If four or five exhaust thermocouples went faulty in gas turbine having twenty four exhaust thermocouples, what will happen?

 

Not good at any time to have so many Thermocouples faulty - why?? Regarding problems, depends on their location but could give you trips in the Combustion Monitor software (Exhaust Temp. Spread).

Best suggestion is replace the thermocouples.

 

A few years ago I had access to a Mark VIe simulator, and I was able to determine that when a third exhaust T/C failed (low!) that a trip was initiated. (And it should be said that the combustion monitor which uses the exhaust T/Cs is virtually the same for almost every version of digital Mark* turbine control system.)

Exhaust T/Cs serve two purposes. First, the average exhaust T/C value is calculated (after rejecting outliers) and used to limit the exhaust temperature to protect the turbine and exhaust components against exhaust overtemperature.

Second, the exhaust T/Cs are used to detect a problem with the combustion system and/or the transition pieces (that connect the combustion liners to the first stage turbine nozzle segments). Combustion problems can lead to high flows through cross-fire tubes which can cause serious problems with liner failures which can lead to metal pieces going into the turbine section. Not good; not good at all--just like glenmorangie says.

This is called 'combustion monitoring' and the way that combustion problems are detected is by looking at the temperature differentials between individual exhaust T/Cs AND whether or not some of the highest differentials are adjacent (next to) each other. In this purpose, it's not a good idea to reject outliers--because they could be indicators of serious problems. Failed exhaust T/Cs degrade the ability of the control system to protect against serious problems which could be caused by combustion problems. And, combustion problems are just as serious, if not more so, than exhaust overtemperature conditions.

Whether a turbine has 13 or 18 or 24 or 27 exhaust T/Cs doesn't mean that any particular number of failed exhaust T/Cs can be tolerated or acceptable. Having more exhaust T/Cs is important for units with higher numbers of combustors and is helpful in detecting combustion problems before they cause more serious problems. But, having more exhaust T/Cs doesn't at all mean that more failed exhaust T/Cs can be tolerated and allow the unit to continue to run.

Unit protection relies on working inputs. If inputs are not working then the ability to protect the unit is degraded--regardless of the number of "redundant" inputs. They aren't just redundant when being used in the combustion monitoring scheme--the individual values are extremely important, and it should be evident that rejecting "failed" T/Cs in such a protection scheme is degrading the protection scheme.

 

A few years ago I had access to a Mark VIe simulator, and I was able to determine that when a third exhaust T/C failed (low!) that a trip was initiated. (And it should be said that the combustion monitor which uses the exhaust T/Cs is virtually the same for almost every version of digital Mark* turbine control system.)

Exhaust T/Cs serve two purposes. First, the average exhaust T/C value is calculated (after rejecting outliers) and used to limit the exhaust temperature to protect the turbine and exhaust components against exhaust overtemperature.

Second, the exhaust T/Cs are used to detect a problem with the combustion system and/or the transition pieces (that connect the combustion liners to the first stage turbine nozzle segments). Combustion problems can lead to high flows through cross-fire tubes which can cause serious problems with liner failures which can lead to metal pieces going into the turbine section. Not good; not good at all--just like glenmorangie says.

This is called 'combustion monitoring' and the way that combustion problems are detected is by looking at the temperature differentials between individual exhaust T/Cs AND whether or not some of the highest differentials are adjacent (next to) each other. In this purpose, it's not a good idea to reject outliers--because they could be indicators of serious problems. Failed exhaust T/Cs degrade the ability of the control system to protect against serious problems which could be caused by combustion problems. And, combustion problems are just as serious, if not more so, than exhaust overtemperature conditions.

Whether a turbine has 13 or 18 or 24 or 27 exhaust T/Cs doesn't mean that any particular number of failed exhaust T/Cs can be tolerated or acceptable. Having more exhaust T/Cs is important for units with higher numbers of combustors and is helpful in detecting combustion problems before they cause more serious problems. But, having more exhaust T/Cs doesn't at all mean that more failed exhaust T/Cs can be tolerated and allow the unit to continue to run.

Unit protection relies on working inputs. If inputs are not working then the ability to protect the unit is degraded--regardless of the number of "redundant" inputs. They aren't just redundant when being used in the combustion monitoring scheme--the individual values are extremely important, and it should be evident that rejecting "failed" T/Cs in such a protection scheme is degrading the protection scheme.

I've had that happen while operating twin 7FAs in a simple cycle. We changed out the T/Cs, and a few days later, it happened on the other turbine. As soon as the system sees 3 out of range "down She Goes"!

 

Such failures are commonly the result of the acceleration damage of the sensor as a result of flow induced vibration. Sensor replacement solves the immediate problem, but not the long term issue.

 

Such failures are commonly the result of the acceleration damage of the sensor as a result of flow induced vibration. Sensor replacement solves the immediate problem, but not the long term issue.

If your failure is due to tip failure of the exhaust gas thermocouple, Conax Technologies has a solution for that. See the link below and let me know if you have any questions.

https://www.conaxtechnologies.com/springloaded/

 

Yes, the spring loading does help boost the acceleration limit to 20+ G's, but does have its own limits. The key is quantifying the accelerations for the expected flow conditions, and sensor mounting.