This week's question is a variation of a maintenance question asked a few weeks ago.

As stated a few weeks ago, equipment maintenance is playing an increasing role in the 2015 Edition of NFPA 70E.

A specific example that is the topic of this week's question is item number 2 in

130.2(4) Normal Operation. Normal operation of electric equipment shall be permitted where all of the following conditions are satisfied:

(2) The equipment is properly maintained.

An informational note follows that states:

[INDENT=2] ...The phrase "properly maintained" means that the equipment has been maintained in accordance with the manufacturer's recommendations and applicable industry codes and standards."[/INDENT]

Here is this week's question:

Based on NFPA 70E's specific text, would you be comfortable stating that your company/client's electrical system is properly maintained?

• Yes
• No

Well, my role is to get the electrical systems to the point where they are properly maintained. And it will probably be a few years before we are really there.

I think the maintenance is pretty good. However, I don't believe any of us would want to make an "official proclamation" based on NFPA 70E's definition which could be used in court should something go wrong. So is maintenance pretty good - yes. Would I answer yes based on the NFPA definition? - no.

This is where we have a definition problem. 70E says "properly maintained" but then does not define it! Neither does IEEE 493. NFPA 70B and NETA MTS define mainteannce standards but they are both quite extensive and at least in the case of NETA, not really implementable without some modifications. A second problem and more practical issue is that "properly maintained" from an electrical hazard point of view is somewhat different from a mainteannce point of view. While normally maintenance condition and electrical hazards go hand in hand, there are many instances where they do not necessarily go together. For instance while it is important to maintain a proper level of inhibitor in an oil filled transformer from a maintenance point of view to minimize detioration of the cellulose structures over time, it does not impact safety in any direct way. Similarly monitoring PCB levels in converted transformers helps to avoid the oil turning into an environmental disposal issue but again, has zero safety risk.

What we need is a set of minimum standards for safety purposes. It is certainly possible to define many of those standards in terms of what is required without specifying how to go about doing it. Certainly failure rates and/or likelihoods can and should play a critical role. It should not be the purview of 70E to define "how" to perform maintenance but should certainly be within the purview to define what "acceptable" maintenance is. One would think that the major items should be:
1. Structural integrity and degree of contamination. This has more to do with shock hazards than arc flash.
2. Proper design and installation to meet short circuit and thermal capacity requirements. Failure here leads to elevated risk of arc flash.
3. Some very minimal periodic inspection of fuses. Fuses do fail to trip for overcurrent but this is not common. Most common problems will be corroded fuse holders which really goes back to #2.
4. Condition of breakers. This is hard to quantify. Breaker and trip unit designs strongly influence failure rates....sealed units such as even the common 15 or 20 A MCCB tend to have very low failure rates compared to designs with "red grease" designed for frequent periodic greasing with fairly heavy trip mechanisms. Microprocessor trip unit reliability is much higher than electro-mechanical designs.

Right now, 70E does just the opposite. "Acceptable" is defined in terms of compliance with a maintenance standard. So either the maintenance standard needs to be more detailed in what is required for safety purposes or else 70E needs to be fleshed out with more requirements instead of referring to 70B or MTS. Section 200-250 go this direction. But 205.4 says to use industry consensus standards (referring to 70B or MTS) without further clarifying this, 225.1 refers to "adequate contact" for fuses without defining how to determine that, 225.2 talks only about externally visible MCCB damage, and 225.3 specifies only that breakers have to be tested after interrupting faults. The latter is a tricky issue because it depends on the size of the fault and what the breaker is really rated for. I think we can all agree that 225.3 is lacking in a lot of practicality. I don' tknow of anyone swapping out breakers immediately after a fault, and only NEMA AB-4 specifies a simply visual inspection following a fault that would not be so onerous as to be impractical.

Debating the new NFPA 70E "Normal Operation" requirements

Prosecution Attorney: Is your equipment properly maintained?
Defendant: Yes
Prosecution Attorney: Do you instruct your employees to stand to the side when switching the equipment?
Defendant: Yes
Prosecution Attorney: Why?
Defendant: In case the equipment fails.
Prosecution Attorney: Then equipment can still fail during normal operation?
Defendant: ??!!##

Completing the argument...

