Is it necessary join separeted earth systems? 3

[Bronzeado]

Hi folks,

In our water supply project (PISF) each 230/6.9 kV substation feeds a pump station that is 200 m away. The substation and pump station have separated earth systems. The 230/6.9 kV transformer is delta/wye-grounded connected.

Is it necessary to connect these two earth systems? If so, how it should be done.


Best Regards,

Herivelto S. Bronzeado
Brasília, Brazil

http://www.linkedin.com/profile/view?id=46319837&trk=tab_pro

 

[Skogsgurra]

Herivelto,

If it weren'y that put the question, I would not care to answer at all. But, knowing you, I think there might be more to it than meets the eye.

These two systems are close enough to have some measurement, data or communication via cable. Then, I would say that it is absolutely necessary to connect the two grounds. Or make sure that every auxiliary system has enough insulation to withstand the worst possible fault scenario in either station. I would do that even if there are no such cables installed today. Someone might just do it later - be it an innocent indicator lamp, CCTV or temporary measurement system. One never knows.

But then again, I would not connect the two systems. And then again - I wouldn't be able to make up my mind.

If the systems were far away from each other - no probs. No connection. And that is what everyone does. Local grounds are the way to go. Universally.

But these systems are so close that you can have optical contact between them. I think that is where the problem is. They are so close that one thinks they need to be connected.

Tricky question. Exactly what one could expect from you.

I understand that the pumps are 10+ MW, or so. If they were smaller - I would try to run a 690 V cable directly between the two places. But that will not be possible at these power levels.

No, I have no answer. Just did some loud thinking.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[Bronzeado]

,
Thank you very much Gunnar for your “loudly thinking”. With your experience you really see much more than “it meets the eyes”.

There are some data or communication between the substation and pump station via optical cables.
For redundancy there will be two optical cables: one aerial and another via steel tubes in the ground separated by connections boxes at every 40-50 m. I think that will be enough to isolate the potential between the two earth system.

As you said t is really a trick and though question!!

I hope some more experts like you could post others “loudly thinkings”


Best Regards,

Herivelto S. Bronzeado
Brasília, Brazil

http://www.linkedin.com/profile/view?id=46319837&trk=tab_pro

 

[Skogsgurra]

Sorry, about the 690 V. At 6,9 kV (as I should have said) - would it not be possible to run a cable directly? Without any extra transformer?

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[Skogsgurra]

Oh, oh. Getting worse and worse. There IS no extra transformer, I think. Or have I missed that, too?

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[squeeky]

Seeing all the alcolades for Bronzeado in other posts, I know he knows the answer already. Here I go with my loud thinking. Soil resistivity and the impedance back to the star point of the secondary of the transformer at 6.9kV needs to be accertained. Then you look at the fault level and the expected fault current. The motors will be running without a star point and there is no earth on any connection so there needs to be some sort of an earth reference and ensure the step potential is covered. The pump station requires an earthing system. 200m is not so far but far enough to make a problem.

I would link the two earthing systems and pay particular attention to access ways between the two as far as step potenital is concerned.

Next is lightning strikes. Look at the strike density and then install the appropriate lightning protection. Your earthing and bonding should take this into consideration.

Ok I'm ready to be marked.

 

[jghrist]

If there is no metallic connection between the two grids, then a ground fault at the pump station has to return through the earth only. This can cause two problems. One, the fault current may be insufficient to operate ground fault protection. Two, the return current through the earth may cause dangerously high step- and touch-voltages at the pump station.

 

[Bronzeado]

Sorry Squeeky but I don´t know the answer.

I have to thinking on that as I have divergent opinions between our experts.

Before I mark you, I would like to "explore" you a little bit more: how to connect the two earth system? Just a wire?

Best Regards,

Herivelto S. Bronzeado
Brasília, Brazil

http://www.linkedin.com/profile/view?id=46319837&trk=tab_pro

 

[prc]

I dont claim to be an expert on these issues.But Bronzeado, please be little more specific. Normally 230/6.9 kV transformers will be Star,N earthed/Delta and not delta/star as you mentioned.If first is correct, unerathed 6.9 kV will be going to feed motor.Then How unearthed system has got N at pump station? By grounding transformer? By neutral of stator winding?

 

[Bronzeado]

prc,

The substation transformer 230/6.9 kV is delta(230)/star-grounded(6.9) (please, don´t ask me why!) feeding a 6.9 kV synchronous motor in the pump station 200 m away. The soil between the substation and the pump station is very dry and rocky.


Best Regards,

Herivelto S. Bronzeado
Brasília, Brazil

http://www.linkedin.com/profile/view?id=46319837&trk=tab_pro

 

[7anoter4]

In my opinion, as per IEEE 80/2000, one has to calculate the touch and step potential at motor site.
If the potential is too high then one has to reduce the current flowing through grounding and a new connection may be required. One has to take into consideration the split factor of the live connection: cable shield or static wire. The grounding grid configuration may be improved too.
Another danger may be the transferred potential from the transformer site.

 

[squeeky]

To join the two earth mats we would be guided by our approved service provider but by experience, they would lay two seperate grids (like a horizontal ladder) bewteen the two with earth spikes at intervals. Parrallel reliablity for leg growing copper. These would be exothermically welded to the spikes and then covered with Empire tape (waxy cloth tape). In hard / stone we have had to drill holes for the earth spikes and pour a special resin into the hole to bond it to the rock. The pipe work would be bonded if the gaskets prove to be to much of an insulation. In the fields we have resorted to stainless steel spikes.

