Generator Earth Testing

[davey345]

Hello,

I have got myself confused with testing a generator installation.

The generator is A 100kVA 650Vac single phase machine. 

The generator is bonded down to true earth via two 16mm - 4ft earth electrodes, and then bonded to the neutral star point of the machine. to create a TNS.

The output is then fed via two 50mm^2 cables to a large stationary metal building. on the side of the building there is a large 10kVA commando type socket rated at 10kVA and 154 Amps, that is then connected to a 650V, 160A MCCB.

The unit is bonded down to true earth via four 16mm earth electrodes.

Inside the unit there are a number of  consumer units all with RCBO type b's fitted to every final circuit.

All the normal tests on each board have been carried out an all fall within the normal parameters and Zs disconnection times for the type of RCBO

As the generator is over 10kVA the manufacturer has recommended the neutral be bonded to true earth.

Do i take the zs reading for TNS as 0.80 and then divide into the 650V to give a fault current of 813A

I have measured the Ze as 0.48, and that was with two earth electrodes, with a fault current of 1354A, more than enough to trip the over current MCCB device.

But for the building do i need to test the earth electrodes ?

David

 

[Sidewinder]

So this is a rail installation?

 

[binky]

I would test the building earth rods. I would have thought something like this would be a TT system?

 

[Andy™]

I would test the building earth rods. I would have thought something like this would be a TT system?

why would it be TT?

you have a earth connected to the star point. no different to a TN system from DNO, there's a rod at their transformer

 

[davey345]

This a rail installation, and i know that rail are exempt from BS7671, however it will be classified as a static unit and will not go into traffic, so the installation is the same as connecting to a mobile static unit and the public will be using the facility so it comes under BS7671 from my understanding. 

I am looking to use Method 2 of the earth fault loop impedance testing, replacing RA with the earth impedance measured at the tns earth electrode, but unsure what to replace "In" with as there is no RCD only a MCCB between generator and train. 

 

[davey345]

the generator is bonded to true earth by two 16mm electrodes and the earth electrode is connected to the neutral star point so the source is a TNS. it is just how to go about proving that sufficient current will flow to cause a large enough current to take out the generator fuses or MCCB isolator, the gen has a fault earth current protection on it set at 10% of full load current, which would become 50A if there was no other discrimination in the circuit, but the MCCB has short circuit and overload protection on it so would trip before the generator. i just need to be able to put all this doen on paper and prove it.

 

[Andy™]

your rod isnt to earth the installation as would normally be required by 7671, its to earth the generator to give a  return path. ideally you want it low, iirc DNO work on 20 ohms. as for your Zs, normal rules apply, if earth loop is low enough then you should get them to tripm just be check the genny is capable of that current long enough for the device to operate

 

[davey345]

I realise i am going to regret this and look a little bit of an idiot, but if the gen star point is bonded to true earth, then that is a TNS ? it then has ONLY 2 50mm^"2 LINE and NEUTRAL conductors to the outbuilding, where the outbuilding is only bonded to ground via 4 earth electrode rods.

I have already taken off the two earth electrodes from the generator and using a 2 wire high current loop impedance test, tested between the 650V AC line and earth electrode

this gave me 14.3 ohms, i repeated this for the other electrode which gave me 14.1 ohms, so using resistors in parallel i got a resultant resistance of 7.1. how do i test a TN system then, G3 only shows TNS, TNS-c and TT, it give disconnection times for TN and TT, i am beginning to loose the will on this one, feeling very thick. 

 

[davey345]

So the Ze with earth electrode disconnected is about 7.6 ohms and then zs with it connected, i would expect a similar value to the ze reading, however lets say i have a zs value of  say 6 ohms, then surely the pfc = U0/zs = 650v / 6 = 108A which would not be enough to clear trip the MCCB or blow and gen fuses? where am i going wrong please?

 

[binky]

why would it be TT?

you have a earth connected to the star point. no different to a TN system from DNO, there's a rod at their transformer

ok, I don't get involved with gennies, so it's an area of work I haven't studied.

 

[Ardet R]

You have your supply generator as a source and the neutral is earthed (resistance to earth 7.1Ω?). 

This earth path is only transferred to the building via the Neutral conductor so it is using a TNC conductor.  I do not know the rules for this under the rail regulations but not permitted on the consumer side of the installation in BS7671 and obviously no RCD will work prior to this.

If at the building the circuits are earthed only via the building earth rods then you have a TT system at the building, if the neutral from the generator is split out to create an earth for the building then you have a TNCS supply with additional earth rods at the building.

The Ze at the building is 0.48Ω? but whether from the generator and building rods or building rods only I do not know.

If the generator can on a short circuit (either Line Neutral or Line Earth) deliver 1350A without stalling / frying then the MCCB will operate.  If it can't generate this then you need a means of protecting the generator from a short circuit.  This could be by low level magnetic trip circuit breakers set above the FLC of the generator but below the level at which the generator would be damaged.  Earth fault protection would be from the generator protection or the final circuit RCDs.