Thermocouple reading shows high and Gas Turbine is going to trip?


Thread Starter


I work in GE Frame 9 Gas turbines of which control system is Mark IV. One of GT 2 (two) times trip due to No-3 Bearing Metal Thermocouple reading high.

No-3 Bearing has 2 thermocouples (Ungrounded): one BTJ31 & other is BTJ32. Though both are 2 elements thermocouples but only one element connected in the Mark IV. Problem is with BTJ32. It is sometimes going to high value and But BTJ31 is showing normal value. So when first time GT tripped we replaced BTJ32 Thermcouple and cables. But GT again tripped as it jumped to trip value but that time BTJ31 was showing normal value.

Could anyone explain me why it is happening as we know thermocouple if it is became defective than it goes to low value? and how to solve this problem? It will be a great help for me & it is urgent.

Mihir Ramkrishna

If I remember correctly, the reason why the reading is showing high in your case is, the system is programmed to do so. This means, in case of sensor breakage or damage the system has been so programmed that it will read high value and trip the machine. This is for the safety of the equipment.

You can check for the thermocouple input card which is reading these values. Try replacing and observing it too. If u have a data acquisiton system connected to the MKIV then u can get a trend of these values. If I remember correctly MKIV system does not have a DAS of it's own.

Since it is a package equipment, I do not recommend any major logic modifications that u can do.

Bruce Thompson

I am not familiar with the Gas Turbines you are speaking of, so my comments are directed toward thermocouples in general.

The first thing you may want to try is swapping the cabling between the two thermocouples and see if the problem follows the thermocouple.

If this is the case, look into the physical dimensions of the thermocouple well. You may be having close tolerances that allow the thermocouple to become grounded at elevated temperatures.
First, the open-circuit thermocouple failure mode in SpeedTronic turbine control panels is for the feedback to "go low", sometimes even negative; that's the open-circuit failure mode, not every failure mode. (The predominant failure mode of a T/C and its associated wiring is to fail in an open circuit condition.) If there are problems with interconnecting wiring and/or terminal boards, the feedback could drift/jump to a high value.

Second, it's not common for GE or any of its former Manufacturing Associates (John Brown, Thomassen, etc.) to trip heavy-dut gas turbines on high bearing metal temperatures. It's generally only an alarm, and any action (shutdown, unload, trip) is left up to the operator. In some cases, Customers and/or their insurance carriers required a unit trip on high bearing metal temperature but it wasn't very common (and your site may be one of those!).

Please confirm the exact cause of the trip by providing all alarm text message(s) annunciated when the unit is tripped AND the logic signal pointnames which cause the alarms to be annunciated.

Most thermocouples (other than exhaust- or wheelspace thermocouples) used by GE and it's former MAs were dual element T/Cs, and only one of the T/Cs (elements) was connected to the SpeedTronic turbine control panel; the second element was a "spare" and could be easily connected in place of the first in the event of a failure of the first element. (This was done for things such as bearings, etc., where the T/C couldn't be replaced very easily.)

Replacing bearing metal T/Cs usually involves removing the bearing shell and sometimes "chipping out" the dual element T/C since it's embedded (epoxied)in the bearing to better sense the temperature. Many of these embedded bearing metal T/Cs had a unique connector/plug to which T/C extension wire is connected and run out to a terminal board, where more T/C extension wire connects the signal to the SpeedTronic T/C input.

Was the embedded T/C replaced (after disassembling the unit and removing the bearing and then reassembling the unit), or was the plug/connector that attaches to the embedded T/C replaced?

Has the second element been connected to the SpeedTronic in place of the first element, and, if so, what have been the results?

If the T/C was actually replaced, then the interconnecting wiring needs to be carefully checked out. (The wiring cannot be meggared while connected to the SpeedTronic.) If only the T/C extension wire plug/connector was replaced, and the same element was reconnected to the SpeedTronic, then the T/C element is suspect and the second element should be connected to the SpeedTronic.

The type of T/C terminal boards typically used by GE for high-temperature locations (turbine- and load compartment) require that the wiring be terminated in a very specific manner. Usually there is a single screw for each T/C conductor with a small square plate under the screw head. The bare ends of the T/C extension wires are slid under the plate and the screw is tightened to hold them in place. THIS IS IMPORTANT--and difficult to describe without being able to draw. Every T/C and all T/C extension wire has two conductors of dissimilar metals. When terminating T/C extension wiring it is critical that one of two methods is used: either the conductors MUST be touching at the terminal, OR they must NOT be touching.

For example, a Type K T/C has a chromel and an alumel conductor. At any interconnecting TB, the chromel conductors MUST be in contact with each other _AND_ the alumel conductors must be in contact with each other. --OR-- The chromel conductors must NOT be in contact with each other _AND_ the alumel conductors must NOT be in contact with each other.

