Mixed Fuel Mode Operation

Hello,

We have several Frame 9E units. Sometimes we're tripping at Gas-to-Liquid changeover, which we perform once per week or two. Usually, the problem is fuel pipes and other volumes not completely filled with liquid fuel. By slowing rate of changeover from 30 sec to 5-10 times more (reducing FXK1 constant), we always succeeding. But, when we (I&T) not at site and operators not allowed to make a simulation, they are in trouble.

Is there a limitation to prevent a hardware damage for working in Mixed Fuel Mode for not 30 sec but, let's say, 3 min? Can we implement this rate constantly?

Thanks.
 
Hello,

This is a very common problem for dual fuel units (gas/distillate) which operate primarily but need to be able to transfer fuels quickly. The OEM and several other companies can provide liquid fuel readiness systems, some of which can be VERY expensive, but which automatically run the liquid fuel forwarding pumps regularly to help keep the liquid fuel lines as air-free as possible, and some even have recirculation valves and orifices to make sure there is flow through the piping to help make sure there is no air accumulating in the piping. Again, these can be very expensive, and some are very effective--but at a steep cost and increase in complexity, I/O and logic/sequencing (not to mention maintenance).

Extending the time for an automatic transfer to take place can help the testing, but what will that do if the requirement to automatically transfer fuels occurs a day or two or a few hours before the test is to be performed?

Finally, when transferring fuels it's NOT advisable to operate for "excessive" periods of time when there is little liquid fuel flow as this can cause combustion trouble and combustion hardware problems.

One of the best things which can be done to improve liquid fuel system reliability is to use pneumatic valves instead of spring-loaded, poppet-style liquid fuel check valves. The usual check valves provided with many GE-design heavy duty gas turbines have been proven to leak in the reverse direction, and this can also lead to opening of the liquid fuel stop valve as well as carbonization of liquid fuel in the fuel nozzles of the combustors with leaking check valves as well as the tubing lines from the liquid fuel flow divider (because it's hot combustion gases that are entering the the tubing lines, and making their way into the liquid fuel supply system when the liquid fuel stop valve is opened by the pressure operating on the liquid fuel stop valve plug).

Using these pneumatic valves increases the complexity of tubing in the turbine compartment (always a not-so-good thing), but it's very effective.

The spring-loaded, poppet-style liquid fuel purge check valves have been proven to leak in the reverse direction when operating on liquid fuel and cause combustion troubles (high exhaust temperature spreads).

These are the two biggest causes of "air" in the liquid fuel system--these and improperly run liquid fuel supply piping. The liquid fuel forwarding stop valve can also sometimes leak in the reverse direction which gives the liquid fuel being pushed backwards in the liquid fuel supply system a place to go.

There's no easy answer to this problem--but one of the best mitigation solutions is to replace the spring-loaded, poppet-style check valves with pneumatic solenoid valves (that would be 14 such valves for a GE-design Frame 9E unit...).

Again, it's not really desirable to operate in mixed fuel mode when transferring fuels (or operating on MIX fuel mode) for any real period of time when the percentage of liquid fuel is very low (as seen on the operating display).

It doesn't sound like someone wants the operators to be able to change Control Constants or force logic. So, that means working with a knowledgeable person or firm to write some logic/sequencing and modify the Mark* and put some buttons on the liquid fuel display to automatically run some procedure to do the tests so the I&T (what does the "T" stand for) personnel don't have to be on site to run the liquid fuel transfer testing.

Either that or change the liquid fuel check valves, at a minimum, to something with a more positive seal when operating on gas fuel.

Hope this helps!
 
Thank you for your reply, CSA.

"I&T" is a typo - I meant I&C (Instrument and Control Maintenance).

We've considered changing check valves with pneumatic valves but this solution was dropped exactly because of reasons you've mentioned - cost, additional hardware in already "clogged" area in turbine compartment, a lack of free I/O in MKV TBs and even lack of spare logic points in the memory.

