Retorque of bolted electrical connections

I am from the camp of retorque when thermal testing indicates. If the system was properly torqued and tested when installed ( most equipment manufacturers have standard torque values for their equipment) there should be no additional problems unless you believe the system has been "stressed". As an energy engineer, any loose connections should be treated as they have a direct impact on electrical costs as heat loss, as much as 2% of an electrical bill can be attributed to these loose connections.

Run1on

I agree with retorque but wonder if other steps typically needed are followed? On buses you would want to loosen and clean before retorque. If a thermal or mechanical sress has caused loss of torque you can be relatively certain that corrosion has started on at least a very thin level. The salts from the corrosion will "snow ball" with subsequent retorque and loss of torque will occur even more quickly.

I was told not to retorque because of something called "cold flow". When you tighten a cable connection to the proper torque, after a length of time the conductor will flow a bit and you will be able to get some additional movement if you apply a torque wrench and tighten to the original value. This relaxation or "cold flow" is accounted for in the original torque specification. Retorquing will increase the rate of cold flow and could result in a poor connection after the conductor has been heated up due to the load causing even more code flow. Based on this, I was advised that a conductor connection should never be retorqued. If there is some indication of a problem, the conductor should be cut back, stripped, and terminated to the correct torque in a new terminal. I have never been able to document this information. If anyone else has heard this or has documentation one way or the other, I'd like to see it.
Don(resqcapt19)

Some manufacturers will specify to retorque after a given number of operating hours, generally after 16 to 32 hours. It is very difficult for a manufacturer of electrical equipment to specify based upon "cold flow" as hardware specs are not generally given for expansion @ N degrees. Expansion of the copper bus will occur much more rapidly, in general, than of the steel bolt. Hence, the copper will tend to compress between the bolt head and nut before the bolt can actually expand sufficiently to relieve the compression. Even without the cooling process that would normally follow on any circuit that isn't run 24/7 at load, the expansion of the steel hardware will allow for some loosening of the joint. Retorquing should take place after recommended running times while the equipment is de-energized, but still warm from running. This will minimize the problem of copper compression during warm-up.

Run1on - Personally I'm just a little bit more concerned about the possibility of the connection failing than about the energy loss. But you've awakened my curiosity... where did that 2% number come from?

Thanks, electricpete !

For the 2% question, which I join.

Before committing a "constructive discharge" from my position of a hired thermographer, I was once questioned by the authoritative office forces of "how much energy a hotel is using, so we can do a sales pitch to the corporation telling them that thermography will help in saving a lot of money by eliminating waste of energy for heating connections throughout the whole chain". I told the sales pro (!) (and presumably an engineer), that any thoughtful maintenance professional will treat such treatise with disbelief, and maybe with suspicion. No 2% level was even mentioned during our exchange.
I stressed, that the "energy waste" is insignificant in relation to the total electrical power being used. Much more important, I continued, is to _prevent_ power disruption (which is a "saving" by itself, sort of) and damage to the system, even by an electric fire. I was pressed for giving the numbers nevertheless...

So I gave them the average number of installed kVA (transformers) in those hotels... thinking - will such argument fly ?
It does, as I hear.
I still think, that _finding_ a problem is much more important than, say, 100 W (assume) of heat being generated by a bad fuse clip, terminal or faulty disconnect hinge. Particularly for a fire pump - and this was a _real_ case of concern to me. Let's count: 10 faults will generate 1 (one) kW. An equivalent of 1 (one) facade illumination lamp... Nothing to talk about, even for a whole chain of hotels.
Much more important is to keep the lights on
for the hotel guests - and this is a reliability factor at play.
Or: infrared scan quality.

My estimation of that percentage is 0.2%,
based on the thermographic anomalies I am
usually finding. Accounting for all other
smaller and not found anomalies let's make it 0.5% .

Any dissenting voice ?

ja2taj@netscape.net