiple can annular combustors (such as are used on GE-design heavy duty gas turbines!) is that a hot spot or a cold spot in a particular combustor or hot gas path travels through the turbine section "intact." In other words--as hard as it is to believe--the hot gases or cold gases do not get mixed with the gases from either side of the combustor(s) having the problem. That makes it possible to detect combustion problems (hot spots; cold spots) and protect the turbine from the damaging effects of excessive exhaust temperature spreads, usually by tripping the turbine (because the operators and their supervisors don't know what to do when an exhaust temperature spread develops and is increasing). The turbine control system will only allow an exhaust temperature spread to get to a certain magnitude, and when adjacent exhaust T/Cs also indicate a true combustion problem it will trip the turbine.

I DO NOT know if you have an exhaust temperature spread problem or not; you have not provided enough information about the location of the red and yellow bargraphs. And, the type of combustion system. And the type of fuel(s). But, if one or more fuel nozzles were plugged--then that would result in an exhaust temperature spread.

And, if the colleague who mentioned plugged fuel nozzles has worked at the site for some time and has had previous experience with plugged fuel nozzles, well, then, it might good to listen to his advice.

Given your information from above on TTXM and trip/alarm points, am I understanding you correctly to say that a trip/alarm will occur if TTXM exceeds the appropriate trip/alarm point but an individual thermocouple reading that exceeds those trip/alarm points won't do anything except light up the HMI screen and possibly provide a "heads up" to alert an operator to pay close attention when an abundance of red/yellow is lit up? If that is the case would it not be better to use TTXM to light up some kind of warning if exhaust "median" is approaching the trip/alarm point and save the red/yellow on the individual bargraphs for something more useful like highlighting TC's with a high spread or something? If so, I may edit our screens to do just that.

BTW our TTXM is described as "Exhaust Temperature Median Corrected by Average". Note to GE: if the calculation CSA described above is correct (and CSA is pretty reliable so I don't have any reason to doubt it), TTXM would be better described as "Exhaust Temperature Average Excluding Outliers".

You understand me exactly 100% correctly. Of course, if several exhaust T/C bargraphs exceed the alarm or trip setpoints the 'Exhaust Temperature Average Excluding Outliers' would be elevated and cause an alarm or trip--but just because one or two bargraphs turn yellow or red does not mean the turbine should be tripped. It's VERY misleading (the changing of bargraph colors based on an "average" value).

And, GE doesn't read posts on this forum. They are much too above that. Individual GE employees might read posts on this forum, and may even reply to posts, but there's no one at GE that is actively trying to improve the user/operator experience for GE-design heavy duty gas turbine owners/operators by looking at World Wide Web forums or feedback from Users Groups. If it wasn't invented at GE, it doesn't exist and doesn't affect anything GE does or might do. Some of us are SHOCKED that someone at GE might actually have read 'The High Performance HMI Handbook' and actually tried to implement some of the recommendations in that book. But, that's as far as it's gotten.

As I was told decades ago, after the Mark IV was developed and went into production, TTXM was originally a median value, but testing and empirical data proved that using the median value from many values wasn't very reliable for protection. So, TTXM actually underwent several iterations before it arrived in its present incarnation--the average of the working exhaust T/Cs after "rejecting" non-working T/Cs and the highest- and lowest values ("outliers"). But, the signal name never changed from TTXM.

Finally, GE has NO standard for longname descriptions of signal names. Anyone who can edit an ASCII text file is free to change or edit any longname description as they see fit, because no one is going to check what was written to see if it's appropriately descriptive or needs improvement or correction. Longname descriptions are the wild, wild West--in the East (well, Western Virginia and Western South Carolina and Upstate New York).
All these GE names do have "Longnames" associated with them, TTXM is originally Turbine Temp. Exhaust Median.

You have a High Exhaust Temp. Alarm followed by Trip, you have a High Spread Alarm followed by a Trip. Control systems are usually designed to not give useless or confusing information, pretty well that is what GE has done here.

Hope this helps
Looks like I'll be making some changes to our screens. I appreciate what they were trying to do, but it just puts us in a "boy who cried wolf" situation and ends up being ignored.

BTW, I've seen GE's "implementation" of HP-HMI and I don't much care for it. The GE folks I talk to all call it the Grayscale HMI. I can see why, because it seems the only thing they took from High Performance was the grayscale color scheme. They still have screens that cram so much information in that they make it difficult to determine what data on the screen should be prioritized by the operator and actually make it harder to read due to shrinking things down to smaller fonts and the color scheme blending too much. Also, the alarms are just so tiny and do nothing to stand out and draw attention (which is supposed to be their purpose). Ugh, anyway, I'm glad we have the ability to change them to suit our sites.
The ActivePoint HMIs are, in my personal opinion, a step up from the previous HMIs--but that's purely a subjective opinion. I haven't used one yet, but I kind of like the grey-scale screens as I feel the others were poorly designed and animated and without the ability to hover over data values and see what signal they were coming from it was all just a mess.

Good luck with your changes! I would think it would be possible to add a TTXM bargraph, and animate it with red and yellow changes for alarm and trip. I would also add a TTRX bargraph (no color animation necessary.) You are aware that when a unit is operating at Base Load that TTRX and TTXM will be equal--which means TTXM is 25 deg F away from the alarm setpoint, and 40 deg F from the trip setpoint. So, it seems kind of hard to alert an operator to potential problems that way.

And, with many machines running in combined cycle mode and/or with DLN combustors TTXM is usually very close to TTRX even at part load, with the same 25- and 40 deg F alarm- and trip differences.

Training and experience are what's really important for operators. I find many operators think the HMI displays are too busy. I know of several who have made their own displays, and ended up with a LOT more data and buttons on them than on the original displays they complained about. But, they usually didn't have a lot of graphics taking up space.

I think what's SORELY missing from GE Mark* HMIs are P&IDs. I have seen a very recent unit which had a few of the P&IDs (GE calls them Piping Schematics) on the HMI, but they were different from the actual paper drawings and were quite confusing. But, I also FIRMLY believe that EVERY control room operator and instrument & control technician should have their own, personal set of P&IDs--with their own notes on them. There's nothing like understanding the systems when it comes to operating and troubleshooting power generation equipment. And, having knowledge of the combustor construction, especially for DLN combustor-equipped units, is also critical to operation and troubleshooting.

Anyway, I digress. A lot. Let us know how you fare in correcting the exhaust temperature display! I'm sure a LOT of people will be interested in what you come up with.