>>Finally, when transferring fuels it's NOT advisable to
>operate for "excessive" periods of time when there is little
>liquid fuel flow as this can cause combustion trouble and
>combustion hardware problems.

We never got an answer from GE what is considered "excessive period of time". But when we run a slow transfer, 10 times of original, CTXF and TTXD readings are very stable and no trouble visible. At the end, a decision to go for this solution is a matter of responsibility which no one on site and seemingly not GE either wants to take.

Thanks again, it's VERY important to hear what experienced people around the world think. It gives a feeling of "enhancement", kind of Mr. Smith in "Matrix" movie..:)
 
jolek,

It sounds like you lucky and have a unit that was packaged by GE-Belfort which has a lot of overly complicated sequencing and a lot of I/O.

I have seen a single solenoid-operated valve to control the 14 pneumatic valves, but that adds a lot of tubing to the compartment. And if you have a unit with on-base enclosure that would be very tight (I presume you also have Water Injection). Even with off-base lagging it can still get very crowded--IF a little planning and forethought are employed.

As for the transfer time length, in NY personal experience and knowledge of having worked through some issues with a pair Frame 9E units many years ago 300 seconds is borderline too long. And that was with conventional combustors with a single nozzle per combustor. DLN units have multiple nozzles per combustor which means the liquid fuel flow-rates during the initial part of the gas-to-liquid transfer is very low....

But if you say there are no spreads--and no load swings--then it sounds as if everything is good. The real proof will be seen during a maintenance outage; if the fuel nozzles and liners are okay, then that's good.

Little or no MW swings, low exhaust temperature spreads and no signs of issues with fuel nozzles (canonization in the nozzle or on the nozzle tips), and no signs of hot spots or discoloration of the combustion liners directly below the nozzles are the desired effects and indication that things are okay.
 
jolek,

WOW!!! My apologies, sir. I am not a good proof-reader of my own writing, and with a new mobile phone the spellchecker is overly ambitious with technical writing corrections--which I didn't catch. So, the message must have seemed very strange.

So, here goes a re-write, with (I hope!) most things corrected. Again, my sincere apologies for any confusion.

It sounds like you are lucky and have a unit that was packaged
by GE-Belfort which has a lot of overly complicated
sequencing and a lot of I/O.

I have seen a single solenoid-operated valve used to control the
14 pneumatic valves, but that adds a lot of tubing to the
compartment. And if you have a unit with an on-base enclosure
that would be very tight (I presume you also have Water
Injection. Even with off-base lagging it can still get very
crowded--IF a little planning and forethought are NOT employed.

As for the transfer time length, in MY personal experience
and knowledge of having worked through some issues with a
pair of Frame 9E units many years ago 300 seconds is borderline
too long. And that was with conventional combustors with a
single nozzle per combustor. DLN units have multiple nozzles
per combustor which means the liquid fuel flow-rates during
the initial part of the gas-to-liquid transfer are ery
low....

But if you say there are no spreads--and no load
swings--then it sounds as if everything is good. The real
proof will be seen during a maintenance outage; if the fuel
nozzles and liners are okay, then that's good.

Little or no MW swings, low exhaust temperature spreads and
no signs of issues with fuel nozzles (carbonization in the
nozzle or on the nozzle tips), and no signs of hot spots or
discoloration of the combustion liners directly below the
nozzles are the desired effects and indication that things
are okay.

-----

Hope that clears things up a little.
 
Thank you, CSA!

Those small grammar mistakes don't bother me at all. Myself, as English is not my first (and not a second, and not a third) language, surely making such terrible mistakes that I'm really glad someone can get to the point of a problem I'm trying to share and describe :). Misunderstanding a technical issue and incorrect advice and explanation as a result are bother me.

Thank you for rereading and rechecking your own replies.

BTW, we don't have Belfort packaging at these units (we do at other two with MKVIe), but upgrade to dual fuel and DLN 1 made by Thomassen.
 
Top