Heating and Imbalance of equa length conductor in cable tray 200" long

[powerproblems]

Good day all,
I am looking for the correct explanation of what is occurring in the parallel conductors in the report below. The conductors are (6) Parallel 500 MCM per phase with approximately 1200 amps of load on them. The pictures show the way phases are laid in the cable tray that they are laid in. I realize that the heating of these cables is related to the way the phases are arranged in the tray and that the only cure is likely laying the conductors out in ABC order.
I would really like the explanation of what is happening in this scenario.

Below is an infrared image of the conductors connected at the 2500 amp feed switch located in the MDP upstairs on the mezzanine. The image shows the amp reading and temperature on the brown (Left Phase A) conductors It shows that the conductors are different temperatures, they are carrying different amounts of current even though they are parallel conductors connected to the same point and the conductors are the same length The amount of current is directly proportionate to the temperature:

[URL unfurl="true"]https://res.cloudinary.com/engineering-com/raw/upload/v1429717183/tips/2383A6EF-D06E-4815-A34B-538BDC96E6BC_hyxi6d.SNAG[/url]

Below is a visual picture of the same conductors:

 
This next infrared image shows the conductors in the cable tray. This image was taken where the conductor split off the main cable tray so that all of the conductors seen in the image are the feed conductors to DS-5. I have also marked the way that the conductors are laid out in the tray by phase. On the left there are (6) 500 MCM Phase A conductors, in the middle are (6) 500MCM phase B conductors and on the right are (6) 500 MCM conductors. These conductors are the same manufacturer and all have THHN insulation. Again they are the same length connected to the same point using the same terminations.
The terminations were replaced when after this condition was first documented thinking that some of the terminations had higher resistance than others.
The temperatures are close to the temperatures recorded in the image taken at the feed switch at the MDP and stay consistent throughout the distance of the cable tray and at the terminations to the DS-5 lugs. It can clearly be seen that the conductors that are closest to the next phase are the warmest and therefore carrying more amperage the right A phase wire and the left B phase wire are next to each other and warmer, The right B phase wire and the left C phase wire are next to each other and warmer than the other conductors.

[URL unfurl="true"]https://res.cloudinary.com/engineering-com/raw/upload/v1429717386/tips/F6E6B619-0209-4DF9-84ED-D8CB1A56A7EC_cinhik.SNAG[/url][URL unfurl="true"]https://res.cloudinary.com/engineering-com/raw/upload/v1429717415/tips/2ABEA761-1A49-4788-9A20-57CD05C0AC32_ubqkad.SNAG[/url]


All of the temperatures recorded in the infrared images were in Celsius. This is so that the temperatures can be easily compared to the temperature rating of the THHN Insulation at 90⁰C. Ambient temperature in the building was approximately 24⁰C (75⁰F) at the time of this study. If the ambient temperature climbed to 35⁰C (95⁰F) the rise in the insulation would be proportionate at 60⁰C. This temperature would be more affected by the load on the rectifiers that DS-5 feeds. The rectifiers are run at higher amperage with larger parts. I cannot say what parts were being run when this survey was performed.
The image below shows where the parallel conductors enter DS-5 near the terminations. This is how we insured that the phasing in the cable tray was phased as stated above:

 

[ScottyUK]

What do we open the attachment with?

 

[powerproblems]

Here are the images in JPG fromat

Thanks for looking!

 

[7anoter4]

The way your cables are laid presents the maximum unbalance. See [for instance]:

http://jmst.ntou.edu.tw/marine/18-2/290-297.pdf

Fig. 1. The unequal current distribution phenomenon of a practical feeder with 7 parallel cables per phase.
Here the maximum is 600 A and minimum 88 A.
The correct way is ABCCBAABCCBAABCCBA

 

[melspuds]

As a fan of cable tray systems you can do as 7anoter4 stated, but you will have to adjust for temperature with no spacing between conductors. Also fault current forces are on the cables will be much higher, see NEC 392.20(C). I suggest that you stack one set of phase A,B,C cables in a triangular fashion and have a air gap between next triangle set of ABC cables of the magic number 2.15 x the OD of one cable for rating of each cable as per table 310.15(B)20 see NEC 392.80(2)d. This method has may advantages 1) temperature rating 2) voltage balance 3) current balance 4) lowers the available magnetic forces on the cables and 5) it is much easier for training the cables off of your buss bars. See 392.60(B) for grounding. If a grd is needed in can be run on outside of one of the sidewalls with at least one connection per tray section. Make sure all cable bundles are tied down every few feet. Another good idea is to color code the phases every 20 ft or so to make it easy to identify the phases if you have cable failure. At looking at images off the buss some of your problem is the way the cables are trained and using the triangle method makes it a lot better for training the cables. I hope this will help you.
Good luck,
Dave

 

[powerproblems]

Thank you 7anoter4 and melspud

I appreciate both the reason this is happening as well as conductor configurations that will eliminate or minimize the issue.