Distribution substation earthing 1

[TimDta]

11/.415kV Distribution substation's(32m long by 15 m wide) propective earth fault current is 27kA but there is limited real estate and the native soil resistivity is very high(~800 ohm-m).There are two incoming 11kV cables to this substation.Running CDEGS software with 13.5kA(assuming 50% current being shared by the distribution substation's earth grid while the remaining 50% by other metallic return paths). With the above configuration,there are hazardous touch potential at the edge of the earth grid(even with grading ring)and hazardous step potential outside the substation. So now accurate % of 27kA current split has to be calculated. First pass will be to model in CDEGS' SPLIT/Right of way module(which I haven't used before;so detail steps or sample cals/steps will be helpful).Without much data coming from client, could you please give estimated screen resistance in ohm per meter(is 0ohm reactance reasonable?)of 11kV XLPE cable screen and reasonable estimates of the 2 remote end's grid resistance in ohms to which the 2 different 11kV cables will be connected? If still safe potentials can't be achieved, then the next pass will be to model 2 separate bondings(using 2*120sqmm bare Cu cable for each bonding)of this 11/.415kV distribution substation's earth grid to the 2 remote end's grid,whose resistance still needs to be estimated at this stage.Since I haven't done such calculations before,any information will be helpful.Any tips on the 1st and 2nd proposed passes,comments,typical input data for CDEGS calculations or manual calculations will be appreciated.Thanks.

 

[Marmite]

It's not clear whether your prospective earth fault current relates to the 11kV or LV network, but it's very high for 11kV way beyond the rating of most 11kV distribution switchgear, so I presume it's the LV figure. In this case regardless of what you think the fault level is, it is physically impossible to get a rise of earth potential more than 240V for a fault on the LV side. Only faults on the 11kV system can cause rise of earth potential at your distribution substation because the source is remote. Most current will flow back along the cable screens to the source, some via the earth. A local LV fault will cause fault current to return via solid metal to the transformer neutral. This will not raise the local earth potential of the substation.
Regards
Marmite

 

[jghrist]

You might be able to find out the utility's standards for maximum substation grid resistance and use that if actual resistance cannot be determined. Only one of the remote ends will be a source, but some current will flow to each. You'll also need to estimate the distance between grounds on the cable screen (if it is grounded between the substation and the remote ends) and the resistance of the grounds. Screen resistance is important and can vary a lot depending on construction. A concentric neutral URD cable will have much lower resistance than a power cable with thin tape shield or drain wires. You should be able to get screen resistance from the cable manufacture.

I agree with Marmite - 27 kA is very high for 11 kV Ø-ground fault current.

A high soil resisitivity will generally result in much less than 50% of the current returning in the earth.

I haven't uses SPLIT, but SES technical support is good; they should be able to help.

 

[Kiribanda]

TimDta,
Since your soil resistivity is high you are getting very high touch & step potentials for the 11 kV return earth fault current.As mentioned by jghrist,the current through soil will be less than 50% due to high soil resistivity too.You can also use the split factors given in IEEE-80 graphs if applicable to your sub.But in my opinion the best possible solution is to BOND each & every 11 kV current carrying equipment using 1x95 sq. mm insulated green earthing cables to the two 2x120 sq. mm bonding cables (per your suggestion) running from your sub to remote end's grid with a good coordinated earth fault protection for 11 kV network.It will ensure a very low zero seq impedance for the return earth fault current enabling the earth fault protection to work faster.
Any suggestions are welcome.

 

[AusLee]

There should be a law prohibiting the construction of substations over high resistive soil.

I think 10.1kA is a more appropriate value than 27kA. Use 13.1kA to be on the safe side if you think it is justified. This will be the value limited by the upstream protection.

