Hi there,

I believe the better maintenance program the lower probability for arc flash to be occurred. I review NFPA 70B but couldn't find a direct guideline to create a good maintenance program for MTM. I wonder if you can share what do you have in place such as primary/secondary injection, Kelman, Relay testing....

Thanks
========================================

You will find the need for proper maintenance specified especially in Article 130 where it talks about the fact that an arc flash study is essentially worthless if the equipment is not properly maintained.

BUT, 70E stops at the point of being a work practices document. It does not specify a proper maintenance program except to specify what should be included. The requirements are all spelled out in chapter 2 (Article 200+).

In that section, it references two additional standards. Those are NETA MTS and NFPA 70B.). Of those, formerly NFPA 70B was terrible. It was riddled with typographical mistakes and had a lot of other problems with it, but the basic requirements were actually quite good. NETA MTS was much better written but is extremely self-serving and has ridiculous testing requirements that are either utterly out of date or simply self-serving. NETA is the National Electrical Testing Association so naturally a document developed specifically for companies doing electrical testing is going to recommend far more testing than is necessary. An even better set of documents, though it takes a lot of time to ferret out the information, are the standards put out by EPRI. They are not only extremely thorough (often nauseatingly thorough) but provide justification including data behind the recommendations. This is completely missing from NETA MTS and sporadic in NFPA 70B.

Justification (why you are doing what you are doing) matters. For example, NETA MTS recommends torque testing fasteners, or using an IR scan, or using milliohm measurements. The IR scan is probably the only recommendation that should be followed by has to be done energized. The problem with milliohm metering is that there is no standard or best practice for it. Ohm measurements can and do both increase and decrease over time, so trending has little value. The torque testing recommendation is even worse. It is essentially fastener engineering 101 that tells you that using a torque wrench to recheck a fastener once it has been torqued to specification is not only a waste of time but will fail to meet the original torque values every single time, even minutes after the fastener has been torqued. Not only that but it permanently stretches and damages the fastener by doing so. I have wasted far too much time having the basics of fasteners proven to me by fighting failing fasteners in a dryer weekend after weekend for months until I understood this.

NETA MTS also recommends unwiring ALL wiring, including control wiring, and doing an insulation resistance test periodically. What a total waste of time. The likelihood of detecting a problem relative to the number of failures that will be created by unwiring and rewiring everything means that reliability will actually decrease rather than increase.

Another example is with transformers. The standards recommend that you sample once a year and increase the frequency based on some kind of rating scale. The reason for increased sampling frequency is to monitor more closely when a transformer is failing. It is not predictive and sampling more often doesn't repair the damage. So its pointless because it just verifies that the transformer has a developing failure, something that the first test proves. Second is with turns ratio testing. What a total waste of time. The turns ratio test is not accurate enough to detect a single shorted turn, but that's the whole reason for doing the test. In reality if there is even a single shorted turn it will create a hot spot that will show up in the oil testing or visual inspection (for dry transformers) well before the turns ratio test detects it.

Another weakness is in the recommendations keeping up with testing technology. For instance at medium voltage, PD testing will find more maintenance issues than IR scans, but barely even gets mentioned. And, despite the fact that lots of documentation proves that hi potting actually causes equipment damage, it is still being recommended.

Another weakness is in addressing actual reliability rates. For instance small molded case breakers have very low failure rates and are generally not tested. This would be anything with a 600 A or smlaler frame size. Yet the maintenance standards recommend exercising these breakers annually and performing a full blown functional test every 3-6 years. Considering that the vast majority of 15 and 20 A molded case breakers are installed in residential applications and the failure rates are so low that insurance companies don't mandate annual exercising and 3 year testing unlike say smoke alarms, this should say something about functionally testing ALL breakers.

Finally, all of these standards have a primary focus on electrical equipment reliability, NOT safety. Arc flash hazards are going to be affected by equipment maintenance in two ways. First and foremost, overcurrent protection devices have a fundamental and very important effect on the calculated incident energy. Fuses rarely fail and when they do, it's usually to trip early. Breakers can go either way. Second, the likelihood of an arc flash can be increased if equipment is damaged or contaminated. Any maintenance tests that do not maintain or verify overcurrent protection performance, grounding performance, or equipment damage or contamination, have no bearing on arc flash or shock hazards. When we worked on a maintenance standard internally, we had to go through section by section and ask ourselves, "what bearing does this have on safe operation of the equipment"" If the answer was none, we eliminated it from the company standard.

Now to answer your question...

PD testing does a great job of detecting insulation failure issues way ahead of time if you do it 2-4 times a year. You can buy a relatively inexpensive meter to do this test external to the equipment as a maintenance test (detect if there is a problem without necessarily finding it). IR scans do a similar job for inspecting energized joints for hot spots.

CT's and PT's can be monitored by checking to see if the values match. In other words if you have 10 A on phase A and B, and 20 A on phase C, but you know that your loads are balanced, this indicates a problem with the instrumentation.

