Would anyone happen to know how the WI reference flow is calculated? We are getting a shut down trying to get wi in service. Actual flow is 3-10lbms and reference is 34. Last time we ran with lf was 3 yrs ago and the reference was at 14. Controls upgraded to mk6e since then. Alarm says wi purge very low psi. Vfd ramps up and looks to put to much water in and then we get a dln high ex spread. Any suggestions or know what constants can change that calculated reference? Thanks.

 

The water injection flow-rate reference is usually based on the fuel flow-rate. It is usually biased by a humidity signal to account for water entrained in the inlet air (from ambient and possibly evaporative cooler or other inlet air cooling means). But if the unit turbine control system was upgraded then one of the first places to look is the Control Constant values for all water injection-related parameters—including the humidity sensor bias. (Historically, water injection Control Constants almost all began with WQKR. Many were arrays of values.)

Also, make sure the humidity sensor is working properly and providing a true and proper signal to the water injection flow-rate reference calculation.

Another often overlooked parameter which can greatly affect the water injection flow-rate reference is the fuel flow-rate feedback from the liquid fuel flow divider. If you have good historical data from before the upgrade you should be able to see if the new control system is calculating the liquid fuel flow-rate to be nearly equal to before the upgrade for a particular load. OFTEN the scaling for the liquid fuel flow-rate feedback is not set correctly for the new control system.

But, without being able to see the application code running in the Mark* VIe it’s near impossible to tell you precisely what signals and Control Constants to check. You need to use ToolboxST to look at the water injection calculations and make sure the inputs to the calculations are correct (liquid fuel flow-rate; humidity; ambient temperature; etc.) and the Control Constants are the same as before the upgrade.

If my memory serves me correctly, the water injection flow-rate reference should be somewhere in the range of 1.1 to 1.3 times the liquid fuel flow-rate, and the Mark* usually calculates the water-to-fuel injection ratio (both the reference and the actual, since a small amount of extra water is injected to ensure the required NOx emission level is slightly below the legal limit). If you have historical data from before the upgrade you should be able to find the ratios to compare to now.

[Usually testing is done during commissioning to calculate the required water injection flow-rates to meet the actual NOx emissions guarantee, and then a small amount of additional water (tenths of a lb/sec) is added to that flow-rate to (hopefully) ensure the NOx emissions are slightly below guarantee so no one goes to jail. So, there’s a required injection flow-rate (water-to-fuel ratio) and an actual water injection flow-rate (and water-to-fuel ratio). But all of this is based on liquid fuel flow-rate and biases for axial compressor inlet temperature and humidity. The inlet temperature inputs should be working fine; but the humidity feedback/input may not be OR the Control Constants might not have been properly checked.]

Again—without being able to see the application code running in the Mark* it’s just not possible to say for sure how the reference is actually being calculated for the machine at your site. And some machines even have inputs from the emissions monitoring equipment to bias the water injection flow-rate reference to a very fine point so as to keep water consumption to a minimum (treating water for NOx emissions reduction is expensive, as is buying the water and disposal of any residual chemicals!).

Please write back to let us know how what you find. Troubleshooting water injection is a lost art. If you’re having purge pressure trips you need to make sure all those transmitters used to calculate purge pressures are properly valved-in (isolation valves opened) and are working properly. Valve line-up, as an old colleague used to say, THE most important aspect of operational preparation and one of the most over-looked, too (making sure all those little stainless steel manual valves are in the right positions is ALWAYS somebody ELSE’S job….).

Happy Holidays!!!! Whoever the jerk was that scheduled this work and upgrade for this time of year needs a couple good slaps up ‘side the head with a blackjack. (Not too hard; but his (or her) ears should ring.) People have families and the weather probably blows where the site is right now and lots of those people who flew or drove to the site may not be able to get home. Great planning—NOT. Capitalism at its … finest.

 

I think you're on the right track as far as determining why the reference is so high--but, again, the reference is typically based on liquid fuel flow-rate. It's usually a series of straight lines which can look like a "curve" over the expected liquid fuel flow-rates.

But, I think you've got something wrong with metering/sensing if purge pressure is tripping you--or something is amiss with the application code. Because unless you have multiple water injection flow lines (a low- and a high-flow system, for example--which is entirely possible) purge pressure shouldn't be a factor when water injection is running. (What I'm trying to say is if the unit is tripping on low purge pressure when water injection is or should be running, that doesn't seem to make much sense, does it--UNLESS the low purge pressure trip is occurring on the high flow system when only the low flow system is in service, in which case it sure seems like something is amiss with valving and/or the application code (logic; sequencing).)

The high exhaust spreads will occur if the water injection flow-rate starts to extinguish the diffusion flame in one or more combustors because the water injection flow-rate is too high to sustain flame. As with DLN, water injection can be a tricky balancing act (trying to get it to come on "quickly" (presuming the reference is correct!) and trying to prevent blowing out the flame while also reducing NOx emissions).

 

Problem resolved. After a couple dozen trips and shutdowns someone located an array block that had points in the wrong spot. After that correction was made we ran up to 160 and WI coupling came apart. Everything seems to be good. Thanks for the advice. And After are upgrade the scaling was off on lf but we did find that when we tried to start up on fg. It tripped us a few times also but not like this issue.

 

Happy Holidays!!!! Whoever the jerk was that scheduled this work and upgrade for this time of year needs a couple good slaps up ‘side the head with a blackjack. (Not too hard; but his (or her) ears should ring.) People have families and the weather probably blows where the site is right now and lots of those people who flew or drove to the site may not be able to get home. Great planning—NOT. Capitalism at its … finest.
[/QUOTE]
Work start in Sept. But I agree never schedule shit during hunting season. The other unit had to come back apart cause of vibs on t2.

 

"After a couple dozen trips and shutdowns ...."

Gotta love it--just keep pushing the START button until it finally starts. Works (almost) every time.

NOT!

People have NO IDEA how much effort went into the Trip Reduction Program for the F-class machines. They just don't like to be tripped. But, whatever gets one to the Business Class Lounge, eh? Guess it's okay to keep tripping the machine in an effort to find the elusive problem (cause it's ALWAYS that damn control system's fault!); at least it looks like someone is doing something, right?