We had field measurements done by a company out of Ontario that measured, GPR, Grid R to remote earth, current split, Step & Touch Potentials using current injection. The instruments were all tied to a data recorder for near simultaneous measurements. They used a 55Hz generator to inject current and filtered all the results to eliminate any 60 Hz influence.
The substation's incoming line was not yet energized and could be safely grounded about 1.5 km away. Fault current was injected A phase to ground and current split measured in the OHGW, phase wires and the local distribution line neutral wires. Current was 10-15A. Step and touch voltages were measured at typical locations.
The measured split came out close to our calculated split. Step and touch came out better than predicted by CDEG's software. Test was expensive. The company claimed they do similar testing on energized substations and lines up to 230 kV. They developed the system when they were a part of Hydro One, the Ontario utility, and then became privatized. (I'm trying to find their name).
The current split depends on the mutual impedance between ground, faulted phase conductor, and shield wire all the way back to the source. I would think that to get an accurate measurement of the current split, you would have to inject the current from the source substation through the transmission line. Injecting current 1.5 km away may result in a higher percentage of the return current flowing through the shield wire (and less through the earth) than would be the case for an actual fault.
rcwilson - the company you're referring to is probably Kinetrics. That used to be the Ontario Hydro Research and Development Division. That was sold off during the break-up of Ontario Hydro.
Marks1080- thanks- It was Kinectrics that tested that project.
jghrist- I haven't looked at the equations recently and could be wrong thinking that on a longer line the impedance of the shield wire to the faulted phase would result in more current in the shield wire and less in the ground. The shield wire is closer to the phase conductor than the ground return path so it should have less impedance and more return current. If that is correct, the test result from the shorter test location yields a conservative split factor with more current into the earth, causing a higher GPR. However, could the additional grounded structures in the longer line provide more paths for the ground return current to get up onto the shield wire and reduce the effective ground return impedance? That might create more ground return current and change the split.
On that project, the tie into the utility system could not be made until the inspector was convinced that the substation area and the adjacent plant were safe from step and touch issues. The plant was still under construction without the final paving and gravel in place. Our CDEG's models showed no hazards, but the inspector wanted actual measurements and was satisfied with Kinectrics' test results. BTW, the inspector was very knowledgeable on substation grounding and was completing a thesis on grounding calculations/simulations for an advanced EE degree.
I looked at a recent design and you appear to be correct. Cutting the transmission line distances in half resulted in the split factor increasing from 7.5% to 29.6%.
