current splitting @ parallel paths thru switchyard - effect on hotspot 4

[electricpete]

Our powerplant has a breaker and half scheme in our 345kv switchyard. Our generator feeds 2200A at into one of the bays and splits in two directions through two generator-position breakers.

We have a hotspot identified by thermography on a disconnect for one of the generator position breaker disconnects.

First reading was 64C rise under 7.5mph wind conditions.
Two days later we had only 10C rise under 2.5mph wind conditions (normally expect decreased wind to cause temperature to go up). Thermographic images are here if anyone is interested:

http://maintenanceforums.com/evefil.../ubb.x/a/ga/ul/8071076041/Y601_Disconnect.ppt

I haven't been able to determine current in these two branches. Power flow through various circuits is monitored with revenue metering but due to deregulation, it is a highly protected secret.

I haven't drawn any conclusions as to the cause of the decrease in temeprature yet. It is possible the connection begins to heal itself or the current has changed either due to other external loading conditions or remotely possible the current has changed due to the resistance of the connection. I rule out the possibility that the OTHER parallel path same phase has a high resistance connection forcing higher current through this phase based on adjacent connection points... this phase is not higher than it's sister phases at the adjacent connection points. Again I am still gathering data and not drawing any conclusions but I have a question:

*** My questions are: how much effect do we effect the resistance of that degraded connection have on the splitting of current through parallel paths. To get to the same point, the two parallel paths have to go perhaps 80 yards distance through the shortest path (with multiple other parallel paths). It is tubular aluminum bus. I am suspecting that the current sharing in a normal circuit would be determined primarily by inductance and resistance would be much lower. If that were true, I wouldn't think the resistance of the hot connection would affect current sharing unless the resistance was extremely high. Any thoughts on this question? What would be the rough X/R ratio of this type of bus? ****

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

No response, so I would like to restate it as simple questios:

1 - What is your gut feel about the effect of high resistance disconnect connection on current sharing in this circuit (two parallel paths of aluminum bus). Is it significant or negligible.
2 - Assuming no abnormal resistances, would the impedance of this type of bus be primarily resistive or inductive?

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[alehman]

1. The resistance of the switch is probably the majority of the impedance. If one switch has relatively high resistance, the other path may take the bulk of the current.
2. Inductive, I would guess. Bus material resistance is available from the manufacturers. X depends on the bus dimensions and spacing. It can be calculated from texts such as Stevenson.

 

[electricpete]

Thankd alehman.

I agree with #2. The bigger the conductor, the higher X/R.

#2 would seem to make #1 less likely (less likely that varying R in inductive circuit significantly affects current). I admit I have not provided enough info for an answer (unless someone has a feel for inferring resistance from temperature rise). I'll try to gather more data on Monday. In particular I will try to determine current split by looking at other portions of the circuit. In the meantime, any more thoughts or guesses?

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[tinfoil]

If you have sufficient heating on a component to show up as a hotspot via thermography, then I would wonder about the integrity of the 'other path', despite what your other readings say.

Properly installed equipment in a yard WITHOUT parallel paths do not ususally create hotspots. When they develop on single-path systems, the result is invariably a loose connection, heavy corrosion, or badly seated switch blade, all of which lead to a relatively high impedance and heating due to I^2*Z effects.

If you are having hotspot problems in one of two parallel paths, then there must still be enough current in your 'hot' path to create it. This implies that the presumably high impedance at the hot point is still 'low enough' compared to the alternate path that appreciable amounts of current still will flow that way.

Are you sure that your other path is good and healthy? Typical impedances for buswork and closed switches in good condition should be down in the microOhm range. A nearly-open condition might NOT show up via thermography if the alternative path allowed near-zero current to flow in your 'good' path.

 

[stevenal]

The current divides in proportion to impedance, but heating is in proportion to the square of the current. Little bit or R slightly decreases the current on the hot path, but the remaining current squared times the R causes the heat. Change in heating is from changing load or changing R. Stop analyzing and fix the disconnect.

 

[cuky2000]

Hi Electricpete:

We should keep in mind that 11/2 breaker switchyard is expected to be designed to operate at full capacity with one bus open. This mean that all components including switches, breakers, instrument transformer, meter, etc should be capable to carry the system full load continuous at rated capacity.

How much effect do we effect the resistance of that degraded connection have on the splitting of current through parallel paths? I expect the split difference to be marginal different since the reactance is the dominant factor by far (X/R~100). The small unequal split have inconsequential effect in the operation of the system and the revenue metering system other than additional small losses do to the Joule effect.