We tend to have a switchyard with a heavy grid of copper 70mm2 bare copper with earth rods. The incoming is either 33kV or 11kV depending on the location. Different countries reticulate at different voltages. This feeds a switch room remote from the switch yard but not more than 70m. The building is earthed and bonded to the switch yard normally on two seperate connections. The pumps are normally very close and not more than 20 cable meters away. Earth bars bonding everything together are installed for test purposes with removable links to the section connected to the earth mat. You break the link and then confirm the earth resistance annually. All of these installations are covered for lightning strikes with aerials and bonding to steel roofs, down pipes etc.

You should find some reference data in IEC 10199:2010 and in table E1, there is a reference to different materials such as aluminium and stainless steel. Copper tends to sprout legs and walk off site. We cannot use aluminium as we run in alkiline soils and it "soaks" away. There are products on the market that have a high tensile steel set of wires in the copper and will break even large bolt cutters but it is too expensive for us we we have started concreting them in.

If you can send a plan view I can have a look and give you some more specfic feedback.

 

[Kiribanda]

Bronzeado,
At the pump station why an earth grid is required for the motor? Earth fault return path shall be taken back to the 6.9kV source
by the armour (if properly selected) or by a separate equipment earthing conductor (PE) running with the power cable of the motor.
If it is wired in that way there are no step & touch potentials at the motor end as mentioned by Jghrist too. That means the bonding
is not a concern.

 

[slavag]

Herivelto, what about some control cables between two parts: trips, interlocks, voltage, current ( for protection ).

 

[waross]

I was going to post similar to jghrist.
I would be very concerned with touch and step potentials if the systems are not well bonded together. Also, with fairly high resistance soil, you will have less fault current and a higher voltage difference between the ground grids. In the event of a ground fault, the service conductors and the ground path act as a voltage divider. The higher the resistance of the soil, the greater part of the line to neutral voltage 400 Volts will drop across the soil between the two ground grids.
I would look at possible step and touch potentials in the worst possible case with the entire 400 Volts dropping between the ground grids.
Will a rain storm cause one area of the earth between the grids to change more or less than another area?
Can you think of any condition which may make the soil resistance uneven?
If you are satisfied that the answers to these questions are safe, then consider running the systems on independent ground grids.
Personally I would interconnect the grids with redundant conductors well able to withstand the possible fault currents.
Remember that while fault currents may be in the thousands of amperes, lethal currents may be in the range of tens of milli-amperes.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[Bronzeado]

Thank you guys for your thoughts! We will join the two earth systems.

I will send a plan view so squeeky could give me some more specific tips. Thank you!

Slavag, all control cables are completely separated. All data from the pump station will be get via optical cables.

By the way, I just realize that all our substation earth systems are metallic connected via the transmission line ground wire!!!!!!!!

Best Regards,

Herivelto S. Bronzeado
Brasília, Brazil

http://www.linkedin.com/profile/view?id=46319837&trk=tab_pro

 

[karls2]

I am agree that:
With an additionnal copper cable between the 2 site (with the good size) and if you also connect the existing cables shield (for power cables) on both side you will increase split factor, and by this way it will begin less dependent of the real value of the earth resistance. (as we sometimes do for Wind turbine generator).
But if we connect both perfectly, do you think that the earth fault on the transformer on 230kV side (maybe the worst case ) can impact the pump station that is normally not impact if there isn't solid connection ?
How we can manage or take in account that ?

 

[7anoter4]

If we should take 230 kV Substation area as 100*100 m [10000 sqr.m] and pums' area as 30*20 m [600 sqr.m] and the Earth resistivity of 1000 ohm.m the approximate R1=4.43
ohm and R2=18 ohm.[Rg=ro/4*sqrt(pi()/Area) as per IEEE 80/2000].
Presumed short-circuit power of 230 kV System could be 20000 MVA-according to IEC 60076-5. [Zsys=230^2/20000=2.65 ohm].
Considering Xsys=Zsys and three times the grounding grid resistance we get 13.55 ohm total impedance and the phase-to-ground current 9.79 kA.Then GPR in 230 kV Substation will be 9.79*4.43=43.37 kV.
The potential decreases with the distance from the center point of the grounding area according to formula:
Vs=ro*Ig/(2*pi()*r)*arcsin(r/x) see: Energy Networks Association

http://www.energynetworks.org/modx/...tc/restricted/Specification_edits/S34/S34.pdf

Table 3 column 3.If r=sqrt(Area/pi())=56.4 m.and x=256.4 m Vs=6.13 kV.
That means the transferred potential will be 43.37-6.13=37.24 kV.
If only 1*75 sqr.mm copper bare conductor will be connected between both grids-neglecting the underground return current –considering only a single pump site- the current flowing towards this site will be approx.2.37 kA the current remaining to be dispersed through Substation site grounding grid will be 12.1-2.37 =9.7 and the equivalent grounding resistance of both sites parallel will be 3.56 ohm and GPR at that site will be 4.4*9.7=42.68 kV. The pump site GPR will get up to 2.37*18=42.66kV.A more accurate calculation will get approx. 206 V difference and Etouch50 for ro=1000 ohm.m =290 V[one second trip time].
For ro=100 ohm.m you’ll need probably 2*300 sqr.mm copper cable connection.Taking Carson return current into consideration the cable cross section area could be different.

 

[7anoter4]

Sorry. My mistake: no return current as the GPR at pump site is less than GPR in Substation if the short-circuit occurs in the Substation -all the time!

 

[karls2]

Thanks a lot for this complete answer, I need time to try to understand it!