Can a second earthing conductor drive the fault current back?

[ChrisToffer]

i have a TT earthing system and to decrease the resistance of the system I'm connecting the rods in parallel. i have the earthing conductor connected but as a mean of security, I'm thinking of connecting a second earthing conductor from the rod to the main Earth bar. Does this cause a any problem? will the fault current go back to my system??

 

[nsaspook]

I don't see a technical problem with redundant wiring 'if' they are on the same earthing bar and follow the same path to the ground rod connection for normal fault currents. It seems like overkill if you've used approved connection methods for the single wire and have a inspection cycle for checking ground resistance and ground connections.

I have no idea about what your local code requirements would be for this service.

 

[jim hardy]

and follow the same path to the ground rod connection for normal fault currents

no idea what's your installation like.

We had an earthing scheme for the instruments in a power plant that incorporated a really big loop, perhaps seventy five feet across with two ground rods on opposite sides. One day a nearby lightning stroke upset the instruments and caused a trip. We surmised it was induction in that loop, though Earth potential drop between the separate ground rods some distance apart couldn't be ruled out. Both would only persist for the duration of the strike.

Routing your wires together per nsaspook's advice will minimize enclosed area hence magnetic induction.

 

[AlephZero]

The downside to this is creating a scaled up version of an "earth loop," which is a good way to pick up electrical noise and RFI where you don't want it: http://www.jaycar.co.uk/images_uploaded/humloop.pdf

Being struck by lightning is an extreme case!

 

[nsaspook]

The downside to this is creating a scaled up version of an "earth loop," which is a good way to pick up electrical noise and RFI where you don't want it: http://www.jaycar.co.uk/images_uploaded/humloop.pdf

This is one of the reasons I set the requirement of the same path for the ground wires. If they are very close at both connections and along the path external signals will be common-mode to both wires and not cause loop currents..

 

[Babadag]

The second connection will not change any thing.

 

[jim hardy]

The second connection will not change any thing.

except that by adding it you've created a closed loop with one turn. Any magnetic flux enclosed by that loop will cause current to flow around it. That's why you want to keep the area of that loop small by running its wires close together, as nsaspook said earlier.

Usually there's not a lot of stray magnetic flux floating about. But around big machinery or lightning arrestors one must keep in mind the laws of Faraday and Gauss.

 

[Borek]

One day a nearby lightning stroke upset the instruments and caused a trip.

Makes me think about step voltage.

 

[Babadag]

I agree with you,Jim, theoretically. Practically the usual induced voltage will be not more than 1 V.
For a distance between grounding rods of 3 m[total 10 m connection length] and 10 m grounding cable length - total occupied area of 100 m^2- a conductor close at 0.5 m, running parallel at entire length of 10 m and carrying 1000 A will induced 12.6 V in the loop.
Usually the area occupied will be not more than 10 m^2 and no conductor carrying 1000 A will be in vicinity of less than 1 m, then the induced voltage will be 0.6 V.
:shy:

 

[jim hardy]

Thanks Babadag

We did a Biot-Savart calculation trying to figure out what had tripped the plant.
I no longer recall the exact numbers.
A lightning protection earthing conductor came straight down from the roof and passed maybe twenty five feet from one corner of our instrument grounding loop. I had seen the lightning stroke hit the rods atop plant and estimated it at 30kiloamps , which is not a gigantic stroke but a pretty healthy one. The instruments physically nearest that corner of the loop were the ones that reacted and tripped the plant.
Would a star grounding system have performed better? Who knows?
I think that having it earthed on both sides held down the induced voltage, else we'd have tripped the other unit too.


So as you observe, induction is not much trouble in ordinary circumstances.
Lightning has di/dt in kiloamps per microsecond , so one ought to consider that when routing lightning protection cables and instrument cables in an industrial setting.

Our main generator connections carried 20 kiloamps at 60hz. With ten turns on a 0.1m^2 loop i could measure 2 volts ac near them, perhaps a couple meters away. A clamp-on ammeter placed around the handrail above read ~75 amps.

So we agree - it's not troublesome for 99.5% of the folks out there.
But it's something to be aware of for that one-in-a-thousand unfortunate event.

