3 Phase power keeps on tripping

[Daveh36]

Hi Guys any advice on this one will be a major help.
Was called out to a loss of power turned up expecting a breaker down as usual but not the case.
so here goes the problem- an apartment block with 315 amp incoming fuses feeding 400 amp mccb set to trip at 220. This in turn feeds a rising busbar with several tap off units and 250amp mccb feeding the apartment fuses and meters. No faults show when tested out but every night it trips taking out the 400amp mccb and all three main fuses.
Anyone any ideas on why this is happening. Please help as at my wits end with this one

 

[R-fur]

Is this a private network or is the DNO coming out to change the fuses? If the fuses belong to the DNO Iw ould expect them to start charging or refuse to replace them until the customer side issue is resolved.
As per UNG why wont the OP answer direct questions?

 

[marconi]

I have just read some posts in the iet forum about mice finding their way into trunking, busbars and switchgear and ‘setting up home’ with baby mice appearing in due course because the TV reception was lousy.

It might be an idea to just open up the switchgear and trunking to see what might be going on Inside.

 

[Daveh36]

I have just read some posts in the iet forum about mice finding their way into trunking, busbars and switchgear and ‘setting up home’ with baby mice appearing in due course because the TV reception was lousy.

It might be an idea to just open up the switchgear and trunking to see what might be going on Inside.

Hi Guys sorry been a tad busy trying to sort this out to reply often enough but I'll try answering some of the questions .
The 400amp MCCB which incidentally has fried today is a Siemens see photos.The supply is from it's own substation and the dno inform me it's a clean supply with 400 amp fuses. I questioned any spikes on their side and they ruled it out although they fitted monitoring on it tonight. The previous issues I've been told about are that a tap of unit had been replaced on the 11th floor a few weeks ago not sure why at this stage and am making enquiries as to why. Going to strip that out tomorrow just to be on the safe side as dont know who fitted it. Really am just clutching at straws now.

 

[marconi]

Daveh36. Thank you for the update on this interesting to me yet fraught for you problem. Looking at the voltage and current plots it seems to me that there is a spike of voltage upwards to 272V immediately followed by a spike in current in all three phases. I wonder if the problem is on the incoming supply side ie causing the spike upwards of 22V but it is troublesome to the block's electricity installation because this very short spike when presented across a very low surge impedance inevitably results in a momentarily very high surge current in all three lines. I mentioned before it may be some quirky trick of fate that the supply impedance connected in series with the building wiring surge impedance are resonating at one of the frequencies excited by the spike. You know that at series resonance the current is only limited by the copper Ohmic resistance which may be very very small because the block is close to the supply transformer and only fed by a short cable. With such large currents flowing there will be significant short term overvoltages which may lead to flashover between exposed copper conductors and effectively a three lines short circuit until the fuses rupture.

Why the voltage disturbance happens may be because of an upstream transformer tap change to regulate network voltage.

I think you could sensibly press the DNO to investigate this observation. The coincidence of the voltage spike, on all three phases and generally overnight at the same time. pc1966 mentioned you should check the peak current specification of the current clamps and analyser because the recorded spike may well be much higher than 2kA.

 

[pc1966]

Looking at those plots, yes there are odd voltage spikes, but in the grand scheme of things they are negligible. I don't see this is a DNO problem.

Looking at the current waveform then clearly it is not overload or too little diversity, this is a major 3-phase short taking place somewhere, with a hint the Y phase kicked off and then its friend joined it. I'm still not convinced the logger has the true peak as ~1.7kA or so won't take out MCCB trip and 3 * 315A fuses, but if it was really 5-10kA then it makes sense.

You need to inspect the outgoing circuit's cable and various DBs as a couple of faults like that is going to paint the inside with copper vapour!

What & why is a mystery, but I can't see anything other than a major short/arc event explaining it.

 

[pc1966]

The 400amp MCCB which incidentally has fried today is a Siemens see photos.

A few high-current faults and that happens, contact life can be as low as a few dozen and max rated breaking. It points to the faults being way over 1.9kA though.

 

[Daveh36]

Daveh36. Thank you for the update on this interesting to me yet fraught for you problem. Looking at the voltage and current plots it seems to me that there is a spike of voltage upwards to 272V immediately followed by a spike in current in all three phases. I wonder if the problem is on the incoming supply side ie causing the spike upwards of 22V but it is troublesome to the block's electricity installation because this very short spike when presented across a very low surge impedance inevitably results in a momentarily very high surge current in all three lines. I mentioned before it may be some quirky trick of fate that the supply impedance connected in series with the building wiring surge impedance are resonating at one of the frequencies excited by the spike. You know that at series resonance the current is only limited by the copper Ohmic resistance which may be very very small because the block is close to the supply transformer and only fed by a short cable. With such large currents flowing there will be significant short term overvoltages which may lead to flashover between exposed copper conductors and effectively a three lines short circuit until the fuses rupture.

Why the voltage disturbance happens may be because of an upstream transformer tap change to regulate network voltage.

I think you could sensibly press the DNO to investigate this observation. The coincidence of the voltage spike, on all three phases and generally overnight at the same time. pc1966 mentioned you should check the peak current specification of the current clamps and analyser because the recorded spike may well be much higher than 2kA.

