Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Splitting power between cables by rating

  1. Aug 10, 2015 #1


    User Avatar
    Gold Member

    How are power distribution cables parallel-ed to split loads optimally? If the cables were exactly identical this would be easy but how is it achieved in general?

    e.g. Say there's a 11 kV to 420 V distribution transformer but the primary load is some distance away. Assume the load increases over time and the Ampere capacity of the original cable from TX to the load is full.

    As it often happens the transformer does have spare capacity so ideally only a second cable would be needed to transfer power. Are they simply paralleled between the two bus bars (TX and load) or is a more complicated scheme needed to be used for balancing?
  2. jcsd
  3. Aug 10, 2015 #2


    User Avatar
    Gold Member

    Say you have two cables in parallel, both made of copper wire but with different cross section area, then the current densities will be the same, and thus the power loss per volume/meter will be the same. The cooling of the cable will be proportional to the surface/meter, so the cable with the bigger cross section area is the one in troubles. I think that two cables, say made of copper/aluminum, will never be coupled in parallel.

    But another serious problem arises when just one cable crosses a district heating tube / dry soil, whatever: The cable will be heated up locally, and thus the local resistance as per meter will be increased. This results in a local increase in power loss, which will increase the local temperature even more, and so on. It's called a "hot spot", and it will burn the cable.

    As for a copper cable, the min. length of such a hot spot can be calculated to be about 10m. I spoke with some power engineers about this calculation, and they were amazed and said:
    You are absolute right: Always 10 - 12m burned cable.
  4. Aug 10, 2015 #3


    User Avatar
    Gold Member

    How do you calculate the minimum length as 10 m? Can you elaborate? Are you including axial conduction of heat?
  5. Aug 10, 2015 #4


    User Avatar
    Gold Member

    Yes. The result should be used when drying some cable in vacuum by means of heating the cable with a 1000 amps current: How close should the temperature of the cable be measured to detect a hot spot during the drying process? When a hot spot was detected: Switch off the current, wait an hour, try again with a decreased current.

    The calculation was done by numerical simulation: I placed two points with some distance, wherein the temperatures were kept constant. Then having a point in the middle where the temperature as a start was set to +1C°. Then switching on current to see what happened using differential equations.

    Now, sequential the distance between the two points was adjusted until a hot spot did not develop into a "thermal runaway":

    Result: Critical distance ≈ 10m.
  6. Aug 10, 2015 #5
    I'm not really sure what you mean by primary load?

    Do you mean the load connected to the seconday (LV) of the transformer?

    The general way to do it is this (attached)

    The tails should be rated so that they can take the full load of the TX. They are pretty short, skimping and putting in smaller tails makes no sense - because if you want to upgrade them you got a world of pain and a big job!

    Loads are connected to feeder cables that are fed from the LV bars either through fuses, or CBs. It makes sense to have spare ways for future needs.

    These feeder cables can all happily be different sizes.

    Say, you had a small load going to a single property, that might be a 35mm cable with 100A fuses

    On the other hand, you might have a 150mm cable with 300A fuses

    Some loads might require two feeder cables. Some loads might have several transformers, each with several feeder cables.

    Another thing you might want to do is have two transformers feeding the same LV bars with an open bus-section between them. This lets you take out one of the transformers for maintenance.
    Or you could backfeed the LV bars though one of the feeder cables (ensuring your LV TX breaker is isolated)

    make sense?

    Attached Files:

  7. Aug 10, 2015 #6


    User Avatar
    Gold Member

    @William White

    Yes. That's what I meant. Sorry, should have been clearer.
  8. Aug 10, 2015 #7


    User Avatar
    Gold Member

    To clarify the arrangement is something like in the sketch below. The question is whether the new cable has to be selected to be an exact match of the old? Or is there some other scheme used to split loads. Although the current cable has a 185 Ampacity the transformer limit & expected load isn't high enough to really need another 185 Amp cable. The max load would only need 250 Amps.

  9. Aug 10, 2015 #8
    250A total (thats nothing) or 250A per phase (which is also not a lot!)?

    The tails you have are pretty small if they cannot hold 250A per phase

    So you may as well do the job properly.

