grounding of MV distribution systems (4-wire multi-grounded neutral

[magoo2]

This refers to the grounding requirements for medium voltage distribution systems outside the substation, that is, along the line. These systems typically operate at 12.47 kV, 24.94 kV or 34.5 kV, although other voltages are still around.

In providing grounds for equipment (distribution transformers, reclosers, capacitor banks, arresters, etc.), our company has a target of 25 ohms. In our area, this usually requires driving 2 - 6 or 8 ft rods. These are stacked with a coupling between them.

In some cases, we can't even reach 50 or 100 ohms with as much as 60 ft of driven rod! This leads me to my questions:

What is a reasonable target for grounding impedance?

Can you relate any experience-base for this selection or is it an opinion?

Does a high grounding impedance have an adverse effect on the performance of the surge arrester? If so, at what value?

Thanks in advance.

 

[stevenal]

NESC does not require a target impedance for individual ground locations as long the other conditions are met. The multiplicity of the ground locations is considered to make the overall impedance to remote earth low enough. Arresters are more effective with a low local impedance. O.C. Seevers responded to lightning problems by simply driving more rods. When no more problems were observed at a location, he moved on. I don't believe he used any particular target impedance.

 

[magoo2]

I'm familiar with what the NESC requires, but my question relates to the performance aspects of proper grounding.

Also, I don't hold OC Seevers' anecdotes as experimental evidence. He provides more of a 10,000-ft view of grounding. There's not much details and data to lead one to adopt his suggestions.

 

[wareagle]

What does the 25 ohm target do for you? Are you trying to
divert lightning?

 

[cuky2000]

There is not possible to provide specific answers to your questions because the nature of targeting resistance value is not a simple deterministic phenomenon.

There is not uncommon for utility to target 25 Ohms and 4 ground connections per miles since low overall grounding help to reduce overvoltage, stray voltage and failure rated of equipment.

It is important to notice that the number of ground connection per miles plays an important role as the footing resistance in the reduction of overvoltage. Study on distribution systems shows that 25 Ohms footing resistance with 8 ground connection per miles limit the overvoltage at the end of line near to 1.3 pu.

 

[Electic]

25 ohms is a good target but sometimes not attainable.


Utility distribution systems often include what is called a common neutral, that is a neutral connected to multiple grounds. A local utility by experimenting determined the common neutral to be low impedance ground path, drawing something like 80% of the ground fault current with earth in parallel drawing the remaining 20% *. (This finding is consistent with magnetic coupling theory and also supports NEC requirements that grounding conductors be with the supply conductors)

If your lightning arrestors are tied to the common neutral, all will work well.

(* this work was published locally by IEEE in the NorthWest United States, not sure if it made international or even national publication)

 

[magoo2]

The IEEE reference sounds interesting. Do you have a title or date? I've got a number of contacts in the Northwest U.S. area, so I might be able to see what was written.

To respond to another comment, the main interest with the ohmic value of ground resistance is in lightning performance. The NESC dictates the minimum 4 grounds per mile and the 25-ohm reference only applies to an isolated ground, so a multi-grounded neutral system doesn't have a specific ground resistance target for each ground rod location.

Thanks.

 

[jghrist]

IEEE Std 1410-1997, IEEE guide for improving the lightning performance of electric power overhead distribution lines, discusses measures for improving the lightning protection performance of schemes applied to overhead power distribution. A method is given to estimate the performance with different ground resistance values.

As cuky stated, it is not possible to give a specific answer to "how good is good enough?" There is no value where the probability of a lightning outage is zero. You can estimate the performance at different levels of grounding and make an engineering judgement based on the cost of driving more rods and the benefits achieved.

 

[wiretwister]

Mr. Magoo2
Go to this site for more info on ground testing.

http://ecmweb.com/grounding/electric_ground_testing_techniques/

wt

 

[cuky2000]

The 25 ohms referred in the NEC, NESC and often targeting values used by many utilities may be controversial maximum limit. The origin of this values may be traced many years back went the electrical power system were still in development. Below are a few interesting remarks:


? Some people believe that 25 Ohm was adopted since is close to an average soil resistivity for most region of the US.

? In 1933 the NEC introduced 25 ohms as the maximum value of resistance for buried or driven ground electrodes.

http://www.nfpa.org/itemDetail.asp?...Technical Milestones in the NEC&cookie_test=1

? Possible influence from the telegraph industry that use Varley = 25 Ohm as resistance unit.

http://www.sizes.com/units/varley_unit.htm

 

[stevenal]

Gotta disagree with Electic. Tying arresters to the neutral without a good local connection to ground is worthless. The high frequency components of a lightning strike see a span of neutral conductor as a very high impedance path.

On reread of Seevers, he did shoot for a target of 10 ohms, but quit after driving the third rod. 20 ohms and most problems went away. Magoo, your OP asked for experience-base, not experiment, that's why I referred you to Seevers.

 

[apowerengr]

I agree with stenvenal, you need as good a local ground connection as practicable. For overhead distribution systems some arrester suppliers recommend secondary arresters as well as primary arresters especially when the local grounding conditions are not good. I had been told by an old timer (not confirmed) that the 25 ohms was based in part on anticipated typical lightning stroke currents of 3kA and typical systems of 75kV BIL or higher to provide some margins. Magoo2, remember that if you're testing ground electrode resistance during installation, the numbers will improve 20-50% after the earth around the electrode settles in (3 months or so) based on my field experiences. High ground impedance won't affect the arrester performance directly, but remember the arrester rating is the voltage which will develop across its terminals at the rated discharge current. If you have to add a 50-200kV increase in ground potential rise to the discharge voltage of the arrester the series combination may be too large for your system insulation BIL and make your overvoltage protective system ineffective. You can take your system BIL, subtract the arrester discharge voltage and see what margin remains. Assume a lightning surge current, divide it by the remaining voltage margin and that simple calculation will tell you what your target local ground resistance needs to be. If you have lightning detection network data and probability of hits you can refine whether you need to increase your insulation levels in areas where you can't lower ground resistance, use better arresters in those areas or both.

