Substation Ground Fault Currents on Sub-Transmission Lines With No Ground Return Path 1

[kmh1]

I am struggling on how to address system ground fault currents on load substations connected to 3-wire 44 kV sub-transmission lines with no ground return path (no neutral or overhead ground wire). The utility quotes a zero sequence ground fault current at a point on the line however I am assuming this value would need to be adjusted based on at least the local ground grid or ground electrode resistance to remote earth when calculating ground potential rise in a station. I have seen grounding reports in the past that quote a GPR that exceeds the system voltage which is obviously not physically possible. Does a current split exist even with no solid ground return path due to capacitive effects? My feeling is that grounding systems with a high resistance to remote earth act somewhat like ungrounded systems under fault conditions (local phase to ground voltage rises above nominal due to neutral point shift) as evidenced by the experience of one load customer that had lightning arresters operate during a local ground fault. Other incidents that have occurred on the utility system are failure to trip for a downed conductor on the line presumably due to a very high fault resistance back to the source. The area around the line in question is primarily sand and gravel over bedrock which are obviously difficult grounding conditions. The main objective is to recommend a minimum number of deep well grounding electrodes but yet ensure safe step and touch voltages in a load substation based on real values. Any thoughts on the subject would be greatly appreciated!

 

[jghrist]

The utility furnished fault current would be based on a zero-sequence impedance calculated with Carson's equations that does not explicitly include the grid resistance at either the load or source substation. You would not have any current split; all the return current would flow through the grid to earth.

IEEE Std 80 ¶15.4 suggests adding the grid resistance into the fault current sequence network equation if the resistance is particularly high. This is not normally necessary, but would prevent the case of GPR being calculated higher than the system voltage.

 

[kmh1]

Thank you jghrist, that is along the lines of what I was thinking but I was not able to locate an example or paper regarding this situation.