Check the enclose relations with rough assumed data:

Split current ratio: I1/I2=Z2/Z1=(R2+[Δ]R2+jX2)/(R1+[Δ]R1+jX1)
For large X/R, ==> I1/I2~X2/X1
_{Where:
1 & 2 indicates parallel path loops
I = current
[Δ]R =Increase in resistance do to hot-spot
R = resistance
X = inductive reactance.}

A quick estimation of X/R ratio is presented assuming a typical 345 kV substation bus clearance of 13 ft with 4” Schedule 80, 6061-T2 Aluminum busbar:
- Resistance at 70oC = 5.284x10^-6 Ohm/ft
- Reactance @ 1ft spacing = 3.96x10^-5 Ohm/ft
- X/R=13*3.96*10^-5/5.284*10^-6.

 

[cuky2000]

See the enclose info from IEEE Std C37. I hope this could help in your research.

http://cuky2000.250free.com/Switch.pdf

 

[electricpete]

Survey today shows still low temperature rise ~ 10C. From review of the thermal patterns, it seems pretty clear to me that
1 - all three phases are balanced for both paths.
2 - There is more power flowing out of our generator through the breaker without the hot disconnect than the breaker with the hot disconnect. Since both locations appear balanced among phases we rule out the connection as a cause. The cause of this power split would seem to depend on the physical layout of the switchyard and where the lines tap in, or possibly also depend on changes in conditions external to the plant... power flowing thru the switchyard.

Cuky - you comments confirm my suspiciouns that we don't affect contact resistance to affect current splitting unless the contact were extremely bad (probably to the point we might see visible evidence of past overheating with binoculars... we don't).

Stevenal - I'm not sure if your comments were serious, but since the disconnect is on the generator side of the breaker, we would have to take 1250MW off-line to work it. Would like to understand risks associated with various options to make informed decisions. One option would be to measure current at CT secondaries inputting to our relay schemes. At this point we elect not to even take that risk because we feel pretty good that this disconnect cannot cause any transient as long as we don't take the parallel path out of service.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[stevenal]

Pete,

I'm waiting for approval, so I cannot see your pictures yet. Sometimes simply opening and closing a disconnect can fix the problem. This can be done hot if a parallel path is available. If the parallel path is present for redundancy, relying on its presence continuously defeats this purpose. In my experience, the planned outages are more desirable than the unplanned ones. If it stays cool enough, maybe you can delay repair until the next scheduled outage.

 

[electricpete]

I didn't realize the link was restricted. There must be some cookie on my computer that allows access. I'll try to figure out a better place to post next time.

Our plan is to continue to monitor. If we should reach 100C rise, then we will open the breaker, cycle the disconnect a few times and reclose, as Steve suggested.

Summer peak is already here in Texas. Any switching carries some risk. No-one seems inclined to cycle the disconnect unless really necessary, knowing that we do have a reliable parallel path.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[alehman]

Let's back up a bit. I can't get to the photos either. I assumed you were speaking of a single disconnect. Are (were) all three phases hot, or just one?

Was there only one reading showing high temp? If so, would it be reasonable to suspect a bad measurement? IR can be tricky. Reflections, sunlight, etc. can mess up readings.

 

[electricpete]

One phase disconnect contact stands out as being hotter than all others. Initially 68C, later only 10C rise. Even at 10C rise, it is hotter than the others which show <1C rise when compared to adjacent bus.

The initial thermographer was qualified and double-checked his result. This shot was shot toward the east in the morning when sun reflection would not be a factor. No other sources of reflection hot enough that I know of.

On the day of the third reading (yesterday - Monday), I used the camera myself. Still 10C rise. Comparing the buswork itself, all three phase temperatures (and I assume currents) are balanced in all paths. However the three-phase path which includes the disconnect appears to have less load than the parallel three phase path. Clear as mud?

It did rain between the first and 2nd reading, prompting speculation by others that the connection was washed. I don't rule out the possibility that even though our generator is constant load, the change in line loading at different points in the switchyard changes the current split. It could be that the three-phase path with the hot connection was carrying the most load at the time of the initial 64C reading. (If the condition starts rising again, that would tend to confirm the latter explanation.... will be getting readings every week)

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[stevenal]

Pete,

Remember that reactance increases with separation. The preferred return path will be through the local adjacent phases, rather than through the parallel path. A little bit of resistance on one phase may not upset the current balance too much because of this.

 

[electricpete]

"A little bit of resistance on one phase may not upset the current balance too much because of this."
I attribute this to ratio X/R only considering self-reactace X.

"Remember that reactance increases with separation. The preferred return path will be through the local adjacent phases, rather than through the parallel path"

I don't understand what you are saying. I guess I am picturing coupling between phases (mutual reactance) to have a fairly insignificant impact on the division of current because of the large spacing that you mention. Is it reasonable to neglect this?

