Line Impedance Calcs

[111R]

When calculating line impedances, what temperature is typically used for the sag calculation? Sag is often listed at the highest temperature rating of the line, but this is rarely the case under normal loading. Is there a standard industry practice when modeling lines or is it usually up to the engineer to choose a temperature that will be a good representation of the line during normal loading or under fault clearing times?

 

[davidbeach]

Fault clearing should be fast enough that the line temperature remains at (approximately) the operating temperature. Pick a number; the variation in R over a reasonable temperature range will have a small impact on fault currents since that calculation is dominated by the line X value and X is independent of temperature.

If you're accounting for the presence of the neutral on underbuilt distribution, you may want to consider a lower operating temperature for that neutral than you do for you line phase conductors.

 

[111R]

Thanks David. I'm more concerned about the sag with different temperatures. Since the sag can increase by feet depending on temperature, this will affect reactance. I think the common sense approach would be to find sag at normal operating temperature, but I wanted to make sure there were no standards in line modeling that state to use a given temperature.

 

[davidbeach]

My thought - there's so many other unknowns that get represented by assumptions in the system models that difference in impedance due to sag isn't worth worrying about.

The rise and fall of the mid-span point can be impressive, but the overall length doesn't change by much at all.

Do you calculate your fault currents with all possible combinations of generation running? We certainly don't, there's a bit of error there right off the bat; fortunately most of "source impedance" is transformers and other lines and not generation unless one is very close to generation. I've found that doing impedance calcs assuming straight line conductors produces results that are "good enough". If you truly want to improve the match between modeled results and actual fault results, I'd concentrate of improving your representation of the zero sequence environment rather than trying to add one more significant digit to the impedance values. The text book approach to zeros sequence modeling may produce a Z0 value as much as 120% of what it actually is, as measured by fault analysis. If you can make a 20% improvement or tweak a fraction of a per cent, which is better? Obviously that 20% improvement is based on a number of factors, some or many of which won't apply in your case; but personally I'd try to get a better ρ_earth value than 100Ωm long before going after sag effects.

But, pick a temperature value that makes sense in your conditions, document what you're doing and march forth. Would you use the same temperature for ACSS as you do for ACSR and AAC? Yes or no are probably equally valid answers depending on your circumstances. Do you want the maximum impedance for worst case voltage drop during high power flows, or do you want the minimum impedance for worst case fault currents?

 

[Mbrooke]

Id recommend this:

https://www.omicronenergy.com/en/products/transmission-lines/line-impedance-measurement/

While above you are measuring the R at ambient, keep in mind that any variations are taken care of via step distance with normal graded zones. R will vary no matter where you calculate it as conductor temps never stay constant. Personally Id be more concerned with fault resistance then conductor R variance.

 

[HamburgerHelper]

There are papers of people figuring out ways to get more accurate line parameters using fault data. I though strongly suspect that there is so much margin built into everything that getting within 10% of the real parameters probably is enough. The reaches on our distance protection has 10-20% margin. We don't adjust our settings due to the ground freezing during the winter (the Westinghouse T&D book makes mention of this but I haven't heard anyone adjusting their settings for winter). Our system is untransposed but we model it in ASPEN and PSS/E as being transposed. According to a SEL paper by Zochell, in his example distance relaying can read a fault on the outer phases as being 13% different than a fault on the inner phase. We don't have separate settings for each phase. We don't change our ground protection based on changes in dispatch or availability of ground sources. I think there is just so much margin that unless you are using some anti-sag high temperature composite or really small conductors, I don't think you'll care too much about how the resistance changes with temperature. The inductance of the lines relates heavily with the spacing between the conductors. If they all sag similarly, I imagine the spacing won't change enough to really affect the line parameters.

 

[bacon4life]

In addition to ignoring the change in sag due to temperature change, the sag itself is not necessarily included other than as an "average" height above ground. I have seen both 25C and 50C used, but as others mentioned the ground resistivity and shield wire configurations make a much bigger difference than the temperature.

I assume your questions is for traditional distance protection. For traveling wave type fault locating algorithms, instead of regular circuit miles, actual parabolic conductor length is used.

 

[davidbeach]

For traveling wave type fault locating algorithms, instead of regular circuit miles, actual parabolic conductor length is used.

Put the relays into service and close the breakers at one end. From the reflections off the open breaker at the other end you can figure out the round trip propagation time. Take half of that and then figure out an effective propagation speed based on center-line miles and time rather than using the calculated propagation velocity based on line parameters.

First line we did was just shy of 100 miles long and the good old oil breaker had enough pole scatter that the reflected wave from the first pole closed arrived before the second pole closed. Took a bit of help to figure that out.