Power Cables in Parallel

[jnims]

I am evaluating unequal breaker loading in an industrial substation. The substation consists of three power transformers and six switchgear groups. Two groups are directly tied to each power transformer secondary by air insulated bus. Tie buses between groups were created by parallel runs of cable in duct banks or bus duct. For normal conditions, the load splits fairly evenly. However, when one tranformer was out of service for maintenance, The load is unequally shared between the four main breakers left in service. I am trying to create a model and use a load flow to verify the model's accuracy. The software program I am using does not appear to preformance the necessary impedance calculations needed to accurately model paralleled conductors (it does perform ampacity calcs). I have reviewed several IEEE papers about modeling parallel conductors, but the calculations are more than I want to attempt in a spreadsheet (cable runs are four conductors per phase).

Is anyone aware of a software program that calculates these impedances so that I can transfer this data to my load flow program?

Thanks for your help.

 

[rhatcher]

I do not know the answer to the question you have asked. However, I understand that you are interested in creating a simulation or model of your system. In my experience with respect to industrial systems, the impedance of the line (conductors) is insignificant next to that of the load and can be disregarded. However, I will add that line impedance may be an issue in cases where the loads are relatively small and the distances between the alternate (tie) sources is relatively large. However, standard practice for sizing conductors will usually make even this case have an insignificant impedance compared to the load. Please provide more info if this does not help.

 

[jnims]

The impedances are small. However, the current is going to divide between the small impedances. It probably was not clear in my first post that the groups are tied together through cables and tie breakers. So any given group's load can be served through its main breaker or its tie breaker and the exact split is going to be related to individual group loads and the small impedances. The magnitude of the cable impedances is exactly why I need an accurate model.

For example, I have observed switching inside a substation where load was being transferred from one breaker to another by the substation's transfer bus. When the breaker's were paralleled, the current split between would be about 66%/33% between the normal/spare breaker. The only difference in the impedances is about 40-50 feet of 2000A bus. I believe the same thing is happening at this plant, but the conductors are not uniform throughout the model - some is paralleled conductors, some is integral web bus, and some rectangular bus.

 

[Dan76]

One thing you may want to check is the wiring on your tie busses, the parallell runs of cable in the bus duct, (I assume they are 3 Phase) Are there all 3 phases per duct or just 1 or 2 ? Every run of conduit should have all 3 phases in it, not one phase per duct. This may be simple but it was the first thing that came to mind. Hope it helps.

 

[jbartos]

Suggestions:
1. Please, provide more info, e.g. voltage level
2. Medium voltage level calculations usually disregard the bus impedances; especially, if they are short.
3. Paralleling subconductors is usually done by dividing the subconductor impedance, Zcs = Rcs + jXcs by number of subconductors, n, to obtain the total cable/bus impedance, Zc=(Rcs/n) + j(Xcs/n) in per unit or ohms, for modeling in Power Analysis software, e.g. SKM DAPPER, A_FAULT, etc. This holds true if the subconductors are of equal size. If they are of different sizes, then one has to parallel them as different impedances in parallel, i.e. Zc=1/[(1/Zcs1)+(1/Zcs2)+...+(1/Zcsn)]. Hopefully, the currents will indirectly proportionally be divided according to subconductors impedances.

 

[jnims]

It is a 13.2 kV system. In my initial system model, I modeled the impedances as 1/n of the single conductor impedance, but my load flows did not match the actual division of current. I started doing some checking and found several IEEE papers that indicate a more complex model may be needed. One of the most compelling is a March 1991 paper “Calculation of Current Division in Parallel Single-Conductor….” in the transactions on Power Delivery. In the actual example sited (6 conductors/phase), individual conductors carried from 10 to 24 percent of the total current. It may turn out that the simple model works fine and there are other considerations.

In response to Dan76, I am trying to get verification on the exact geometrical layout of the conductors.

 

[jbartos]

Please, which software are you using for modeling?

 

[jnims]

Dapper

 

[jbartos]

Have you checked with SKM, Inc. regarding this and possible alternatives?
Generally, if you have papers and suggested equations for possible use, then you may need some better mathematical software such as MathCAD (

http://www.mathsoft.com),

MATLAB (

http://www.mathworks.com),

etc. to perform more difficult modeling/calculations easy. Certainly, you may prefer to have it modeled by someone who is very proficient with those software.