Or are you saying that if my hotspot high resistance shifts current to the paralllel path, the other phases on the same disconnect will also experience reduced current? If that's what you're saying I would like to understand better.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[stevenal]

Pete,

I guess if I thought it was insignificant, I wouldn't have mentioned it. There is a mutual coupling between supply and return conductors that favors the closer lower impedance return path. Conductor tables generally list the reactance for 1 foot spacing. This value is then adjusted according to the actual GMD, and goes up as the GMD increases. Without crunching numbers, my guess is that that a small bit of extra R in one phase has two effects: Do to the extra X in the longer path, return current favors the short path even through the extra R tending to keep the current in that phase high. The other effect would be the one you mentioned, considering the two good phases as the return paths all three phases will have a tendency to drop together. The two effects together would put the current somewhere between normal and where it would be if all three blades had the same extra R. Unbalance would depend on the actual R and geometry.

 

[electricpete]

Thanks. I am just trying to understand a litte better. You mentioned two effects.

1 - "Due to the extra X in the longer path, return current favors the short path even through the extra R tending to keep the current in that phase high."

I read this sentence 5 times and still don't understand it. I think you're assuming high resistance in short connection (like I assume). Then that extra R to the extent it effects impedance (very small) will tend to decrease current in the short path and increase in the long path. So if the extra R (in long path) increase current in the short path (same phase), what is the meaning of the phrase "even though". I may be overanalysing your comment. If I replace "even though" with "and" it makes sense to m.

2 - "The other effect would be the one you mentioned, considering the two good phases as the return paths all three phases will have a tendency to drop together."

The effect I understand but again at maybe 10 or 15 feet spacing I have a hard time imagining a difference. Self inductance must be orders of magnitude higher. Now if the high resistance were out on one phase of a transmission line and the parallel path went miles through another transmission line, I could see the cuumulative effect might be significant. But here the parallel path is only within the switchyard. I haven't done any calcs but my gut says it is not significant. If it is I'd like to know/understand because it changes the way I look at the problem. Steve and others - do you have any more comments on whether mutual inductance is an important part of the problem?

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

One correction to my last post (as if it wasn't already rambling enough!).

I read this sentence 5 times and still don't understand it. I think you're assuming high resistance in short connection (like I assume). Then that extra R to the extent it effects impedance (very small) will tend to decrease current in the short path and increase in the long path. So if the extra R (in long path) increase current in the short path (same phase), what is the meaning of the phrase "even though". I may be overanalysing your comment. If I replace "even though" with "and" it makes sense to me.

should be changed to:

I read this sentence 5 times and still don't understand it. I think you're assuming high resistance in long connection (like I assume). Then that extra R to the extent it effects impedance (very small) will tend to decrease current in the long path and increase in the short path. So if the extra R (in long path) increase current in the short path (same phase), what is the meaning of the phrase "even though". I may be overanalysing your comment. If I replace "even though" with "and" it makes sense to me.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[cuky2000]

Unequal currents in the generator bay could be driven by the load flow demand in the internal switchyard circuits. You may check this in other bays without hot spot or determine the current through the center breaker.

To illustrate this case, consider the power flow in the bay with a generator and transmission line as shown in the enclose sketch.

The generator will input current to the bay and the line will deliver out the power. The current in each node under normal conditions do not have to be equal.

The effect in each bay of the mutual impedance may not be significantly different to change the current flow do to the physical symmetry of the substation with respect to the center circuit breaker.

Please notice that the Std C37.30 allows to the switch to be at 70 oC and other individual components can safely capable to handle temperatures from 75 oC up to 125 oC depending in the material (Cu, Al, silver & alloys) and the parts such as mechanical joints, contacts, connectors, etc.

I under the impression that you still are operating the disconnect switch within the allowable temperature.

 

[stevenal]

Pete,

The word I used was not "though" it was "through." I may have been unclear what I meant by "long" path. I wasn't referring to extra conductor length, but greater separation.

I'll rephrase. We have parallel paths, but phase B of path number one has a somewhat resistive connection. Current flowing in phases A & C of path 1 normally will return on B phase of path 1. Due to the added resistance, one would expect current returning on phase B to be diverted through path 2. But GMD comprising the A & C phases of path 1 and B phase of path 2 is much higher than the GMD of a single path. So current returning on phase B sees added resistance on path 1 and added reactance on path 2 and divides accordingly.

But phase B is not in isolation, the added impedance also affects phases A & C, which also have path 2 available. A & C currents diverted to path 2 along with B see the lower GMD and reactance of a single path.

Overall effect is less unbalance than one might expect if resistance only is considered.

A similar effect is seen during ground faults. The whole world is available, so one might expect the current to spread out and use the whole cross sectional area of the planet. But current straying too far from the supplying conductor sees increased reactance. Unless there are some nice metallic pipelines nearby, most of the current will return in the ground directly beneath the faulted line even when the pole line zigs and zags.