We had a salesman come by discussing the benefits of using ultrasonic testing inside of MCCs and switchgear as a means to detect early on if there are problems building up (i.e. corona voltage present). Ultrasonic can detect problems earlier than IR scanning (and also earlier than accelerometers, oil testing, and laser alignment for bearings, steam traps and other non-electrical systems). I was wondering a few things:

(a) Do you agree that ultrasonic (US) is an effective tool for electrical diagnosis? Or perhaps you find it doesn't live up to the hype / isn't as effective of a diagnostic tool as one would like to think? In other words - you bought it for your staff but it sits in the box not being used.

(b) I'm reading that US is primarily for medium voltage (1000V or more) applications but other literature seems to indicate it could be effective on 480V systems. At what voltages do you see it being effective?

(c) If we detect corona, then we can assume that insulation (cable insulation, insulating devices, insulation around transformer windings, etc.) is going bad and deteriorating, right? Or perhaps it indicates other problems in addition to insulation trouble? Perhaps using it in conjunction with IR scanning we can determine if lugs have simply come loose or other minor adjustments are needed? I believe corona is an irreversible deterioration of the system performance but perhaps if you get the hint of it at a junction in the system you can look via. IR to see if the problem is as simple as torquing connections?

(d) If we detect corona, then what? Is there some literature on time to failure that one can surmise based on the amount of corona voltage noise the scanning process records? I suppose we'd want to record periodically over time to see if it continues to get worse. The salesman made it sound like the electrical panel either 'passes or fails' the corona test but does the presence of corona mean we have to act immediately? The bean counters aren't going to believe just my testimony but if there is industry literature & records of time to failures to anticipate that would be great.

(e) In general, when you run across corona problems is it usually the cable insulation failing or not? Or the insulator device failing or not? In other words is it almost always an issue with 'X' as opposed to 'Y' or 'Z' components?

Thanks.

I know nothing of this type of testing but did find this video plus several more on YouTube:

https://www.youtube.com/watch?v=D-gdWSRUb40

[yt]D-gdWSRUb40[/yt]

I've used it in the past. It is useful for open, exposed equipment such as looking for corona issues on a pole mounted structure for instance. But the RF ("TEV") tools are vastly better in terms of early detection of PD in switchgear. The same tool can also usually be used with a high frequency CT to measure PD in cables. Furthermore if you have DC motors around, FORGET IT. The commutator floods the area with ultrasonic noise so the meter becomes totally useless. However ultrasonic can also be used to detect air leaks, hydraulic leaks, and other things.



PD really doesn't occur below around somewhere around 2000-3000 VAC so even on 4160 V systems it is marginal on phase-to-ground arcing and surface tracking. It works really good when it works. So it won't find insulation faults on 480 V equipment. That's also the difference between UV and IR inspections...IR will detect loose connections for instance because it's high resistance but does not detect insulation faults at all whereas UV (again at medium voltage) works well finding insulation faults and tracking but IR works better for loose joints.



Now we get to the crux of it. See comments above. Ultrasonic works well finding valve issues, compressed air leaks, hydraulic problems, and bearing issues (overlapping with vibration analysis). It detects medium voltage (true medium voltage...4160 V and above) insulation problems. But it doesn't detect loose connections, overheated/overloaded equipment, and loose connections. That's where IR comes into play. You can now buy a decent IR camera that plugs into your cell phone for $400 so there's no longer really any excuse other than training not to do it.

Another interesting and extremely useful application is ultrasonic greasing which we do. Basically you connect the probe and listen as you pump grease into the bearing until it quiets down as much as it will. This provides the exact optimal amount of grease to the bearing. We have several customers that are very happy with our service and some that have bought their own as well. The only odd situation is when the ultrasonic sound level doesn't change at all. This means that either the grease or purge line is stopped up, or the bearing is already greased.



It's like IR...it doesn't tell you how severe, only where and what the problem is. You then have to investigate what is going on. There is some correlation with the amount of corona that can indicate how severe it is though. The ultrasonic guys claim that they've managed to develop a correlation but the TEV/HFCT measurements are a bit more precise and can more directly indicate the severity. On a cable with an HF CT and effectively a TDR (time domain reflectometer) version of a PD meter, it can look for both a pulse and the reflected pulse and by knowledge of their relative times of arrival, it can pinpoint on the cable very precisely where the problem lies. In Europe for instance where most utility lines are buried, it can find the location within a few feet over a couple thousand feet of cable, eliminating the need to jackhammer up several blocks to find it.



That depends. Cables are frequently a problem where they are exposed to traffic and heavy equipment. Except for some early 1970's vintage polymer cables, cables have practically an unlimited life and rarely fail on their own without some outside influence (corrosion, physical damage). The same can be said for most other components, except that workmanship is a big factor. IAEI did a "study" of sorts on this. Most electricians tighten connections until it is "tight enough" by feel. IAEI's results show that the "feel" approach is grossly off most of the time, and usually undertorqued. I can't tell you how many times as a field service guy I end up fixing bad connections later. Second issue is when it comes to medium voltage cable terminations, few people doing them have actually had any kind of training. As an engineer whether I stink at making terminations or not (I'd like to think I'm pretty good at it), my eyes were opened a lot from taking a medium voltage cable termination class about not only how to do it but how medium voltage cables and terminations work in general. I feel a lot better now when I'm working "without a net" doing some kind of strange tape splicing when I've run out of options as well as fixing a lot of manufacturing defects. The "big names" frequently screw this up too even though you'd think they have the resources to stop it. Also although I don't accept a lot of things that he claims, Ben Lanz has made a lot of money basically finding bad terminations and splices in Duke Power's systems among others. One of the big things that he frequently finds is when people do really stupid things with the semicon. If you don't cut it off square, clean off all the residue correctly, and leave everything a smooth transition, you are asking for a voltage stress riser and it will eventually fail the cable. None of my terminations have failed. Either I'm lucky or I was awake in class that day!

Keep in mind though...this is a medium voltage problem. You don't have PD on 600 V cables. You don't have shields or voltage stress (at least at the cable level...board level is a different matter). In 600 V cables the big problems are overheating either from overloading or bad workmanship (loose connections) or sometimes being stupid like burying the enclosure in dirt or putting way too many cables in a small area, physical damage, or corrosion. I could write a book on this stuff...I guess that's why I'm a service engineer. Fixing people's screw ups is what I do for a living.

Thanks for the responses!!

I have used it a good number of times on 480 vac systems and medium voltage primary equipment, and still do on occasion. I've not discovered anything, so far. I have had great results using it for air system leaks, roof and vessel leaks. Air system leaks was interesting. The ultrasonic likes to echo off building surfaces and finding some leaks involved following the 'sound' around the room, up to the ceiling, down the opposite wall, over there...

Same thing happens with electrical because it's generally a surface acoustical wave. In fact when you are near electrostatic precipitators and DC commutators, it becomes basically useless. So the RFI versions work better in those conditions.