I KNOW THERE ARE AT LEAST A FEW OF YOU WHO WILL TAKE OFFENSE when I challenge the constants in our industry. So be it. You will have the opportunity to challenge the comments that follow. Editorial Director Rick Bush has offered to print your rebuttals in an upcoming issue of Transmission & Distribution World.

Too often, our industry fixates on plugging historical engineering constants into equations and then living with the results as if they were gospel. We are smarter than that. We now have the tools to apply sound engineering analysis and experimental testing to verify or debunk earlier rules of thumb. This should enable us to extact extra value out of our T&D systems by applying designs that more closely reflect field reality.

WHAT HAS ME SO STIRRED UP?

Our industry's insistence on mindlessly sticking with 40 Ω as the input to fault impedance calculations is an example of what I'm railing against. Modern test values for fault impedance are nowhere near the 40-Ω value used by many utilities today. Measured minimum resistance values are on the order of 1500 for dry asphalt, 500 for wet sand, 190 for wet sod, 305 for dry grass, 170 for wet grass and 95 for reinforced concrete.

Using the 40-Ω value, as many utilities do, creates a complex nightmare for the protection engineer. Utilities not only experience more false operations, they find their equipment underrated, which affects reliability. The 40-Ω value also has inadvertently put severe restrictions on relaying options while adding nothing in terms of increasing safety.

ORIGIN OF 40 Ω

The origin of the use of the fault impedance value of 40 Ω (or 30 Ω or 20 Ω) is apparently the result of an AIEE paper, “Overcurrent Investigation on a Rural Distribution System,” written in 1949 by G. Lincks, D. Edge, W. McKinley and J. Leh. This paper describes measurements taken from 1944 to 1947 and is especially impressive considering the limited monitoring capability at the time the data were taken.

The paper describes many aspects of overcurrent protection, but adds information on fault impedance almost as an afterthought. There is little description of the data used for this determination except the following: “The assumed 40-Ω fault resistance used in this investigation proved to be more than ample for determining minimum fault currents and might have been reduced to 30 Ω.” Maximum fault current on this system was about 500 A.

There have been many, many tests on downed conductors performed by utilities, manufacturers, universities, EPRI and consultants. The results have been consistent at all voltage levels: The use of 40-Ω impedance provides virtually no level of protection for high-impedance faults. Tests have shown nothing to the contrary.

MY RECOMMENDATIONS

High-impedance fault current levels are very low and should almost always be represented by an impedance of 80 Ω or more (e.g. 80 A of fault current is approximately equal to 100 Ω of fault impedance on a 13.8-kV system or 90 Ω on a 12.47-kV system). Fault impedances of 200 Ω or more would have to be used to simulate average fault levels caused by most high-impedance faults. All the data that could be found, which represents the past 30 years of research, suggests that the use of 10 Ω, 20 Ω, 30 Ω or 40 Ω has virtually no value in helping detect high-impedance faults. No research in the past 40 years that I am aware of — or could locate via a survey to more than 500 utility engineers — supports the use of these values, and there is no evidence that fault impedance varies depending on primary distribution voltage level or distance from the substation. My five-year study of more than 50 feeders throughout the United States concluded that faults that were detectable showed fault impedances of 0 Ω.

There is simply no data to suggest that using 40 Ω has any beneficial value for detecting downed conductors with normal protective-device operating levels. All test data as well as recorded data obtained by utilities using digital recordings to locate faults suggest that detectable faults have very close to 0 Ω of fault impedance and those that fall on virtually any surface (wet or dry) are undetectable (high impedance) since their fault impedance is usually on the order of several hundred ohms or more. (I personally use 0 Ω of fault impedance for coordination studies.)

As an industry, it is time we challenge engineering rules of thumb that were made before we developed the many testing and engineering tools we have today. Otherwise we will find ourselves resisting progress but without a technical leg to stand on.

---

Jim Burke is an executive consultant for Synergetic Design. He holds a BSEE degree from Notre Dame University and has more than 40 years of experience as an application engineer and engineering consultant. jimburke@synergeticdesign.com