    Put in a single set of tails that are good for the job, prefarably that are future proof, and abandon the bit of wet string you've got doing the job at the moment.

    otherwise you have got all sorts of hassle with your protection on the two CBs (I assume they are there for some sort of earth unit protection)
    Last edited: Aug 10, 2015
  10. Aug 11, 2015 #9


    User Avatar
    Gold Member

    @William White

    250 A per phase. The transformer is rated for 250 kVA.

    What do you mean when you say "tails"?

    I'll try to explain the situation more. The transformer is located some distance away from the loads & there's a distribution room ( bus bar + CB ) close to the loads. It is the cable running from the transformer to this distribution room that is falling short on capacity. This is the 300 feet long cable. Extra capacity needed is only 30%

    In increasing order of cost the options are:

    (a) Add a smaller cable in parallel to existing cable such that both together will satisfy the total demand

    (b) Add a similar capacity cable in parallel to existing cable such that both together will have a spare capacity remaining in excess of the load demanded as well as what the TX can supply

    (c) Get rid of the rather new and perfectly sound current 300 feet of UG cable and replace with a new and larger Cross Section cable that can supply the full demand by itself.

    Basically (c) is the most expensive solution. But if the other two solutions are technically iffy then we must do (c) I guess. Another downside of (c) is that it is a non-redundant design. i.e. Any future damage to this one cable will shut down the show.

    PS. In this specific case, since the currents aren't large and the distance is only 300 ft the cost difference between the three options (a), (b) and (c) isn't a killer. But a part of the question is me wondering how this sort of problem is solved in general for larger cases where the cost might be significant. e.g. 3000 ft line supplying 1 MW that needs a load increase to 1.2 MW.
  11. Aug 11, 2015 #10
    Tails is the term to describe the conductors connecting the secondary windings to the LV TX CB. By using the term tails, you have less confusion. Tails, feeders, pilots etc...they are all conductors, but you use different words so you are clear what you mean, because they are for different things.

    When you look at the cost, have you considered everything? Digging up land, installing, jointing, testing, comissioning, reinstating land.

    Running another parallel set of cables just seems like a bodge to me. You are better putting in the right cable for the job.

    The person that installed this needs shot. You have a 250A load. If this is total, split across three phases, then what size is the cable, 35mm? (more to the point, what size is the TX? Is it just a tiny 100KVA sat up on a pole or in a little cabinet?

    Bodging the job again will only give you grief in the future.

    you say if the cable (c) is damaged, it shuts down the show - what happens if you lose one of the cables if you have two sets? Same problem (well you would lose some load). I assume they will be in the same trench, so if one gets damaged, its possible they both will be. And what is going to damage them? What sort of land is the cable crossing? Are people likely to be digging around the place? Put the cable in some rigi-duct for some protection, with plenty of yellow tape six inches above that.

    Your last question. If the load was 1MW and this was via the 11kV network, you would either have two 750KVA transformers or a large 2 MVA transformer. Previous option is best, because during maintenance, you can take one side out (reducing the load to essentials only so the TX can take it). Larger TXs (above 750KVA) give you larger short circuit currents, and you need better, more robust protection. All adds to the cost.

    But in general, if the load increases, and the original supply is not up to the job you get a new supply and pay for it. Simple as that. Its why you need to think about the future when installing your network, and ensure you have a few spare ways on your LV board and a TX with plenty of spare capacity. The different price of the hardware is not the real cost. The cost is the man hours and time installing it.

    I had a job a few years ago where a land owner is Sussex wanted a water pump for his polo fields. The nearest available line was a 33kV line a mile away across the downs and through a forest. It cost him hundreds of thousands of pounds and took over a year to get through wayleaves.
    Last edited: Aug 11, 2015
  12. Aug 11, 2015 #11


    User Avatar
    Gold Member

    Thanks @William White

    I am convinced that swapping the entire cable is indeed the best alternative.
  13. Aug 11, 2015 #12

    do it properly.

    Put in 120mm (300A) or 150mm (340A).

    Depending where it is, you might not need the expense of an armoured cable.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook