Hi paulengr - Your first line was of great interest to me. It seems like you don't have high regard for thermographic testing. I'm going to take this thread to a more appropriate part of the forum and pick your brain on it. If thermo testing isn't all it's cracked up to be, I'd love to hear your opinion, and would really love to be pointed to any documentation. I see too many clients spending thousands of dollars on it
It’s not that it doesn’t work, far from it.
To begin at the joint bolt pressure and joint compound and sanding if used crack or remove the oxide layer forming alpha spots. Higher pressure creates more and mushes the metal of already formed spots. These are cold welds. Even as preload releases you have to almost totally lose contact pressure to get any appreciable increase in resistance. If you have access to one try it. Bike together two bus bars then loosen the fasteners while watching a micro ohmmeter.
Basic fastener theory shows why if the initial installation is done correctly the joint is good for life. There are always signs of corrosion, stripped threads, and other obvious issues with failing joints. Those are just as obvious on initial installation. And if you find one there are always more. It’s a static load. Fasteners self loosen via the Jost effect. Those conditions do not exist on bus bar connections and rarely even in mechanical power transmission. Google it and Junkers machines.
Second issue is the 800 lb. gorilla in the room is constriction resistance. Going back to alpha spots less than 10% of the total joint area is actually in contact. At low currents this is not an issue. So if I have a 1” piece of busbar and I put a 7/8” notch in it how much heat do I get with 10 A? None. If I go to 1000 A how hot will that notch get? That’s constriction resistance. The joint is only going to overheat when you need it most under near capacity loads. An IR rule of thumb is 25% of capacity. With a typical MCC spec of 800 A horizontal 300 A vertical are you really going to get 200+ A on the horizontal splice plates? I can easily rig a breaker tester to produce 800 A of current and do a millivolt drop test in the shop but it’s not practical. But overall it does pick up early on issues.
Third issue is design. Once upon a time we built outdoor overhead substations. We used large open frame knife switches. Protection was mostly a lot of chain link fences and danger signs. Those days are long gone. Now we build equipment where the controls, switching equipment, and bus are in physically separate compartments. The only energized access is controls. Access to components requires tools and is interlocked so that it often can’t be accessed energized. Even in de energized testing often I have to disassemble key interlocks. Bus compartments are bolted and impossible to access if the panels are back to back or against a wall. Contact fingers are always obscured by the equipment. So as a consequence direct measurement is usually only practical in uncommon cases. You can add windows but these are limited. It takes multiple large windows to be effective, if it is even possible.
There are three methods of detecting loose connections typically mentioned. The first is torque testing. Every fastener engineering book will clearly state not only the fallacies of this method but why it’s a terrible idea. Within even minutes after tightening a bolt the bolt relaxes and preload drops. Reloading overstretches the bolt. And 75% of torque is overcoming thread friction. Dirty fasteners, improper wrenching techniques, improper installation (flat washers between bus bars...don’t laugh I’ve seen it) stripped or galled threads. The issues are there on day one.
Micro ohm readings measure DC resistance using up to about 10 A on a DLRO or Ductor. So it directly measures resistance but since we need constriction resistance and alpha spots tend to remain until right at failure it doesn’t change.
Threw other methods not often used or mentioned are fiber point sensors, distributed fiber measurement, and millivolt drop. Fiber point sensors run several hundred dollars each. They are rugged and have no issues with voltage limits. The downside is the overall cost. Distributed temperature uses nonlinear properties of single mode fiber plus time of flight to measure temperature at any given distance on a single fiber but the technology costs hundreds of thousands. Millivolt drop just measures the voltage across the joint. It is very simple and limited only by access and safety issues. It has kind of fallen out of favor.
So coming full circle IR scans of bus are looking for an issue better served by other methods. It is the exception not the rule that the bus is visible. Fortunately copper is so thermally conductive issues can be detected indirectly: But the issue stems from initial installation issues. Self loosening is a myth, doubly so in electrical pressure contacts where alpha spots cause hysteresis even if loosening occurs. IR scans so frequently detect loss of spring force in fuse clips and breakers, overheated bearings and motors, worn contact tips, and many other issues compared to rare bus bar incidents.
IEEE 493 gives fm failure rate data on disconnects of 10E-12 and similar data exists to the point that IEC 61511 for process safety systems and similar standards simply ignore bus and wiring failures. While I believe they exist human caused failures dwarf any age related issues.
Going further I reviewed every arc flash incident in the investigation database OSHA maintains from 2009 to 2014. There was a single incident though it supports my point. At a parking deck an electrician covered up a lighting panel opening by sandwiching loose metal between the covers. When an attendant later operated a breaker the metal fell into the bus initiating an arc flash. So bus was involved though not loose bus and poor workmanship was the cause.
Thus it is far more likely that issues will b found with contact resistance tests. They are just as effective on spot checks on say annual PMs. By nature you get a visual inspection for free: They can be trended. They are done safely offline. They are not affected by load. I am a strong believer in IR just not specifically for bus.
So I think it’s tilting at windmills. It’s not really the best way to detect the concern and the concern is not very likely. As to the OP in a division 1 area energized work is definitely a bad idea. I’m really not worried about equipment temperatures internally but the fact that human error rates are vastly higher. Taking a cue from two large LNG storage facilities in my area locate all starters and switchgear outside the division 1 and preferably 2 areas. Purge and pressurize instrument panels. Open only after sniffing and only if necessary when there is no sign of gas present. At that point you can have an overheating contact but it doesn’t ever become a hazard. Infrared cameras for this are a solution looking for a problem.
