Small Substation Grounding 2

[RSChinn]

We are very successful using WinIGS to design our substation grounding systems to comply with IEEE80. However, on small substations (80'X 80' square) we cannot get the touch and step potentials to be lower than the allowable. Small areas tend to have this problem regardless of how much additional horizontal copper or ground rods we install. Any solutions to getting small substations to meet IEEE80? Worst case we think would be to install an expensive grounding well. Other solutions?

 

[7anoter4]

It is a big problem, indeed. If the soil resistivity is between 200-400 ohm.m [let's say the upper layer-10-14 ft height- would be 400 and the lower layer 200 ohm.m ] , the I"k3=20 KA and the return current through static wire or cable shield would be 25-40% you could get good touch and step potential if the horizontal electrode dia could be increased up to 8" [using 8" galvanized steel or aluminum conduits or by means of low resistance material as bentonite or Erico GEM compound see:

http://www.erico.com/public/library/fep/LT0359.pdf

A lot of vertical electrodes [90-100] of 3/4" dia and 50 ft long you have to drive all around the substation area.The distance between the horizontal electrodes would 15-16 ft.
Of course crushed stone of 3-4" height is required.
Also the Grounding Grid has to be extended 3 ft outside the fence [and the fence has to be grounded if you cannot isolate it].
If the soil resistivity is high the problem is more complicate, of course.

 

[SubstationDesigner]

Dear RSChinn,

In case you have buildings, try to using structural electrodes (dedicated steel mesh 5mx5m in reinforced concrete floor, wire wrapped to re-bars for electrical continuity & finally brought out to be connected to your earthing system). See pic on

http://files.engineering.com/getfil...e4d4-4786-83bb-f994a69decd5&file=P9220025.JPG

This should provide better control of touch & step voltages by making the floor more equipotential.

Near corners of the fence you could try to install pre-made Cu meshes (1.5m x 1.5m, Furse supplies something similar) at shallow depth.

 

[jghrist]

Are you taking current splits with shield wires and neutrals into consideration? Often with small substations, the higher resistance of the grid forces more of the return current through shield wires and neutrals, making step- and touch-voltages lower.

 

[RSChinn]

Reply to ALL- we are in the Seattle area where the shield wire is not typically used due to low isochronic levels. Also, we do not have counterpoise wiring so that the return current does not have a current division through those two elements. We had actual measurements of the substation grid of .35 ohms with the overhead feeder neutrals disconnected and .39 ohms with these neutrals connected (our feeder neutrals are grounded at every two poles). So the current split due to the neutral contributions is only about 11%. The 8” piping seems very expensive… maybe a grounding well is cheaper. Yes/No?…I will check it out. In the past, we have installed pre-fabbed copper mesh in areas most traveled within the substation when we could not pass the Touch and step criteria… we may do it again. Thanks for all the input. (By the way, although the empirical Rg value is .39, the WinIGS calculated value was 1.19 ohms and failed the Touch and Step. I will now inject the actual values in and hopefully get passing results.) If there are further comments, please reply. Thanks to ALL.

 

[TurbineGen]

Just some general ways I have seen used to help with this issue:

-Use breakers with shorter clearing times
-Import soil of lower resistivitity
-Drive ground rods deeper (I heard of a company on this forum claiming they could get 100' in bed rock)
-Use Erico GEM ground enhancing material

Is it allowable or possible to extend your grounding system beyond the fence to give you a greater area?

Just some ideas.

------------------------------------------------------------------------
If it is broken, fix it. If it isn't broken, I'll soon fix that.

 

[jghrist]

(By the way, although the empirical Rg value is .39, the WinIGS calculated value was 1.19 ohms and failed the Touch and Step. I will now inject the actual values in and hopefully get passing results.)

How will you "inject the actual values in"? There are two possible problems that could explain the large difference between measured and actual grid resistance. First, the feeder ground connections may be influencing the grid resistance measurement even if they are not connected. Second, the soil model may be incorrect because of faulty measurement of the soil resistivity or use of homogenous model where layered model is correct.

In the first instance, the actual resistance may be larger and there may be more current split. In the latter instance, how do you change the soil model to correct for the resistance measurement? A change in overall resistivity may not reflect the actual case. If the soil actually has a lower layer with lower resistivity, lowering the resisitivity of all of the soil may give an unconservative calculation of touch-voltages.

 

[RSChinn]

To jghrist: Thanks for the interest and I absolutely appreciate your help. I divided all of the soil measurements by 3 (this was arbitrary; need your smarts on what the pitfalls are) let WinIGS calculate what the new Rg is. WinIGS now recalculated an Rg of .367 ohms which was between our actual values of .34 ohms with the neutrals connected and .39 without the neutrals connected. Then we let WinIGS do the follow up re-calculations on Touch and Step and we appear pass. The reasoning behind "faking" the soil sampling was because there are only input two elements (correct??) that affect the Rg: 1) the soil resistivity and 2) the physical configuration of the grid. Since we drew the grid configuration into WinIGS, the only other thing left to change was the soil resistivity.

 

[jghrist]

Unless the soil is homogenous, there are an infinite number of input elements that can change the calculated resistance to match the measured value. I realize that this isn't very helpful, and your approach is reasonable and likely the only practical approach.

How did you measure the grid resistance with the neutrals connected? There are ways to do this, but it will be difficult to get outside the zone of influence of the ground system with a simple fall-of-potential measurement.

 

[rcw retired EE]

RSChin - At Fort Lewis Army base, just south of Seattle, the soil is mostly glacial till -lots of gravel with high resitivity. (Broken 13.8 kV lines laying on the ground did not draw enough current to trip the ground relaying.) To get low resistance ground rods at some munitions areas, the contractor used a boring machine, similar to a well digger. The drill head had a hole in the middle that was used to guide the rod into the bored hole as the bit was being retrieved. With proper set up was fairly quick work to bore a 50' hole, insert the ground rod and pump a bentonite slurry into the hole. The rods tested at 5 ohms each. Normally they would be > 50-100 ohms in that soil.

Maybe adding 2-4 deep rods will drop your resistance.

 

[RSChinn]

To jghrist: The existing substation is 40 years old has homes and businesses built around it. The engineers that actually took the soil resistance measurements did it in a park across the street from the sub since they could not do it in the sub due to the interence of the city infrastructure (streets, parking lots, private property). They did this hoping their sample would be representative of the sub's soil resistance... so there is some chance of inaccuracies here. When they did the actual Rg measurement after the installation of a new grid (the old substation had none but a few horizontal conductors for a grid), they only route they could do (to approach 5 times the sub's diagonal length) was go down the adjacent street. The street has distribution poles on the opposite side of the side where they took measurements using the fall-of-potential method. They did this with and without the neutrals connected. The neutrals on the distribution poles are grounded at every other pole with pole span about 200 feet. We have this situation at many of our substations and struggle with credible ways to analyse and document concurrence to IEEE80. If totally in doubt we have the grounding well as a last resort.