There may be nothing you can do at the edge of the property, use Argon software to calculate the probability of a fatality at this location, usually less than 1 per million is socially acceptable as it becomes more likely to get struck by something else ... the only number you will need is the estimated number of faults per year. Look at pages 96-97 of the following document:

http://www.ena.asn.au/udocs/2010/07/EG-0-Power-System-Earthing-Guide-for-website.pdf

 

[ScottyUK]

That sounds like a practical law. Maybe most of the middle east should be prevented from building substations?

I'm curious how you arrive at such precise figures - why 10.1kA and not 10.2kA? or 13.1kA instead of 13kA. Not saying it's wrong, just interested in how you reach those numbers.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[jghrist]

I think 10.1kA is a more appropriate value than 27kA. Use 13.1kA to be on the safe side if you think it is justified. This will be the value limited by the upstream protection.

How do you know that the upstream protection limits the fault current? Usually, the fault current at 11 kV is interrupted, but not limited. Fast interruption will increase the allowable step- and touch-voltages, but will not reduce the values during a fault.

The linked ENA guide for determining the probability of a fatality is interesting, I've never seen this analysis before. I don't know how practical it is. The one per million acceptable rate is arbitrary. Any estimate of the probability of someone touching an object during a fault is arbitrary.

 

[TimDta]

1)Just to clarify,11kV earth fault value given by client was 27kA. Even though this value is not realistic,did a quick
manual calculation of earth fault current split to earth grid and cable screen.Need help to verify whether the folowing derivation of earth grid fault current of 1700A is correct and realistic.

11kV cable is 400sqmm Cu Olex(65sqmm being the cross sectional area of the screen).
1a)11kV cable screen(235 m long)resitance for cable to sub 856: 115 milli ohms
1b) 11kV cable screen(140 m long)resitance for cable to sub 857: 68.6 milli ohms

Earth resistance(inclusive of soil and Cu earth grid)of modelled substation's earth grid is say 3.5 ohms.

Current thru earthgrid with 115mohm screen =27000/((3.5/.115)+1) which is 858.9212 Amp
Current thru earthgrid with 68.6mohm screen=27000/((3.5/.068)+1) which is 519.0271 Amp

For a 27kA fault level,worst case(assuming 68.6mohm screen is disconnected and 115mohm screen is connected)current flowing through 861’s earth grid is 859 A.
Allow a 50% Safety margin and the earth grid fault current is 1700A.

2)Also attached is "FCDIST Configuration Picture of Substation3.5 Ohm under study.bmp" but the following FCDIST output is not looking alright(Total Fault Current...........: 13000. Amps,Total Neutral Current.........: 13000. Amps, Total Earth Current...........: 0.30082E-06 Amps).Probably some mistakes(which I can't figure out) have been made in the inputs to FCDIST module.

TERMINAL GROUND SYSTEM (Magn./Angle)

Term: 1 Total Earth Current...: 0.13010E-05 Amps / -0.91654E-02 deg.
Earth Potential Rise..: 0.13010E-05 Volts / -0.91654E-02 deg.
Term: 2 Total Earth Current...: 0.65049E-06 Amps / -0.91654E-02 deg.
Earth Potential Rise..: 0.13010E-05 Volts / -0.91654E-02 deg.
Term: 3 Total Earth Current...: 0.13010E-05 Amps / -0.91859E-02 deg.
Earth Potential Rise..: 0.13010E-05 Volts / -0.91859E-02 deg.
Term: 4 Total Earth Current...: 0.13010E-05 Amps / -0.91859E-02 deg.
Earth Potential Rise..: 0.13010E-05 Volts / -0.91859E-02 deg.
Term: 5 Total Earth Current...: 0.93638E-07 Amps / 90.000 deg.
Earth Potential Rise..: 0.13242E-12 Volts / 135.00 deg.
Average Resistivity...........: 350.00 Ohm-meters
Grid Impedance................: 4.3248 +j 0.0000 Ohms
< Magnitude / Angle >
Total Fault Current...........: 13000. Amps / 0.0000 degrees
Total Neutral Current.........: 13000. Amps / 0.0000 degrees
Total Earth Current...........: 0.30082E-06 Amps / -0.91654E-02 degrees
Ground Potential Rise (GPR)...: 0.13010E-05 Volts / -0.91654E-02 degrees

