Generator cubicles : thermal imaging and EMI test

[SolarPrestige]

I have a generator which has a cubicle on each side, one called the line-side cubicle (LSC) and the other called the neutral-side cubicle (NSC). Just thinking about the LSC, it has connections coming in from the generator bushings and goes through to the isolated phase busduct on the other side. Inside it has busbars, voltage transformers (VTs or PTs), current transformers and surge arrestors. There are lots of joints, applied insulation and leads inside this box. The NSC is similar but with added neutral impedance equipment and with no surge arrestors.
Here’s the problem. We cannot undertake thermal imaging in these boxes. We do other routine inspections and tests but we cannot do the thermal imaging. One way to see the temperatures inside would be to take covers off and view while the generator is running, but we would not do this due to safety concerns. Another method is to install IR windows, but there may need to be a lot of these. Or we can just say low risk so let’s not worry about it. I don’t immediately like the do nothing answer, but installing windows would need a pretty good justification.
A flash over inside one of these cubicles would be a very high energy incident as the generator cannot be turned off instantly. There’s at least a big initial event from the generator and grid, followed by the generator winding its excitation down over some seconds while continuing to feed the fault current.
Has anyone seen issues with these types of systems related to incipient faults (e.g. poor joints, PD activity)? Anyone know about these systems failing?
And a related question, can EMI testing be used to find the same type of issues? How is an EMI test set attached to measure on this type of system? I am reading the Doble brochure for PDS200 and EMI-Surveyor and they say no physical connections on one and no HV connections on the other.
Thanks, Greg.

 

[davidbeach]

Fixed IR cameras inside the cubicles?

 

[SolarPrestige]

David, hmmm. Pity those Flir cameras cost an arm and a leg to buy, and I would need how many to see everything I want to see. I have used the cheap ones and they are not so nice.

 

[davidbeach]

If the cameras cost “an arm and a leg”, how would you describe the cost of the generator damage you are looking to prevent? How many minutes of revenue?

 

[Hoxton]

Are these indoor or outdoor?

I have seen some spectacular failures of line side cubicles outdoors due to water ingress.

Who made them?

 

[SolarPrestige]

Regarding cost, like all plant, once it is agreed what monitoring will be built in from new and then it is built, it is usually a battle to get new monitoring added. The outcome of an event could be expensive to repair, and lost generation opportunities might have a high cost (depends on the time of year and how much water is in hydro storage lakes), but the chance of events like this occurring will be very low. From a fleet perspective, how many LSC's blow up each year? Probably typically none so maybe it's how many years between each failure. Even the bigger fleet of all generating plant that use LSC's, the number of failures per year is small. Actually finding data on failures is not easy, so maybe I don't really know. I certainly haven't heard about concerns through the grape vine. I have used IEEE Gold Book and its risk tables, but they don't exactly have a "LSC" category.

We have two of these generators installed outdoors, now 10 years old. The LSC's have not given us problems, don't have water ingress issues. I recently got the LSC modified to reduce the chance of arcing faults occurring. Replaced the installed VTs which had exposed fuse and fuse clips on the top, with a VT that had a fully enclosed fuse holder and was able to insulate the one exposed fuse terminal after installation. This was the only thing inside the LSC that had uninsulated terminals, and changing this was from a project that looked at arc-flash risks on the plant.

This LSC was manufactured for GE by a company called "Farmer's Marine Copper Works" in LeMarque Texas. I don't know anything about these people.

 

[GJY]

Hi,
It is difficult to use IR cameras at this location because usually the cameras are disturbed by theelectromagnetical field. (>10kA)
A good thing to do is a regular visual inspection every 5-10 years and to apply thermotabs directly on electrical connection to be sure the temperature at the contacts doesn't exceed 105°C.
And of course let the electrical connection be done and prepared by a specialized company during maintenance (specially with aluminium, you have to take care about aluminium oxyde).
The occurence of fault at this location is most of the time bad connection with local increasing temperature who damages the materials in contact and increase the losses by Joule Effect, till the moment the aluminium palm (if any) on the busbars locally melt (>700°C)

I have worked in the past for a company who was manufacturing seach GLAC/ GNAC (Generator Line Accessories Compartment)/(Generator Neutral Accessories Compartment) for GE as a subcontractor in Europe.

 

[SolarPrestige]

Hoxton, can you provide the names of places where failures have occurred? I can then search what is available on the web.

GJY, thanks for that suggestion. We have thick insulation everywhere so its a bit of a guess what the external temperature would be if the internal temperature was reaching 105°C. So maybe we would look for temperatures around 80°C plus or minus 10°C. Try it and see what happens. I can see the main benefit being the ability to see an increase suggesting that there is an issue at the joint.

Also is 105°C based on a standard, or is this what the company you worked for used as the upper limit? I have recently been reviewing temperatures on terminals of LV contactors, and there the acceptable terminals temperature is anything up to 120°C depending on terminal plating and other factors (like cleanliness), but is often quoted as 105°C (this is for LV equipment so I wouldn't use it to characterise HV equipment).