The majority of problems with breakers break down into two areas. First, the very complicated protection relays driving them often have a number of electro-mechanical related issues that requires repairs. But since it is electronics, the failure rates are essentially random so periodic testing to verify the functioning of the device takes care of the problem. If you have electromechanical or solid state relays, get rid of them and upgrade to digital microprocessor controlled relays. The failure rates and self-checking features are much better. Regardless the last problem is the trip coil itself. There are "trip coil monitors" and you can make your own fairly easily. All these devices do is to put a signal onto the coil and monitor for open circuit conditions, indicating that the coil has failed. Simply wiring a high impedance digital input such as on a relay or PLC card in parallel with the output contact of the relay performs this function. Since the input current is very low (milliamps), the current through the input is not enough to drive the coil. So long as current is flowing the input is energized and the coil is good. If current stops flowing then the coil is bad. When the output contact fires it simply bypasses the input and during this operation the input needs to be ignored.

The second area with breakers has to do with bearing surfaces. The breaker sits in the same position for extremely long time perios which means that the lubricant on the bearing surfaces gets pushed out of the joint over time. This causes slow operation or sometimes even seizing up. Simply exercising the unit (operating it) periodically helps a great deal. But unless it is maintenance free (such as a molded case breaker), it has to be periodically disassembled and regreased. The grease matters, too. Greases are emulsions. In older greases, the oil in the grease evaporated and the grease would harden and have to be replaced. Newer greases such as fluoropolymers don't have this issue and last much longer. At least one manufacturer (Powell) has an 8 year maintenance schedule on their breakers. Considering that a breaker design life is around 20-30 years, this means it may only have to be shut down and maintenance done perhaps 2-3 times over the life of the breaker.

Pretty much everything else is just simple visual inspection.

Keep in mind...there is plenty of test data out there that clearly shows that drawout switchgear is 4 times more likely to fail compared to say a panelboard. The breaker designs are essentially the same and it has everything to do with the drawout mechanism itself. ABB reports that 80% of arcing failures on drawout switchgear are with the drawout mechanism. So inspecting this portion of the equipment, and/or doing as little pulling out or inserting breakers while energized as possible, is the surest way of reducing the risk of arc flash with drawout breakers.

Paul,

Impressive response !!!! Thanks.

You mentioned about the portable PD testing meter to perform maintenance test. Can you provide me the name of that device? I'm not a fan of PD since it is complicate, not able to re-produce the problem and most the time proprietary to the manufacture. I'm looking to extend the use of hand-held IR camera not only to increase system reliability but also employee safety. We currently run Trip unit test, primary/secondary injection but thinking of deploy thermal inspection and Kelman.
====================================

HVPD is the one I've purchased and seen elsewhere.

I can tell you from real world testing that it works. I agree that PD is complicated and at this point there are no standards except for an IEEE preliminary standard and some long established ICEA specifications for cable manufacturers. It is very much a lot of snake oil. And even if you do find a PD problem, the little $5K-6K maintenance testers only indicate presence/absence of a problem. You can often find the problem on your own. Start looking for presence of white powder residue (insulation breaking down) or signs of tracking. Beyond that you need the $20K+ testers that use time-of-flight principles to narrow down the exact location of an issue.

So here's my experience. We bought one. We took it around doing a preliminary "what can we find". We found that DC motors easily trigger the things in terms of ultrasonic noise which is expected since the commutators are effectively arcing sources. Everything else was quiet, except that out of about 30 metal enclosed substations that were tested, three "pegged" the meter. Further investigation showed that the throat "bushing" between the substation transformer and the switchgear housing was leaking water. This section used sections of "C" channel steel as a "seal" around the joint. Water would pool up on top during a heavy rain and then run down to the bottom of the channel filling it up until it entered the throat. Some duct seal eliminated this. The second one used 15 kV uninsulated cables on a 23 kV system. The manufacturer said that we could make up the difference with "air" which is partially true (not going to get into what happens at interfaces) but also the cables were excessively long and laying directly against the metal housing instead of carefully routed, resulting in white powder residues wherever they touched the housing. Simplest fix was to go to 25 kV shielded cables and ground the shields. The third and final sub has a problem in an air cooled PT. It seems that something is too close inside the transformer. It's not terribly critical so we are just running to failure.

All three problems were found EXTERNALLY on metal enclosed substations without opening doors. We simply walked up to the unit and used the magnetic (RF) pickup to detect potential problems inside the enclosures.

If I had found problems everywhere, or no problems at all, this would have basically thrown PD testing into the scrap heap in my opinion. But the fact that we were able to quickly and easily detect real problems before they became actual failures is worth it in my opinion. This is similar to IR scanning...it is valuable if it allows you to detect and correct impending equipment failures.

Paul,

Does HVPD stand for High-Voltage-Partial-Discharge? Does it work on LV or only between 3.3kV and 45kV? IR is good but still require opening the case and doesn't work well with indirect-view objects. Have you used it on secondary cable?

Thanks
================================

Not sure what hvpd stands for other than the name of the company. The tester I have has 3 modes. It uses a Rogowski coil to pick up PD on cables. It uses an ultrasonic detecter to pick up PD at the surface, similar to corona (UV) cameras. This is fairly directional but you can pick up some PD by aiming at the door seal, but not reliably. It also has an RF detector that uses the enclosure metal (door) and induced high frequency pulses to detect PD on the inside of switchgear. These are all online tests. Voltage is immaterial to the test. I have tried on cable but since you need to install the Rogowski coil on the cable ahead of time, this function is hardly ever used. I don't have much underground cable and don't have the $20k+ locator so it is of limited use on cables.