 

[TimDta]

Thanks to all of you who have replied to this post.
A few more questions:
A)Wondering how FCDIST's output shows Total Neutral Current.........: 13000. Amps when no info related to neutral was entered in FCDIST?[13000 is the 11kV earth fault current and there is no neutral conductor path(11kV ie primary side of 2000kVA transformer is delta connected).]

B)Is FCDIST suitable for non-transmission line networks or SPLITS/other method/hand calculation is better for the attached

http://files.engineering.com/getfile.aspx?folder=46e824e5-283c-4ae6-b928-14

configuration ?

C)To mitigate touch voltage issues related to metallic fence/structure,1 meter wide bitumen strip is often used. Is 1 meter adequate or has anyone encountered a slightly wider e.g. 1.5 meter wide bitumen strip in few installations?

D)Secondary side of this 2000kVA transformer is star connected with solidly grounded neutral. For an earth fault on the 415V side,I have assumed zero 1ph-E fault current will enter the distribution substation earthgrid as almost all 1ph-E fault current will return via neutral connected to transformer's star point. Wondering how accurate is this assumption,what small %(eg 5 to 10%) might still flow to earth grid,and any exceptions/variations to this thumb rule? Kind regards. Thanks in advance.

 

[jghrist]

Why are there five terminals? You only mentioned three, the station being analyzed and stations at the ends of the two incoming lines.

There's something wrong with the FCDIST output. There should be earth current. Do you have ground connections specified at the remote terminals?

I haven't used SPLITS, but my understanding is that it is more complete than FCDIST. FCDIST should be adequate, however, if the inputs are correct.

We normally extend crushed rock surfacing 5 feet outside the station. The bitumen needs to be extended far enough so than anyone that can touch the fence will be standing on the bitumen.

You don't have to consider secondary faults because the source is local.

 

[TimDta]

Regarding Jghrist's question "Why are there five terminals? You only mentioned three, the station being analyzed and stations at the ends of the two incoming lines.",terminal 1 & 2 are 2 other earth grids to which station being analyzed is proposed to be bonded,
terminal 3 & 4 are indeed the stations at the ends of the two incoming lines, while terminal 5 is the 13kA earth fault current source only as shown in the above attached file

http://files.engineering.com/getfile.aspx?folder=46e824e5-283c-4ae6-b928-14.

Regarding "Do you have ground connections specified at the remote terminals?",for each remote terminal,I had keyed in FCDIST:
a)in the 'ground impedance section', ground impedance(ohms) values of Re 1 ohm + jXe 0 ohm.
b)in the 'defined circuit data,selected 'impedance method' and in the 'series impedance section' keyed in ohmic value of bonding conductor for termnial 1 and 2, cable screen for termnial 3 and 4.

Wondering what further analysis or mitigations are typically practiced in such a situation, because of hazardous transfer potential issues arising from connecting the station being analyzed(whose earth reistance is 3.5 ohms)to 4 remote stations?

I am planning to use other CDEGS engineering modules and would like to post in this thread if there are any issues.Cheers.

 

[jghrist]

The links in your second 10 Jan post and 16 Jan post come up blank. From the sketch in FCDIST_Configuration_Picture_of_Substation3.5_Ohm_under_study.bmp, it appears that the only fault current source is local to the substation. No impedance is shown between the E/F source and the 3.5 ohm Substation. If this is the case, there won't be any GPR, touch- or step-voltages.

If Building 1 and the Treatment Center are close enough to the substation to be within in the zone of influence of the ground grid, I don't think FCDIST would be an appropriate program to evaluate the currents going to those ground systems. Maybe SPLITS could be used, if there is a remote ground fault source.