Dear All

I have observed a very peculiar behavior of the allowable exhaust spread limit in the GE Mark-V of our 25MW(MS5001) Gas Turbine.

Generally the values of some variable are as under :
TTXM=403, CTDA=289 deg C
We have constants as under :
TTKSPL4=0.12,TTKSPL3=0.08,TTKSPL5=26.11 Deg C.

So TTXSPL(Allowable Spread) is calculated to be 51.59Deg C or Appx. 52DegC (Say).

The value of TTXSPL in Mark-IV is showing around 52 DegC generally. But sometimes we see a peak in the trend of TTXSPL up to 134/135 Deg C. The value is hold up there for about 1 Min and comes back to 52 Deg C gradually. This peak repeats irregularly i.e. don’t have a certain trend. (Comes 5/6 times in 8 Hrs)

Can anybody help me out understanding the phenomena please.

Rameswar

 

The combustion monitor software has built into it a facility which will increase TTXSPL by +200 Deg.F every time there is a system disturbance. Most frequently this disturbance is an adjustment to the turbine governor set-point. As soon as the disturbance is finished, TTXSPL will maintain it's set-point at normal + 200 Deg.F for two minutes then ramp the value slowly back to it's original setting. This is what you are seeing.

If you would like a copy of a GE document which completely explains this, send me your Email to: bbobjohnston [at] gmail.com

 

This is absolutely not abnormal behaviour. Trend L70L and L70R in addition to the other signals you have been trending and you will find that whenever the allowable spread is bumped up that it occurs whenever L70L or L70R transitions to a logic "1".

This is done to bias or "bump up" the allowable exhaust temperature spread out of the way during a load change. Anything that causes L70R or L70L to go to a logic "1" will work to bump the allowable spread up to maximum, and then after a time delay (usually 60 seconds or so), the allowable will start ramping down to the value which is a function of the allowable exhaust temperature spread.

Have a look at L2SPMBx and L2SPMBn and L83SPMB (where 'x' is an alphabetic character and 'n' is a digit). Also follow the signals TTKSPL8 and TTKSPL9 in the CSP and you can see the bias and how it's added and subtracted from the calculated spread bias.

I'll bet the unit is operating on Base Load, exhaust temperature control and when there is a change in CPD or the exhaust temp then a RAISE or LOWER load signal (L70R or L70L) is initiated. A RAISE or a LOWER is considered to be a load change, whether initiated manually (by an operator) or automatically (by exhaust temperature control or load control).

 

Abnormal Behavior in Gas turbine

well as CSA and Bob have said there is nothing abnormal here. the equation for calculation of TTXSPL or the allowable spread is given by the formula

TTXSPL = (TTXM*TTXSP4 + TTXSPL5) - (CTD*TTXSPL2) + TTXSPL8(1 - e^(-TTXSPL9*t))

Where t is the time from which L83SPMB changes state from 1 to 0

From the control constants in the application code the formula can be rewritten as

TTXSPL(deg F) = (0.12 * TTXM) + 47 - (0.08*CTD) + 150 ( 1 - e^ -100t)

from the above you can see that the term 150(1 - e^ -100t) is a exponentially decreasing term. when L83SPMB goes to 1 , the 150 gets added to the TTXSPL value. this stays on till L83spmb is 1. so when L83SPMB is 1 the formula is


TTXSPL(deg F) = (0.12 * TTXM) + 47 - (0.08*CTD) + 150

L83SPMB goes high when there is a change in the fuel input to the turbine. ie if the speed/fuel raise signals L70L and L70R remains high for 2 sec L83 SPMB goes high. then after the L70L or the L70R loes low a timer runs for 5 sec before L83SPMB goes back to zero. when it goes back to zero , the equation becomes

TTXSPL(deg F) = (0.12 * TTXM) + 47 - (0.08*CTD) + 150 ( 1 - e^ -100t)

thus the value of ( 150 (1-e^-100t)) exponentially decreases thus in effect reducing the TTXSPL value. if you trend the signals TTXSPL and L83SMPB , just force L83SPMB then unforce it , you can get a clear pic of the exponentially decreasing TTXSPL value.

why this is done ??

Whenever there is a fuel increase / decrease to the turbine , the combustion pattern changes , the control system allows a time period for the machine to settle down in steady state , this is the reason for increasing the allowable spread and after the fuel change is over it is reduced back exponentially to its actual value . thus in the formula the TTSXSPL without the exponential part is the min allowable spread value.

 

This is an old thread, I know... But if anyone is watching I have a followup question:

What percentage of time is it appropriate for this "normal" behavior to exist? I have noticed that our allowable temperature spends what seems like quite a lot of time at the biased level (5 out of 30 minutes at the peak with the rampdown taking the remainder of the time).

We have a 61B with Mark VI controls (HRSG on the back end). Steam injection for NOx control. We operate in preselect mode frequently to prevent us from exporting power.

May be a dumb question, but is the smart logic putting us at greater risk by trying to prevent spurious trips?

Thank you.
Ricky