Startup Acceleration Too Low

M

Thread Starter

Maythem

Hello all,

we are operating simple cycle frame 9E and i realized since a while now that the following alarm appeared in alarm screen without affecting anything in the routine operation and loading "start up acceleration too low" and according to that i have several questions as follow:

1- what does this signal means in terms of turbine control (how it is generated?), what are the possible causes of such signal to appear?

2- does that signal have any effect on turbine life cycle? is it kinda a signal that needs a great deal of attention?

3- what are the possible solutions to prevent this signal from appearing?

best regards,
Maythem
 
Maythem,

Good questions. Since GE in Belfort, France, took over the responsibility for the Frame 9E (and Frame 6B and Frame 9F) units a lot of what was once standard sequencing has changed. This is not an alarm I'm familiar with and without being able to see the sequencing/application code running in the Speedtronic turbine control panel I can only make some educated guesses. Hopefully one of our posters with more recent experience with Frame 9E units will add to the discussion; we have several who have dived very deeply into Frame 9E control philosophy.

I can also suggest you locate a document called "Control Specifications" and look at Sect. 03, which is usually the Speed Control section, for more possible details about what's actually contained in the Speedtronic turbine control system at your site.

1) Most GE-design heavy duty gas turbines controlled by Mark V and Mark VI and Mark VIe control systems use a varying acceleration reference during start-up to accelerate the turbine from warm-up to FSNL. The name of the acceleration reference is usually TNHAR (Turbine Speed-High Pressure Shaft Acceleration Reference), and it's compared to the actual turbine acceleration rate, TNHA, and fuel is adjusted to make the actual equal to the reference. During staring of a GE-design Frame 9E heavy duty gas turbine the unit requires a torque assist from the starting means until the unit reaches approximately 60% of rated speed. The reason is that if fuel is used solely to accelerate the turbine it will require a LOT of fuel meaning the internal hot gas path temperatures--as well as the exhaust temperatures--will be excessively high, so a torque assist from the starting means is used. (It's also required at low speeds during starting, firing and warm-up when there's not enough heat/pressure from the combustion gases to sustain rotation.) So, possible causes would likely be an inadequate torque assist from the starting means and/or an inability to increase fuel fast enough to make the actual acceleration rate equal to the acceleration reference. I suppose under some conditions a dirty axial compressor and/or a high exhaust duct back pressure might also contribute to slower-than-normal acceleration.

Typical Speedtronic turbine control logic also has an upper limit to fuel control during starting and acceleration. And it's not uncommon for this value to be set too low during commissioning or for changes in fuel make-up (BTU content; LHV/UHV; Wobbe index) to also cause the Speedtronic to have to increase fuel flow-rate until it reaches the acceleration limit--so fuel make-up can also be a possible cause of this problem, in conjunction with a dodgey acceleration rate limit value. One should not just change Control Constants without understanding why they need to be changed--and any knock-on effects of changing them. So, we'll table this one for now. But, if the fuel changes from time to time at your site this could be a possible cause--or one of several possible causes.

Without being able to see the sequencing/application code running in the Speedtronic at your site, it's impossible to say exactly how the difference is being calculated to generate the alarm. Knowing the Belfort Bunch, it's unnecessarily complicated (Belfort's philosophy seems to be why make something simple when it can be very involved and unnecessarily complicated), but it could also be as simple as the unit is not achieving some speed setpoint in a certain amount of time after a START or flame is detected, or something like that. Again, without being able to see what's running in the Speedtronic, it's impossible to say for sure how your machine is operating and determining if the acceleration is low.

2) Well, if the Speedtronic is having to put extra fuel just to try to make the unit accelerate at the desired rate, and this is resulting in abnormally high internal- and exhaust temperatures, then, yes, it is adversely affecting parts life.

There is also something called "bog down" detection which looks at the speed and if it decreases during the period when it should be increasing during starting then the turbine is tripped. This is to protect against firing a machine which isn't accelerating but should be.

As far as how critical this "slow acceleration" alarm is, well, it could be just trying to alert an operator to impending problem, that not dealt with, could eventually result in a problem that could trip the turbine during starting in addition to needlessly increasing machine temperatures during starting (when there's not much air flow because the axial compressor speed is low, the compressor bleed valves are open, and the IGVs are closed).

3) The possible solutions are dependent on the causes. Right? One needs to do some troubleshooting to understand the cause, and then one can solve the problem.

Hope this helps!
 
Dear CSA,

thank you for taking time to write me the answer. i will dig up to bring more details. i am thinking of put some related signals on a trend and then try to analyse it. i am kindly requesting to advise me which signals should i register on trend to get the most of it.
 
Maythem,

You want to trend:

TNH (or TNH1)
TNH_RPM
THNA
TNHAR
FSR
FSRSU
FSRACC
FSRMIN
FSRN
TTRX
TTXM
CSGV
CSRGV

But, you also need to have someone look at the application code to see how this low acceleration is being determined--is it that actual acceleration is less than reference, or is it that it's taking longer than some time setpoint to reach a certain speed, or ???

Also, if you know that the gas fuel is coming from a different supply/source, and you can get a copy of the fuel analysis from the supplier to compare against the previous fuel's analysis and you discover that the heat content of the new fuel is lower than previously then that suggests another possible cause or contributory factor. BUT, you still need to know how this alarm is being calculated or determined--which is key to any analysis.

Hope this helps!
 
Maythem,

Since my main project is a Frame 9E, I have some experience with them and the Belfort design. I also have MarkVI software from various sites with 9E. The problem is that I don't see in any of these software an alarm similar to yours.

Can you please tell us the name of the signal that trigger this alarm? I could try to identify in my software where your alarm is generated based on it. Of course, since the guys from Belfort are very "creative" lately, the conditions for raising an alarm may vary from unit to unit.

Even better, if you know how to identify in the software the sequence where the signal is generated, make a screenshot of the sequence, upload it somewhere, and include the link in your reply.
 
Top