1. Hard pan clay that will not allow driving a ground rod deeper than 5 feet.
2. Solid rock that is not porous to water.
3. Even solid rock that is porous to water presents a problem.
4. Clay that is nonporous to water.
5. What is known in Pennsylvania as pyrite shale, which when exposed to air and water turns into clay and expands. There is a section of the PA Turnpike between exits 3 and 5 north of Pittsburgh that have fences in the rock cuts because the pyrite shale landslides every time it rains.
6. 100% bentonite clay backfill. This will expand so much as to exclude water and turn into an insulator. You can get a prepared backfill from Erico that has a small amount of bentonite clay that produces moderate compaction.
Worse case soil model for grounding
thread238-124763
Hi Rezaa -Welcome back to the forum.
Probably you meant soil model accuracy instead safety. If there is inaccuracy in the soil model, probably the natural way to balance this is overdesigning the ground grid to be in the safe side.
Soil for grounding purposes could be modeled as a single uniform layer, two, three or "n" layers. However, more than two soil seldom provide significant better results do to non homogeneous nature of the soil.
In any case, soil model is an approximation of actual soil condition and can be used to perform a safety grounding design.
Here are some quotes from the IEEE Std 80 that could be applicable to your question:
… Soil resistivity varies laterally and with respect to depth, depending on the soil stratification. Seasonal variations may occur in soil resistivity due to varying weather…
.. The most commonly used soil resistivity models are the uniform soil model and the two-layer soil model. Two-layer soil models are often a good approximation of many soil structures while multilayer soil models may be used for more complex soil conditions.
….. While the most accurate representation of a grounding system should certainly be based on the actual variations of soil resistivity present at the substation site, it will rarely be economically justifiable or technically feasible to model all these variations. However, in most cases, the representation of a ground electrode based on an equivalent two-layer earth model is sufficient for designing a safe grounding system.
I am considering soil resistivity cases due to seasonal variations. For example in our case, the worse case soil model is considered to occour in spring when the top crushed rocl=k layer is wet from melted snow while the ground grid remains in frozen soil.
The surfacing crushed rock may be considered as an insulator with wet resistivity around 2000 to 3000 Ohm-m.
The soil layers that should be considered are the one under the crushed rock.
Double check if at the depth of the ground grid and ground rods are in the frozen leyer since 10 ft (~3 m) is normally below the frozen depth in most cases ouside permafrost regions.
Probably will be a good idea to check with a geotechnical consultant with knowledge and experience in this area to verify the worst case condition.
As general sample, check in the enclose figure will provide you ideas of soil measurement, model and seasonal variation of resistivity.
I don't think you can generalize on what would be the worst case. You have to analyze the possible soil models. The lowest grid resistance is not necessarily the best in terms of step- and touch-voltages because with higher resistance, more of the fault current will flow through the earth and less through shield wires.
Your frozen depth of 0.5-1.2 meters with a thawed top layer of 0.2 meters implies at least a 3-layer model, counting the soil below the frozen layer. You could probably ignore the top thawed layer. How does the top layer get thawed?
The top layer is wet from melted snow. We are assuming a 3-layer model. The worse case soil model is considered to occour in spring when the top crushed rock layer is wet from melted snow while the ground grid remains in frozen soil( surronded with higher soil resistivity). For the sake of the simulation (using CDEGS) we can ignore the top layer but for safety limit calculation we can not ignore the top layer.
If the wet crushed rock layer is equal or less resistivity than the frozen layer, you cannot consider it as an insulating layer in calculating allowable step- and touch-voltages.
See thread238-68202 for more discussion of frozen soil.
In general, the GPR of an extensive grounding system is influenced more by the deeper layer soil resistivity. For Simulation we can ignore the seasonal varation of the surface layer soil resistivity and consider the reresentatives of only deeper layers(using either summer/fall /winter/spring soil model a 2-layer model).For safety criteria (calculating safety limit levels for touch and step voltages), the worse case soil model is considered to occour in spring when the top layer is a thawed layer resulted from melted snow(low resistivity)with a small thickness that follow with another soil model(high resistivity)with a small thickness as well (Using 2-layer soil model).
from melted snow and follow with a thawed layer while the ground grid remains in frozen soil.