Thanks for that info. I've replaced the MCCB again this morning and re- energised and now that the DNO have monitoring I'm hopeful of a resolution although I'm expecting my phone to ring at about 2.00ish in the morning. I'll keep you updated when I get some results back from my monitoring or the DNO

 

[plugsandsparks]

This is indeed very odd. It looks like a typical direct short but suspect very high current (higher than recorded).
What is odd and has been mentioned you would expect the fault to still be there.
Not sure how many apartments there are and how doable this would be but.
If you could isolate all the apartments and switch off the main incoming isolator and IR test to see if there is residual evidence of L-L, L-N and L-E resistance such that it is low for normal expected result but high enough for you to replace the fuses until the next event.... e.g. you may get one reading that is significantly different to the rest
Keep us posted, its a good one

 

[pc1966]

Thanks for that info. I've replaced the MCCB again this morning and re- energised and now that the DNO have monitoring I'm hopeful of a resolution although I'm expecting my phone to ring at about 2.00ish in the morning. I'll keep you updated when I get some results back from my monitoring or the DNO

Really this looks like a bigger fault inside the installation!

What does the MCCB feed?
How is it cabled?
How are the flats fed from the main supply?

However, if you are going to inspect stuff please make sure the supply MCCB is off! You really do not want to disturb something with a history of high kA faults as it can lead to life-changing or life-ending injuries!

 

[Daveh36]

Really this looks like a bigger fault inside the installation!

What does the MCCB feed?
How is it cabled?
How are the flats fed from the main supply?

However, if you are going to inspect stuff please make sure the supply MCCB is off! You really do not want to disturb something with a history of high kA faults as it can lead to life-changing or life-ending injuries!

Think I'm a little wiser than to work with out turning the MCCB off but hey.
There are 66 apartments spread over the 12 floors. Rising main with tap off box ever other floor feeding fuse cabinets for the apartments. Which are in turn fed via split concentric to individual boards. Bear in mind these were fitted around 9 years ago so no surge protection. Interestingly though I was informed today that the apartment block next door it one had power outage yesterday as well might be coincidental but who knows

 

[pc1966]

Think I'm a little wiser than to work with out turning the MCCB off but hey.

Its a public forum, we can't assume anything!

I would rather risk hurting someone's pride but underestimating them than have to live with overestimating somone by not saying something that ultimately led to an injury.

There are 66 apartments spread over the 12 floors. Rising main with tap off box ever other floor feeding fuse cabinets for the apartments. Which are in turn fed via split concentric to individual boards. Bear in mind these were fitted around 9 years ago so no surge protection. Interestingly though I was informed today that the apartment block next door it one had power outage yesterday as well might be coincidental but who knows

It would take a very high level of lightning hit to trigger an arc, and this is not a country that sees that every day!

I'm guessing the flats are all on 60A fuses and if so they have total selectivity with 100A or more, so anything downstream of those fuses cannot be responsible for blowing the incoming 315A fuses.

So from my armchair view it comes down to a fault the rising main or in one of those boxes on the supply side of the flat's fuses.

Or there is some secret unknown big 3-phase load somewhere!

 

[pc1966]

Incidentally have you been able to speak to folks in various apartments? Has anyone heard a loud noise at the time things tripped out?

Might just help get to which floor(s) are close to a fault.

 

[Spaldingsparks]

someone hasnt got a funny kind of farm in the flats have they?

 

[marconi]

pc1966: I don't feel so ready to dismiss something on the DNO side of the fuses because the timing and regularity of the outages hints at something happening during the quieter hours or lower load periods - so something being done for technical reasons and necessary during these periods and regularly. My money now is on a DNO overvoltage mechanism triggering a busbar arc fault perhaps at an insulator defect. (I discovered by the way high current arcs between busbars tend to move away from their point of initiation due to the usual electrodynamic forces at play.) As ever I could be wrong...

I got myself immersed in this problem to avoid the miserable news about what is happening in Ukraine. This is what I read this afternoon:

https://www.studiecd.dk/cahiers_techniques/Fault_arcs_on_busbar_sets_and_switchboards.pdf

 

[pc1966]

pc1966: I don't feel so ready to dismiss something on the DNO side of the fuses because the timing and regularity of the outages hints at something happening during the quieter hours or lower load periods - so something being done for technical reasons and necessary during these periods and regularly. My money now is on a DNO overvoltage mechanism triggering a busbar arc fault perhaps at an insulator defect. (I discovered by the way high current arcs between busbars tend to move away from their point of initiation due to the usual electrodynamic forces at play.) As ever I could be wrong...

Very true the regular timing of the fault suggests something is a trigger for it.

I can't see high over-voltage (i.e. 8-10kV as mentioned in the article) being the cause as otherwise we would have lots of flat residents complaining of damaged electronics. Still, it would be a good idea to consider having some decent-sized SPD fitted to the incoming supply (e.g. a "type 1+2" and some 100A fuses/MCCB or similar for fault protection).

But the idea that DNO changes are causing electrodynamic forces in a busbar somewhere is very intriguing to me. I was pondering on how you could have a system working perfeclty for a period and then BANG! go the incomer fuses. I was wondering if the rises or one of the fuse cabinets was adjacent to a lift or something that was occasionally hitting it? But maybe there is a loose bar and DNO tap-changing is making it flap?

I got myself immersed in this problem to avoid the miserable news about what is happening in Ukraine. This is what I read this afternoon:

https://www.studiecd.dk/cahiers_techniques/Fault_arcs_on_busbar_sets_and_switchboards.pdf

Yes, a very depressing event all round