designing for a distribution system for worse case, i.e with a stand by generator in circuit, how do you establish your source information

Ask the generator manufacturer.

    - Andy.

Taking up on some earlier comments, using RCD as a solution, raises a few other issues I.e. by doing so you are making the system, particular when maintaining essential supplies less secure.

as an example on one project, a UPS was used to maintain lighting on mains failure. The lighting design had a number of distribution boards housing 5amp fuses for final sub-circuits. the UPS manufacturer was asked to prove that during FAT, under an fault condition would the 5amp fuse disconnect, L-N and Zs impedances for the circuits were given for simulation, they could not provide evidence.

i guess has anyone got any empirical evidence on how these issues were overcome.

UPS - indeed any inverter supplied system - poses interesting challenges. The current that most electronic inverters can supply into a fault is extremely limited - perhaps only 2 or 3 times their nominal output. Obviously that's going to call into question using conventional overcurrent protective devices for fault disconnection. In some circumstances a fault of negligible impedance (or even an approximation to that) will cause the output voltage to collapse to such a low level that the dangers from shock are removed anyway. Maintaining a supply to fault-free circuits is challenging. If sufficient supply remains to operate RCDs you could have something like RCBOs (rather than 5A fuses) on each final circuit - or if that won't work, have multiple UPSs - so a single failure only affect a subset of all the lighting circuits.

   - Andy.

It might, exceptionally, be considered acceptable to use engineering judgement to permit of a longer disconnection time, under generator power, than would ordinarily be allowed.

Consider the following. Q why do we require that earth faults be disconnected in some specified and short time ?  A to reduce the danger from touch voltages during the fault. During an earth fault on a standard mains supply, it is likely that the supply voltage will be maintained at close to the nominal voltage of 240 volts.

Presuming equal voltage drops in the line conductor and in the CPC, That will give about 120 volts on exposed metal work during the fault. Dangerous, hence the requirement to disconnect such faults promptly.

Now consider an earth fault on generator power. During the fault, the generator output voltage will NOT remain even roughly constant but will decay to a very low figure. The touch voltage will be half of this much reduced voltage. 

This only applies to circuits such as sub mains that are large in relation to the generator capacity. On small final circuits the generator should operate the OCPD device promptly.

There may be special cases, see BS 9991 and BS 9999, where in a fault (which could be L-N or L-L, and not related to ADS or other methods of protection against electric shock) it's better to keep a generator supply running - but these are special supplies to parts of safety services. And even there, the standards don't say it's acceptable to injure people by electric shock (which could be achieved by other means than ADS), and the measures would be part of wider risk assessments in any case.

Anthony Smith said:

And the UPS reason fo existing I.e. maintaining essential supplies is thwarted 

Many UPS are provided for business continuity, rather than for reasons of safety.

When we are talking about safety services, other things (like higher integrity wiring systems etc.) come into play, and a basic UPS may not cut the mustard (for example certain UPS could be used to provide a CPS but they have to meet all of the requirements of BS EN 50171 and not simply BS EN 62040 series.

Anthony Smith said:

as an example on one project, a UPS was used to maintain lighting on mains failure. The lighting design had a number of distribution boards housing 5amp fuses for final sub-circuits. the UPS manufacturer was asked to prove that during FAT, under an fault condition would the 5amp fuse disconnect, L-N and Zs impedances for the circuits were given for simulation, they could not provide evidence.

This is a supply to safety services, and should comply with the relevant standards. It's a specific case. See Regulation 560.5.3, which leads to a design decision to perhaps consider IT system for those circuits? Where BS 7671 applies to this type of installation, Regulation 560.7.4 effectively tells us to provide an appropriate number of circuits so that OCPD operation does not impair other circuits.

BUT

See earlier post - this isn't simply a UPS, but a CPS - and the requirements of BS EN 50171 and Regulation 560.6.10 of BS 7671 apply to the "UPS" itself - but the whole system ought to comply with the BS 5266 and BS EN 1838.

AJ says ask the Generator manufacturer for parameter details, that is clearly the first place you would go, they will give you a Prospective Short Circuit value, transient and sub-transient value how would these values be used to provide a Ze, a true Prospective short circuit current value to start to evaluate system impedances

It will be a very brave genset maker that gives you any sort of single figure for the PSSC !! The output impedance of a genset and its ability to supply current is very load dependent - if is running at say less than 1/4 capacity and you apply full load - not a short, just full load mind, as a shock, you will at best collapse the output until the regulator recovers, or at worst stall it. (*)  What happens in short is similar, but you will not get out much more  than about twice full load current even as a short transient, without the output collapsing, and nothing like the ten times FLC or so required when trying to blow fuses in a sensible time.

Readings made with a loop tester/PSSC meter will be about as much use as a fruit machine - all it tells you is what happens if you scale up the voltage drop from a small test current, what would the current need to be as if it was 100% voltage drop. But that assumes of course the relation between current and volt drop is linear and unless the currents in consideration are small compared to the genset full load then it simply isn't true.

So if the genset has a full load of 500KVA, call it 800A per phase, the largest fuse or MCB you should expect to be able to always fast blow, in all load conditions, is in the range 80-150A . The absolute upper limit for setting  earth leakage trips would be similar, but less 10A and some delay is more credible.

Mike

(That is not quite worst case, I have not so fond memories of squeezing in to an enclosure never intended for it, alongside an engine that would not have looked out of place in a really large bus or some piece of farm machinery, to replace a shear-pin between the motor shaft and the alternator. That unit was probably 1/4 MW or so.)

Mapj1, thanks for the reply I knew these issues. But trying to make decisions when designing a power system which has a Stand-by Generator, what parameters would you use for Ze and PSC, given that these values need to be assessed for disconnection times at all points of the installation, which would be worse case in the design, clearly max. Fault level will be determined by the main Transformer characteristics. I have designed numerous Power System arrangements, but the Stand-by Generator parameters still cause me a problem