We have frame 5 machine using in power station, having mark 5 control system. We using 4 flame detectors in combustion chamber, each using 315v dc supply and all are connected to ptba terminals 13&14, 15&16, 17&18, 19&20. We removed all the detectors for combustion chamber inspection and before re installation we checked all flame detectors with UV torch. It was working fine, but after running the machine one of the flame detectors is not working.

How can I check online whether this flame detector is healthy or not while machine running condition from my ptba terminal side? Also I want to know about the step by step procedure regarding this ptba board replacement ie, what are the precautions I have to take before replacing this card while machine shut down condition.is it required to power off <p> core before replacing this PTBA board? Can anybody help me?

 

SHYAM,

Geiger-Mueller flame detectors are polarity sensitive devices, meaning that the 315-335 VDC applied to them must be connected positive-to-positive and negative-to-negative. If they are connected backwards, nothing untoward will happen <b>UNTIL</b> the detector sees a high-intensity light ("flame"), and then they will fail.

It's not clear how the flame detectors were tested, nor where they were connected when they were tested. But, it's a safe bet that either the terminations of that one flame detector are not good (poor contact) somewhere along the circuit, OR that the wires were connected incorrectly (reverse polarity) when they were re-installed.

There is no way to test that flame detector without removing it at this point. The good thing is: There are redundant flame detectors so that in the event one (or two, in this case) fail that the turbine will continue to run until such time as the unit can be shut down and the trouble resolved. However, that does mean that if any other flame detectors fail in the interim (before the one that is currently failed or seems to have failed--can't tell which until the circuit and terminations can be checked) then running reliability is reduced, because any subsequent failures will likely result in a trip due to loss of adequate flame detection.

I presume that the exhaust temperature spreads are well within reason, and that's how you're certain the flame detector in question is not working. A very high exhaust temperature spread AND a seemingly failed flame detector could, in fact, be an indication of a combustion problem (poor flame; loss of flame in that combustor).

But, there is no SAFE way to check the flame detector while the turbine is running. Some (older) machines had isolation valves between the flame detector and the combustor, but that's only a single point of isolation and doesn't fit the requirements of most safety programs these days. Also, working on an energized circuit about 24 VDC is usually frowned upon these days, even though the current on these circuits is very limited (a shock will definitely get your attention, but the current is very low--but, still, caution should always be used when working with live circuits, whether or not the turbine is running).

Wish the news were better--but, still, with redundant flame detectors it's not that bad!

Please write back to let us know how you fare with resolving this problem.

 

We tested the detector using uv torch,our one of the terminal having problem (ptba).so kindly provide the procedure and prequations for replacing this ptba card.

 

shyam,

The turbine has to be shut down, off cooldown, and the Mark V has to be powered down--to be completely safe when replacing this card. Also, some Mark Vs had CTs (Current Transformers) connected to the <P> PTBA and if yours does, then you have to short the CT circuits to prevent a huge explosion.

Then, just disconnect the wires and cables, remove the card, install the new card, reconnect the cables and wires, and begin the process of powering-up the Mark V panel, one processor at a time.

Note that <C> will not go to A7 until at least one of <R>, <S> or <T> goes to A7. So, I always recommend powering-up <T> first, waiting until it gets to A6, then powering-up <Z>, and waiting until <T> gets to A7 before powering-up <C>, and then powering-up <S> and <Y>, and then <R> and <X>. And then applying the power to all of the discrete input and -output circuits.

Hope this helps!