Fall-of-Potential(FOP) measurments should be compared to a simulated(using a software)FOP. For simulation of FOP we need to define a soil resistivity model. Usually 4 soil measurment traverses are used. 2 short traverses( representation for shallow layers soil model)within the station and two long traverses ( representation for deep layers soil model)outside the station. We can combine all 4 traverses points to develop a single three layer soil model using a software like RESAP modele of CDEGS software( RESAP is normally used to determine equivalent earth structure models based on measured soil resistivity data) or any other software. We can use this combined soil model to simulate FOP. My question is what would be the base
for modifying the comined soil model data to obtain agreement between Fall-of-Potential(FOP)measuements and simulated FOP?
I'm not sure what you mean by traverses. We would normally measure soil resistivities along one or more lines, each with several probe spacings. These would go into RESAP to determine the optimum soil model, homogenous, 2-layer, or 3-layer (or more, but this would be rare).
If you simulate the FOP based on the soil model and grid design, and then measure something different, it isn't clear what to do. It may mean that the grid was not constructed according to the design. It may mean that the original resistivity measurements were not correct or complete. How would you vary the soil model? Change the layer resistivities? Change the layer depths? Some combination?
I have done a lot of FOP testing and I now beleive that most of the test results were bogus. The FOP method assumes that the ground is homogenous, at least horizontally. Vertical differences do not have as much of an effect. This means that the area around your grid must not have anything else metallic in it, like fences, pipelines, creeks, foundations with rebar, overhead pole lines with grounded anchors or grounds, underground duct banks, telephone lines, etc. Any of these objects, especially the power lines, will pick up your test current and return it to the grid so your potential probe doesn't read all of the current. The results will be skewed.
To get a good FOP test, your remote current probe must be placed a distance at least 5 times the longest diagonal of your grid. This means fa 150'x 150' grid, the C1 probe is at least 1060 feet away. If you are closer, you won't get the flat area on our curve to indicate the actual impedance.
Most substations and grids are not in a virgin area They are surrounded by these other structures that can affect the results. So trying to modify your soil model to match a field test may be fun academically, but you're just comparing an approximation from a calculation to questionable test result. Which one is more accurate?? I no longer trust either.
It might be useful if you could run several tests on a virgin site and try to adjust the model to match.
We now verify our designs by impressing a voltage on the grid and measuring the step & touch potentials. I don't worry about trying to optimize the grid to save copper because by the time I bond all of the substation equipment, I already have more copper than the program claims I need.
The soil resistivities were measured along 4 traverses (lines), each with several probe spacings of course. Then we combined all those 4 data and treat them as one data by eliminating bad data. For example the shallow data associated with long lines are not a good representation for the area within the station(due to localized site improvement).
The curve we get from FOP measurements is smooth without any dip. This means there is no indication of metallic objects such as fences, pipelines, creeks,....
The FOP that uses the soil model can be resulted in lower or hihger grid impedance. So a realastic soil model is imporant in the calculation of the grid resistance.