 

[Borek]

I can't find any estimates of the step voltages, but if they are high enough to kill, they should be able to trip the plant, don't they?

 

[jim hardy]

I can't find any estimates of the step voltages, but if they are high enough to kill, they should be able to trip the plant, don't they?

It takes surprisingly little to upset an instrument and cause a partial trip signal.

We found no damage to any of the instruments involved.
They operate on a one to five volt signal, derived from a 4-20 milliamp current loop.
So if , via the Earth grounding scheme,
just a couple volts were induced into a signal line, a trip circuit actuation is in theory possible.
It's a "two channels out of three must agree" to actually trip the plant
and the two channels on that side of the control room nearest the strike agreed.

BiotSavart, long straight wire: B= \frac{μ_{0}I}{2\pi a} , a being distance from the wire


in center of the loop perhaps 15 meters from lightning rod earthing conductor

B = \frac{4\pi\times10^{-7}\times30,000}{2\pi\times15} = 0.4 milliwebers/m² , which X 100 square meters yields \Phi = 40 milliwebers

not much flux
but when you consider it all appeared in maybe ten microseconds
[STRIKE]d\Phi/dt = 400 webers per second[/STRIKE]
[STRIKE]which in one turn induces 400 volts.[/STRIKE]

oops!
d\Phi/dt = 40 milliwebers / 10μsec = 40X10^-3/10^-5 = 4000 webers/sec
which in one turn induces 4 kilovolts

I found a newer reference on lightning than the old book i had at the time, which was from 1950's.
It suggests lightning di/dt in range of a few hundreds of kiloamps per microsecond

With respect to the current steepness, positive lightning can be omitted because the current normally does not have a high rate-of-rise. The highest current steepness is found in nega
tive subsequent stroke as can be seen from Fig. 13.
The average current steepness between the 30 % and the 90 % current level of negative subsequent strokes is about 200 kA/μs.

http://www.iclp-centre.org/pdf/Invited-Lecture-3.pdf

which of course gives higher result than [STRIKE]the 400[/STRIKE] i got above.
But the absence of damage suggested actual transient voltage presented to the instruments was [STRIKE]somewhat [/STRIKE] quite a bit lower.

i hope my rusty old math is okay above, further corrections are welcome.

old jim

 

[Borek]

Just to make sure we are talking about the same thing (could be these are different sides of the same coin). My understanding was always that the step voltages are not induced (that is, they have nothing to do with existence of the loop), they are just because two points on the ground are at different voltage. The simplest explanation being that if the point where the lightning strikes has some potential, and the ground has some resistance, voltage of the surface points around changes with IR, the further from the strike we are the lower the potential. Something like that:

and that:

http://en.wikipedia.org/wiki/Earth_potential_rise

(sorry if I am talking about obvious things, I am just not convinced we refer to the same effect).

BTW: check your LaTeX formulas now (right click them, and select Show math as Tex command)

 

[jim hardy]

Ahhh Thanks ! I mis-understood your meaning of step to mean a step function 1/s ...

We surmised it was induction in that loop, though earth potential drop between the separate ground rods some distance apart couldn't be ruled out.

your 'step' was my 'earth potential drop'.

The conclusion was it was one or a combination of the two effects.

I learned from the design engineers who came down from Bethesda to help investigate the event:
Lightning protection has its own ground rods, separate from plant's ground mat. The lightning grounds go way deeper than the wire mesh ground mat that's buried under the plant. The idea is to dump lightning charge well below the ground mat so the mat can act as an equipotential "magic carpet" that rides up and down with the local Earth potential. All plant equipment is tied to the ground mat so it can float together with little or no 'step voltage'.

The design folks said they'd addressed 'step voltage' by previous paragraph
but said they hadn't figured induction when routing their lightning ground cables. This was a mid-60's design.

Inspection found a few lightning rods missing and some wires in disarray from construction work atop the dome. The lightning system was put back in tip-top shape and we went back to making cheap electricity.

 

[jim hardy]

PS Thanks for fixing up the formula. I will learn to use that feature !

i seem to be inordinately awkward at these things...