Earth leakage circuit breakers in series

I have an unusual TT set up for a large multiple unit filling station. By the time I get to the final circuits there will be five RCDs in "series". 400A main intake, 200A to separate buildings on site 200A incomer to panel board in each building, various in the panel board 160 to 63A  to the final circuit distribution boards then individual MCB/RCD for each final circuit.

On the 400 and 200A units I have adjustability 0.03 to 10000mA and time delay 0,60,150,310mS but on the 160A-63A units 0.03 to 5000mA and 0,60,150,500,1000mS. Schneider Compact NSXM all 4-pole. 

I am unsure how much reliance can be placed on the exactness of the time-delay but the maximum time delay on the larger units is only 310mS. The larger units are already in place by others. We are responsible for one of the units thus commencing with a 200A module.

The 400A module is set at set at 10000mA /310mS and first 200A at beginning of distribution circuit is 3000mA and 150mS. So I need a 200A incomer to the panel in our building and 160/63A for the distribution circuits in the panel in our building. I am struggling to see how I can achieve appropriate earth leakage selectivity. Earth loop is currently 2.36ohms at our unit intake and much the same at main intake so around 100A of possible earth fault current. 

I also need 2-pole or 4-pole RCD protection for my final circuits to avoid neutral/earth faults being left in. Cost is mind-boggling!

Just thought this was an interesting set up that probably few would experience. 

  • You also have a good earth at the substation I hope to operate that 10A setting Astonished promptly - do we consider a 9 amp fault to be safe forever at that level  that's the after diversity demand of a small semidetached house, as earth leakage ?

    An interesting challenge.

    On the odd occasion I have had to set up installations with nested earth fault protection, and I will say never a fuel station (!),  I have used a rule of 2 or 3 - so 30mA instant is backed by 100mA 0.1 seconds, in turn backed by 300mA and delay of 0.2 to 0.3 seconds so far this has always worked without a hitch, but I am not sure how much one can squeeze the settings up a bit before things start simultaneous tripping.

    In your shoes I'd be asking what the various levels of disconnection do, and how serious the collateral damage of the next layer up tripping as well would be - if I knew the designers well I may ask why they think they  need so many levels of 'defcon' .

    All but a lowest and fastest levels of  trip should indicate an attack on the infrastructure, damage to buried submains and so on, rather than a faulty component, and it may be that in such a special  case a wider shut down is OK.

    It may be worth a call to the maker of the earth fault relays  about the discrimination.

    Let us know how you get on.


  • I've probably not pictured the setup entirely correctly, but it sounds suspiciously like there's an RCD on both ends of the same cable in some instances (200A ones at least) - arguably the 2nd of these isn't strictly necessary and could either be omitted or at least need not discriminate reliably with its upstream sibling. So maybe you only have three or four tiers to achieve discrimination between rather than five - which might be a little easier.

       - Andy.

  • Hi Lyledunn,

    Given I have specialised in filling station electrical installations for the last 40 years and sit on the Tech Committee that produces the tech standard ( Blue Book) for that unique industry, then please send a pm if you have specific questions. On a typical UK filling station normally three levels of time/earth leakage current cascading, after all it will be a TT earthing system, so depending on the technical spec of the client? usually the principal protection device is set at between 1A and 3A earth leakage (Not usually any higher than that) and a 0.5S time delay or thereabouts. The submain protective devices usually set at 300 - 500mA earth leakage with a 150 - 300mS time delay and then your final circuits 30 - 300mA with instantanious delay.

    Yes, cost is expensive................

    Regards GTB