Yes, equipment can still fail during normal operation. There are some cases where there is such a thing as an absolute guarantee of safety but these are the exception, not the rule. The most advanced, modern safety codes such as IEC 65108 instead take the approach of quantifying the likelihood of a failure and then taking steps to lower the likelihood to an industry accepted standard. This varies from industry to industry but generally speaking if we are talking about major injuries or fatalities, the industry accepted standard is around 1 in 100,000 to 1 in 1,000,000 depending on how many people are involved and depending a little bit on industry norms. Even the U.S. EPA takes this approach. They are looking for odds of less than 1 in 1,000,000 for many cases where an absolute guarantee of no injury cannot be given and this is the accepted environmental standard for things like toxic chemicals causing injury to the public.

The past approach was extremely prescriptive and took either an absolutist or opinionated approach. For instance most safety codes used to require hard wired interlocks. PLC's were simply not allowed. This made sense years ago because the first generation of PLC's was not all that reliable in the first place. A stock hard wired relay generally has a failure rate of around 1:100,000. PLC's barely made the grade with that level of reliability. The only reason for the massive conversion in the 1970's from relay logic to PLC logic occurred is because this level of reliability for a PLC occurred just one time. With relays as the number of layers of logic increased, there is a rapidly decreasing reliability. For instance logic depending on proper operation of 2 relays decreases reliability to 2:100,000, or 1:50,000. There was fear of new technology. Today's safety grade PLC's are operating with failure rates down close to 1 in a billion (1,000,000,000). The limitations on safety and reliability at this point are driven by the outputs, especially valves.

In fact, according to IEEE 1584, though I am not aware of a case where this has happened yet, there is a 5% chance that the arc flash PPE will also fail. This is documented directly in IEEE 1584. This is due to the statistical "long tail" issue when it comes to predictcing arcing currents, which can result in abnormally low or high arcing currents, and resulting potential for extended trip times. Even if we take the approach to simply wear arc flash PPE, that still does NOT give an iron clad guarantee of anything. The only reason I can imagine that we have not actually found a failure yet is that with roughly 0.1 arc flashes per 100,000 workers per year (ESFI number). Out of those, we then have to consider the number of cases which are operating very close to the threshold such as equipment that has an incident energy rating of 39.9 cal/cm^2 and selecting 40 cal/cm^2 clothing. Most of the cases involve not even knowing or wearing the proper PPE in the first place. From a statistical point of view even if we were to have a "perfect storm" situation once a year, at a 5% chance, an arc flash injury of 2nd degree burns or greater would average only one case every 20 years.

This issue of absolute vs. acceptable risk is rife throughout the industry. Whether you want to use the class/division rules or the zone rules for hazardous locations, taking the division rules, division 1 is for areas where hazardous conditions normally exist, division 2 is for areas where hazardous conditions sometimes exist under abnormal operation, and unclassified areas are those in which hazardous conditions do not exist, but again, this is in a limited sense.

I could go on for quite a while, but the argument would then be:
Defendant: We cannot provide absolute reliability of any system. We can only provide reliability equal to or better than existing national consensus safety standards.
Prosecution attorney: What standard was used in this case.
Defendant: NFPA 70B, Article 200-250.
Prosecution attorney: Did you comply completely with those standards? Can you provide proof that those standards were met.
Defendant: Yes, we met all the requirements in those standards. We have PM documentation which we have submitted into evidence showing where and when the equipment was inspected to ensure that it meets those standards. We have trained employees that perform this work, and documentation of training records.

This last line is the rub...we need something more concrete, akin to NEMA AB-4 for instance, to show compliance. The big problem though is for instance with AB-4, it specifies WHAT to test and even specifies what is considered passing or failing. But it doesn't give schedules, and we don't have a standard other than IEEE 493 for maintenance intervals. These are site specific (depends on operating environment) so we kind of need a good standard for determining this as well. The rub here is that frequently failure rates are very low, frequently longer than expected operating life of the equipment. So we need a better way to track this and adjust according to known environmental factors. IEEE 493 just provides a guess (double for "bad", halve it for "good").

I wonder how the revision process of NFPA 70B and other "industry practices" will go once they realize they are now in the NFPA 70E spotlight. Will they become more focused with detail or more fuzzy and open to interpretation - making the yes/no call regarding maintenance even more difficult?

I think this comment pretty much sums it up. Most try to achieve proper maintenance and keeping equipment in good condition. The intent of this statement in NFPA 70E is a good one. However, in the legal driven lawsuit crazy society that we live in (at least in the U.S.) stating that the electrical system is "properly maintained" could come back to haunt someone if there is ever an injury.