 

[magoo2]

I'm not sure what problems went away (Seevers) when you got below 20 ohms. Was he dealing with equipment failures, arrester failures or what. In my read of Seevers, I found that he was so hung up with certain parts of a cow's anatomy that I couldn't take him seriously. I'd find it more believable if he'd have concluded for example that you notice a noticable % reduction in equipment problems (or fill in the blanks along similar lines) when the grounds were 20 ohms or less.

His book dealt with stray voltage. Did his stray voltage problems go away with this grounding impedance range? I would doubt that since there's many other factors to consider.

In some IEEE references, I found the stability of the ground resistance is dependent on the depth of the water table. In many cases, you had to go to a depth of 30 feet or so to be sure you have a stable ground resistance over the year.

We use 8-ft, 5/8-inch diameter ground rods. Applying Dwights formula with a rho of 100 ohm-m, this shows that you need 2 stacked ground rods to achieve 20 ohms. This seems reasonable. By contrast, it would take about 5 ground rods to get below 10 ohms.

In certain portions of our system, rho is closer to 300 ohm-m. In these cases, you need 5 ground rods to meet 20 ohms. 10 rods would get you down to 15 ohms, so 10-ohms would be out of the question with the higher rho conditions.

Based on the stability combined with the range of rho values, we usually employ 4 or 5 ground rods.

I'm not sure I agree with apowerengr in the issue about round potential rise and arrester protective characteristic. The ground potential rise affects the neutral. If the neutral potential increases due to GPR, the protection across the transformer winding shouldn't be affected. Am I missing something?

I agree that the grounding improves as the soil becomes more compacted. I know of a consultant that used bentonite in their ground rod installations. He claims that the bentonite clings to the rod and makes the initial as well as long term reading lower.

I also have used the AEMC/Yokagowa clamp on ground resistance tester. Not only does it permit faster checks with utility grounds, it also gives lower readings that the crank-type meggars. I think the frequency of the clamp-on tester is around 5 kHz, so you're in effect comparing a high frequency reading (clamp-on) to a dc reading (crank-type).

I appreciate all the good comments.

 

[dpc]

Bentonite and other soil amendments have been used for many years to lower resistivity at ground rod/soil interface.

 

[stevenal]

Wrong book. The one I'm referring to is titled Power Systems Handbook. The problems he speaks of are "arrester failure concurrent with fuse blowing and/or transformer failure" The fill in the blank figure is 4/5ths. Lightning damage expense was reduced to 1/5th the amount spent prior to the grounding program.

 

[apowerengr]

Magoo2, the protection afforded by the arrester to the primary winding may not be affected, but the protection provided to the system including pin/post insulators and the secondary windings may be compromised by high GPR.

 

[magoo2]

Hey dpc, can you tell me who uses bentonite in their grounding?

Stevenal, I glad you clarified the book. From what you described as the problem he solved, it sounds among other things like the damage due to a direct hit above the arrester capability. Did he offer any conclusions why the lower grounding resistance lowered the failures?

One of the problems often overlooked is that lightning activity can vary considerably from one year to the next. If he initiated some improvement program, the improved performance could be due to decreased lightning activity in that area the following year. As an example, Tom Short in his Electric Power Distribution Handbook relates an example of line protection using arresters at LILCo. As I recall, the control group, the one in which no improvements were made, showed better performance than the ones where they installed arresters. The problem is that he based his evaluation on a single year.

We've found similar conclusions. With the availability of data on flash density today, you can see why a multi-year evaluation is necessary to 1) decide what areas to work on and 2) to properly evaluate the results. A single year or snapshot evaluation can be very misleading.

I'm still having trouble understanding how GPR is a real issue here.

 

[stevenal]

He attributes the failures to lightning induced surges on services and secondaries with no mention of the possible use of secondary arresters as a cure. Path is to the pole ground. IR drop across the grounding resistance causes potential rise on the grounded end of the primary arrester, putting it at a higher potential than the phase end of the arrester. He then claims that arresters are not designed to work in reverse to conduct under this situation resulting in overvoltage on arrester and primary winding. (Not sure I buy that claim). The grounding program he wrote of spanned eight years.

 

[Electic]

Clarification: The common neutral described in my first post is tied to ground at each pole. It was not suggested that pole grounds be eliminated.

The utility I previously worked for put lightning arrestors only at underground termination poles and line deadends. Often times the line deadend arrestors would be blown (pieces of porcelain around the base of the pole) by distant lightning strikes, that had travelled down the line. For medium voltage power distribution systems this was ample evidence the system worked sufficiently (though we didn't like the durability of the arrestors) despite high measured individual pole ground resistance.

I would expect such neutral to be near zero volts prior to a lightning strike, and though there may be substantial potential rise during a strike, the impedance was low enough to allow the surge arrestors to valve most of the energy away from the distribution line which was the primary objective.

 

[dpc]

magoo,

Check with Erico for bentonite (GEM)users:

http://www.erico.com/products/GEM.asp

In a past life, we specified for some utility applications in Southeast Alaska, where they are sitting on big rocks, basically.