Hotspot detection on trenched DC bus - metals smelter

[kitosepa]

Gents and Ladies. I have searched the forum to find any info relating to hotspot detection on busbars.
I have a problem where we cannot perform any busbar inspections on our 1000V +300kA busbars for our smelter pots. These busbars are trenched, and where they run between potlines, these trenches cannot be opened due to production vehicles movement and other associated dangers. And then also the old reason: we cannot shut down for inspection due to production reasons.

I am looking for a way to detect hotspots forming on the busbars. A suggestion was to measure the inlet and outlet air temperatures on the forced air cooling system to detect any hospot forming in the trenched busbar sections that cannot be inspected. Due to the very strong magnetic fields optic fibre communication would used. This, coupled with the amount of sensors and labour required proved to be very expensive. Personally, I have doubts that this might work at all.

I thought about measuring voltage drops (or resistance, which should not change significantly unless faced with a problem) across sections of the lines. There are many variables that can influence this, but I really need some advice on this issue.

Any ideas out there? Who has encountered similar problems, and can Thermography cameras work in those strong magnetic fields?

 

[rcw retired EE]

At the smelters I worked in, we used millivolt measurements across bus connections including the cross over busses in trenches. In retrospect, we probably had as much error from induced potentials in the leads as we had valid readings. We marked the measuing points on the busbar to make sure the readings were taken at the same same physical location each month. The technicians would alsorecord the DC amps when the measurements were made. We were looking into doing the amp measurement via radio connection to the main Celtrol computer and doing automatic data logging of millivolts, amps and equivalent resistance.

The problem with millivolt measurements is you are looking for a localized problem but you can only read values that are averaged over a long piece of bus. Also the values are so low that it is difficlut to extract good data. You end up with trend informaton which may or may not be accurate or useful.

We tried thermal imaging but the cameras available then did not work well in the high magnetic field and/or high temperatures.