Mark IV GE Gas Turbine Generator Reverse Power


Thread Starter



I 'am very please if anyone could share same experience or advice related to the problem in Mark IV gas turbine generator control. We have Frame5 P for generator drive control by Mark IV speedtronic. Recently we face generator breaker open after few hours/day of unit running by reverse power. Our investigation to the problem found that this reverse power due to 2 co incident of problem related FSR.

1. Computer "T" problem, indicated by Red Activity LED was Off after reaching FSNL & FSRACC freeze at 16%

2. FSRACC in Computer "S" (Computer "R" was OK) ramp down to 17-19% /below Full speed no load FSR (FSR FSNL 20%)& recover above FSNL FSR

Related to the item #1 finding we already change HMPF card in Computer T, HCMA card in comp "C", try to disconnect card except HMPF, HXPD & HAIC & even replaced all card in comp T, but problem still remain

Related to the item #2 finding, we suspect this is related to the HMPF card due to some time FSR ACC in computer "S" only that drop below FSNL FSR. but problem still remain after replacing this card.

many thanks for any feedback.

Reverse power means the fuel has been reduced to less than what's required to keep the turbine-generator spinning at synchronous speed--and power from the grid is coming in to the generator (real power--watts, kW, MW) to keep it and the turbine spinning at synchronous speed. That's not a good condition for the grid, but it really doesn't hurt a gas turbine-generator very much (presuming there's still flame in the machine and excitation on the generator) as long as it doesn't continue for long.

You didn't tell us what fuel the unit was burning when the reverse power trips occurred.

When a Mark IV loses a processor the servo-valve outputs from that processor are lost. That means the other two processors are the only ones still providing power to the servo-valves. (I believe the servo-valve outputs are shorted when the processor is not functioning/powered off.) This means the servo-valve output currents from the two remaining processors are summed in the servo-valve and the resulting amp-turns determine the position of the device being controlled.

You can look at the servo-valve outputs when all three processors are running to see if there are any large mismatches--and if there are, then you need to work backwards through the sequencing/logic to understand why. Because if there are large mismatches during normal operation, then when a single processor is lost there will be issues--such as reverse power.

Things to look for include LVDT feedback mismatches (difficult to spot in a Mark IV, or a Mark V--but not impossible). The Mark IV uses one LVDT of a dual LVDT device for one processor, the second LVDT for a second processor, and the high-selected value of the two for the third processor. (See the servo-valve output sheets of the Mark IV Speedtronic elementary for details.) Also, because there are only two LVDTs on most devices (some older units only had a single LVDT on the SRV (Stop-Ratio Valve)), the power for the LVDT comes from one processor, and the power for the other LVDT comes from a second processor. There is no redundant power supplies for LVDTs (or RTDs, if so equipped). So, if one processor is lost and it's powering one (or more) LVDTs and the remaining LVDT feedback is not close to the other's when it's operating that can cause a problem which can result in low fuel flow-rate (if we're talking about gas fuel).

Also, one should always consult the Diagnostic Alarms when troubleshooting problems like this. If there are Voting Mismatch Diagnostic Alarms when all three processors are running, predominantly on one processor, and another processor is lost then those Voting Mismatches can cause problems while running on the remaining two processors. So, check and resolve any existing Diagnostic Alarms, because while they may not cause problems when all the processors are happy and healthy--they can cause problems when a processor is lost.

There's this big belief that when a TMR Mark IV loses one processor, one of the two remaining processors "takes over"--that's only true for logic signals (where possible). The servo-valve outputs of the remaining two processors are not switched to the "designated voter" processor when one is lost. So, if they are mismatched that can cause problems.

I also believe that if you're having problems with FSRACC that the turbine shaft speed feedback for one processor might be intermittent or low or failing. But, this should be annunciated as a Diagnostic Alarm.

If the unit is running on liquid fuel, there may be an issue with one of the Liquid Fuel Flow Divider speed feedback signals. But, this, too, should be annunciated as a Diagnostic Alarm. Again, Diagnostic Alarms may not seem like they are important, especially if the unit is seemingly running well--but they indicate the potential for real problems if one processor is lost.

Hope this helps! Please write back to let us know how you progress in resolving the issue.