Ground Grid Design

[tctctraining]

We are building a small substation ( including a transformer and a padmount switchgear) in adjacent to a wind farm substation and connecting the new substation ground grid to the existing grid. The existing grid has a measured grid resistance of .5 ohms. we are using Cymgrd to model our grounding grid and assumed a parallel resistance of .5 in the model which has dropped the GPR to below 5K.
However, step and touch potentials are still high. Do we need to model the existing ground grid as well as the new grid to accurately calculate step and touch potentials or are we missing a step?
Thanks for all responses.

 

[Marmite]

What is the voltage of the installation? A single line diagram would be helpful. You need to model the existing grid to be accurate as the area of the earth grid has a big impact on GPR. Don't forget Kirchoffs voltage law applies so you can't get a GPR more than the driving voltage for the earth fault current case you are looking at.
Regards
Marmite

 

[pwrtran]

To avoid over design of your ground grid it is reasonable to take the parallel Z into account. GPR depends on the fault current, ground resistance, system grounding method, the location of the fault and fault current splitting factor. Once your GPR is satisfied, other things you can do to control the step and touch potential are:
1) select different surface material, i.e. high resistivity material like clear crushed stone, asphalt, etc
2) use unequal spacing method to lay finer grid around fence and other metal structures

 

[7anoter4]

I agree with pwrtran. See: IEEE Std 665/95 for power station grounding. You may calculate, also, the common grounding grid resistance following
"DESIGN OF SWITCHYARD GROUNDING SYSTEMS USING MULTIPLE GRIDS-by Stephen W. Kercel" method.

 

[tctctraining]

GPR level is acceptable. Step and touch are not within the limits, so my question is whether modeling the existing substation ( which is taking a long time) and connecting it to the existing grid would change the S&T potentil values of the new subtation or as pwrtran indicated we just have to try different methods?
Thanks

 

[pwrtran]

GPR is not in the equation of step and touch potential calculation. Also see IEEE Std 80, you will find step and touch are only related to the body weight, fault duration and the surface resistivity, in other words, step and touch is nothing to do with the GPR, they largely deal with the body tolerance during a fault on a certain surface. You can have very high GPR but very low step potential difference as a bird sit a a wire, and very low touch potential difference as the person is insulated to any conductors. On the extreme side, you could have a person electrocuted just by touching 120V and you can think about the 120V is the result of GPR on a metal structure.

 

[jghrist]

If the new grid is close to the existing substation, you can't treat them as separate parallel paths to ground. There are mutual resistances to account for between the two grids. To accurately determine both GPR and touch voltages, model the interconnected grids as a single ground grid.

Does the existing substation have a HV/MV transformer, with a grounded MV winding that feeds the new transformer? If so, the source of MV ground fault current will be the existing substation transformer and it will not flow through the earth and contribute to GPR or touch voltages. A HV ground fault will produce GPR and touch voltages on both the existing substation and the new small substation.

 

[odlanor]

it is not easy to find a grid design data of an existing substation and integrate it with a grid design of a new substation. I never got !

 

[jghrist]

it is not easy to find a grid design data of an existing substation and integrate it with a grid design of a new substation. I never got !

That's true, but you can probably assume a perimeter ground around the existing station (confirm the the fence is grounded). This would be better than nothing.

 

[odlanor]

it is not enough to calculate mutual resistances neither acurately determine both GPR and touch voltages and model the interconnected grids as a single ground grid.

 

[karls2]

I am agree with jghrist, If you have the size of the existing substation and maybe the location of the building the value of 0.5 ohms for the existing grid you have a lot of informations.
With cymeground If I remember you can simulate the both cases connect or not and see the impact of your design.