At any TB, if the chromel conductors are in contact with each other and the alumel conductors are not in contact with each other, then a "cold junction" is introduced into the circuit and in the presence of heat or cold in the junction box an erroneous millivoltage will be introduced into the circuit. Sometimes this only results in one or two or three degrees of error, sometimes it's much more. (The same thing can happen if the alumel conductors in contact with each other and the chromel conductors are not.)

The T/C TBs used by GE in high-temp locations are designed such that the chromel conductors are to be placed on either side of the screw under the square plate and the alumel conductors are also to be placed on either side of its screw under that screw's square plate. The conductors are NOT to be touching! In this way, a cold junction is introduced into BOTH circuits--and they cancel each other out. If the chromel conductors are both placed on one side of the screw touching each other under the square plate, and the alumel conductors are placed on opposite sides of its screw such they are not touching, then a single cold junction has been introduced and will cause a slightly erroneous reading, depending on the temperature in the junction box.

These same T/C TBs have been known to become brittle under high temperature and with age, and break causing the terminations to come into contact with the metal junction box housing/cover. Quite often, when this is discovered, an inadequate terminal board (i.e., one not designed for the heat of the location) is used as a replacement, and the practice of isolating T/C conductors is not possible or not observed.

Sometimes, a wire nut with a set-screw was used to make T/C terminations. It's important that if this is the method being used that the T/C conductors be twisted together tightly before being placed in the wire nut and the set-screw tightened. An insulating cover is then screwed on. Quite frequently, vibration causes the screw-on cover(s) to loosen and allow the metal connector to touch junction box metal. Heat over time can also cause
the screw-on covers to become brittle and break.

Many times cables are damaged during construction, and it takes several years for moisture to seep into the cable and cause problems. This can be very difficult to determine. Problems with cables in foundation-embedded conduits or underground conduits can be very difficult to troubleshoot and confirm (unless it's obvious that water is or has been in the conduits for some time--or, worse, that oil and water and diesel fuel or crude oil has also been in the conduits).

Check ALL the interconnecting junctions in the circuit between the bearing and the SpeedTronic to be sure they are tight, dry, and properly terminated.

If you suspect wiring problems, use a piece of proper T/C extension wire, properly terminated, and temporarily run it along the ground outside the unit to connect the T/C to the SpeedTronic and monitor the results. Usually, if GE provided the interconnecting cables, there are several spare T/C pairs in the cable.

Another common problem with older units occurs if they were built using cables provided with plug connectors at one or both ends. Moisture CAN and DOES get into these connectors and causes all kinds of UNUSUAL problems. It's happened many times, that a "Pyle" plug or "J-cable" plug as they are commonly known looks reasonably good from the outside, but when disconnected it is found to be FULL of moisture and white powder and corrosion from repeated wettings and dryings. If the unit has these cables at either end of the interconnecting T/C cabling, be sure to check them carefully (the unit will have to be shut down when disconnecting any one of these cables...).

Are any Diagnostic Alarms related to T/Cs being annunciated when the "trip" occurs, or prior to, or afterwards? Please report back any and all Diagnostic Alarms being annunciated presently. There's always the possibility that there is a problem with the T/C input card in the SpeedTronic, or with the analog card processing that particular input, but there are usually Diagnostic Alarms associated with card problems (not always--but usually). There are test cables/procedures for performing diagnostic tests of Mk IV hardware; see the Mk IV Maintenance Manual.

One last thing to consider: shield grounding. Over-all cable shields and individual shields should only be grounded in one place--and that's usually the SpeedTronic turbine control panel. Grounding shields at both ends (which happens all too frequently) can introduce ground loops which can cause electrical noise problems.

Many installations outside of North America required armor-shielded cables, and these armor shields were grounded at both ends. The armor shield is something of an overall cable shield, and grounding it at both ends can lead to--and has led to--electrical noise problems if not done properly.

Hope this helps! There are so many possible configurations and so many possible modications that it's difficult to cover them all in a brief write-up (and this isn't brief!). Add to that any possible modifications made over the years by Customers and service personnel, and the possibilities become just about endless.

The basic concepts will still always apply--do not introduce cold junctions in the interconnecting T/C cabling/wiring. Make sure all interconnections (T/Bs, plug connections, wire-nut connections, etc.) are all tight and dry. Make sure all shields (twisted-pair shields, triad shields, over-all cable shields) are properly grounded.

And, _most_ SpeedTronic panel problems are annunciated with Diagnostic Alarms, or, in the case of the Mk IV, can be troubleshot with cables/procedures in the Maint. Manual.