Trouble with generator RCD

[panelman]

Brain picking time (with apologies for length)…….

We have a big three phase generator (BG) & a little single phase one (LG). During the day BG runs and supplies a three phase load and a critical single phase load. At night LG runs and supplies the single phase load only.

Both gens have a 30mA RCD on the output and are connected to a panel which starts the BG if LG isn’t running. Workers stop LG when they arrive (panels starts BG) and start LG when they leave (panel stops BG). If LG stops in the night then panel starts BG.

Panel gets power from both gens and, via contactors, energises a three phase and single phase output if BG is running and single phase only if LG is running. It has two control transformers (240/110), one fed from each gen, with the 0V of the secondaries tied to earth. This is the only connection between the two systems, contactors have timers to give a 20s dwell between the two single phase contactors at changeover time.

Problem we have is that when BG stops its RCD is tripping. It does it maybe 60% of the time and only if LG is running i.e. if we fiddle it so the panel only thinks LG is running then the rcd stays in.

Both gens claim to have star/neutral connected to earth

Any thoughts welcome

 

[ScottyUK]

There should only ever be one, and only one, connection to earth on the system otherwise neutral current will flow in the earth conductors causing misoperation of the protection. One option is to connect both neutrals to a neutral bus and earth the neutral bus at one point.

I'd be tempted to use a larger time-delayed RCD for the generator E/F protection and provide individual RCDs or RCBOs for personnel protection on the circuits requiring it. You can still meet maximum circuit disconnection time requirements using, say, a 300mA / 100ms delayed RCD but that would not be suitable for personnel protection.

----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[panelman]

Cheers Scotty, I was leaning towards the time delayed rcd myself but without having a firm handle on what is actually happening.

To clarify, the sequence is….

BG running and supplying both loads
LG is manually started and powers up its control transformer
Timer fed from this drops out both BG contactors
NC contact on BG contactor powers another timer which brings in LG contactor
NC contact on LG contactor opens the run circuit for BG which then stops.
More often than not as BG stops the rcd trips

 

[waross]

I suspect that you have an unexpected connection between neutral and ground somewhere. Part of the neutral current from one or more of the single phase loads is flowing on a ground conductor. This connection may be deliberate (possibly the control transformers, hard to say without being there) or it may be an accidental connection between a neutral and earth.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[panelman]

Bill, thanks for your thoughts but there is more to it

The control transformer link to earth is on the 0V of the secondary side, no load (apart from the transformer) is connected to BG when the trip happens and the trip only happens when BG is stopping and LG is supplying the single phase load.

 

[waross]

Have you put a meter on any conductors (ground or neutral) to look for a current spike?

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[ScottyUK]

The genrator neutrals are almost certainly bonded to the generator frame earth and connected to the earth electrode. Do your contactors switch the neutrals or are the neutrals permanently connected together in the panel / switchboard? Look for possible path where a neutral conductor is paralleled through the generator earthing conductors and possibly the mass of earth. There are two good reasons not to allow this: it's a violation of the wiring regs because neutral current passes through a protective conductor, and it leads to the symptoms you're seeing.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[panelman]

Yep, contactors switch the neutrals so they are kept isolated from each other at all times

 

[waross]

Look for a single phase contactor or circuit that is NOT switching the neutral.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[panelman]

Bill

There are none, both generator neutrals come direct to the panel, are used for the control transformers (both completely independent apart from the 0V of both secondaries being tied to earth) and then go to their respective contactors where they are switched (only one in at a time and with 20s dwell on changeover) and leave the panel to feed the critical load.

Re your earlier comment about spikes…..I haven’t investigated this yet as site is 200 miles away (a long way for us in the UK). I have see spikes before on generators as they “cough” to a halt but don’t see how this would bring out the rcd given that the only thing connected to the gen when it happens is the control transformer

 

[waross]

When you get on-site it may be productive to temporarily lift the transformer secondaries from earth to see if this is the source of the problem.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[ScottyUK]

30mA instantaneous RCDs are pretty sensitive to nuisance tripping on energisation with loads such as switch mode power supplies. If you definitely rule out parallel neutral paths through earth then I suggest a higher E/F protection setting and a short delay, with supplemetary protection for personnel as required. There's rarely any need to have such sensitive protection at the origin of the supply. Somthing like Merlin Gerin's Vigi relay with a CT on the neutral-earth bond is one of many possible solutions.


----------------------------------

If we learn from our mistakes I'm getting a great education!