Disconnection times on a system if a stand by generator is used

On a fully designed power distribution system, clearly the fault levels, earth loop impedances and ultimately fuse disconnection times are determined by the system parameters Transformer %Z etc one or two in parallel for min and max fault levels levels. What needs to be considered when bringing on a stand-by generator, clearly all Fault levels, Earth loop impedances and consequently fuse disconnection times would change because of the Generator parameters. How would you design in these new parameters, bearing in mind that disconnection times on a large system may no longer be achievable 

  • Earth loop impedances and consequently fuse disconnection times would change because of the Generator parameters.

    Correct. So your design needs to cope with the characteristics of the supply when the supply is the generator alone (where that's one permitted combination). The requirements for disconnection times for ADS are the same regardless of the kind of supply. In some cases that'll mean you can't rely totally on overcurrent protective devices to provide ADS - so some kind of residual current operated devices may be necessary - RCCBs, RCBOs or earth trips on larger devices.

    With very small generators it may be permissible to adopt some other measure than ADS to provide shock protection - e.g. a separated system - but often that's only practical for very small systems - e.g. a portable generator feeding a single Class I tool.

          - Andy.

  • I would agree totally, but on a large distribution system, with a stand-by generator used at source, this would be impossible and I believe you are right, but I have never seen this in practice, to maintain disconnection times for stand by, cables sizes would be so large as to make this impractical

  • Wait to see what the "big boys" have to say (my personal experience is on smaller scale systems that yours sounds to be), but I would have thought that increasing cable c.s.a.s wouldn't usually be the way to go - but rather choosing different protective devices and/or sacrificing an amount of discrimination (worst possible case would be an RCD or MCCB+earth leakage trip on the output of the generator).

       - Andy.

  • on a large distribution system, with a stand-by generator used at source, this would be impossible

    RCDs may be used for ADS. .

    but I have never seen this in practice, to maintain disconnection times for stand by,

    The design may require more RCDs, but ultimately, Andy is correct

    Further ...


    Regulation 551.4.3.1 requires ADS according to Chapter 41 to be provided for circuits supplied by the generating sets, and Regulation 551.4.1 requires fault protection to be provided for each source, or combinations of sources.

    So, if disconnection times not met (one way or another), then BS 7671 is not complied with.

    Therefore, if disconnection times are not met and the installation is supplied by a public supply network, this is potentially a breach of Regulation 21 of the ESQCR, which requires the installation to comply with BS 7671 where a switched alternative to the public supply is provided.

    Some people at this point may at this point discuss the operation of UPS (and application of BS 7671, in particular the two Regulations quoted above, and Regulation 551.4.3.2.1.

    However, I think it's very tenuous to determine that a UPS is not a "switched alternative" to the public supply, because, in bypass mode, the grid supplies the UPS loads, whereas in "inline" and "mains fail", the generator (UPS inverter) supplies the loads, powered by the DC 'charger' in 'inline' mode, or the battery in 'mains fail' mode. So, I think Regulation 21 of ESQCR ought to be applied to UPS (as well as relevant parts of Section 551 of BS 7671, including Regulation 551.4.3.2.1).

  • Generators have a very different behaviour under fault than the almost limitless street supply - initially the frequency may fall, and once it drops out of voltage regulation, the effective impedance rises dramatically, and may be near infinite if it stalls. It is not safe to assume that there will be enough 'oomph' to clear a large fuse  or breaker (i,e, one near the full power rating )reliably.  The solution as others have alluded is a suitably sized earth fault relay to ensure ADS against low impedance earth faults,  (LN overloads are allowed to take their time.,, ) and if required then co-ordinated RCDs of varying delay on any sub-mains whose ADS would not reliably  clear in time for safety on earth fault.

    Mike

  • I agree increasing cable sizes is not the way forward, but still interested to know how you would design your system, regardless of size on a stand-by generator parameters, as these parameters are clearly different to transformers. And I’m not sure how the transient and sub transient reactance would be plug into a power system were you are trying to establish fault levels, Ze, etc

  • I know I am maybe not explaining my issue very eloquently, but just say you have a 1.5MVA transformer at say 4.74% imp. It is clear the fault level is 31.5 MVA and from there system impedances can be determined, designing for a distribution system for worse case, i.e with a stand by generator in circuit, how do you establish your source information

  • Ze is the one thing that may be lower - the local genset wll be TNS. Do not confuse it with electrode resistance - that may be quite high - several tens of ohms. You may not relay on the street earth when running on genset, as the main may be cut-off.
    Typically set the earth fault trips at 1% to 10% of the full load current for that branch if 30mA RCD function is not needed. Going the other way, final ccts will have to be 30mA RCD, now to reliably discriminate with the next layer up you need at least a factor of 3, so  be 100mA 0.1sec, and then 0.3A 1/3 second. It is not common to have more than 3-4 layers between genset and load.

    Mike.

  • Well that level sets the upper level fault handling of your design, and may need death or glory fuses near the origin, or surprisingly large front end MCCBS etc. But, you also have a second limit - with the rather uncertain and load dependent source impedance of genset it must also operate ADS properly,  Usually you cannot do both of these  on MCCB and MCB alone. Enter the earth fault trips, as above ;-)
    Mike.

  • It was mentioned earlier, but UPSs have a similar issue for me, re. Disconnection times and I feel from what I have seen in practice these issues are ignored, if a fault occurs when the UPS is being utilised, (I know the likelihood is extremely low) because the disconnection times are not met, the UPS can shut down totally. And the UPS reason fo existing I.e. maintaining essential supplies is thwarted