Some variable speed controls are enclosed in sealed enclosures, but often they are mounted in a Rittal or Hoffmann enclosure that uses external cooling airflow. Air is drawn in from a fan at the bottom of the enclosure and exhausted through a grill at the top of the enclosure. I did the calculations for a 2000Hp regenerative AC inverter. With a working distance of 18" there is a maximum of 13.21 cal/cm2. At the door of the enclosure, based on the additional distance from an arc flash source, it is reduced to 8.24 cal/cm2.

I would think that the intake and exhaust vents would just act as arc flash chutes. I would venture to guess that there are hundreds of thousands of variable speed controls in various types of enclosures installed in operation in the US alone. How would one address the arc flash hazard that exists given a vented door likely provides no protection?

I know there should be a label indicating the Arc Flash danger, but to read such label, in many cases, one would have to be within the Arc Flash Boundary to do so.

Since the enclosure is not arc resistant plus with the vents, I would not take credit for it. I would label it for 13.21 cal/cm2.


One can walk up to a panel and read the label without needing the PPE specified on it as long as you are not interacting with the device in such a manner as to possibly cause an arc flash.

The table of standard working distances in IEEE 1584 (which is probably where you got your working distance number from) are based on commonly available electrical gear. In your case since it is custom gear, you need to determine this for yourself. It is not the distance to the opening of the box. The distance is the closest distance from a worker's chest/face to the bus that you are calculating for. It turns out that panelboards, MCC's, switchgear, etc., are remarkably similar within a voltage class in terms of placing the "line side" bus bars at the back of the cabinet in about the same distances, so thus was born the "standard" table.



Frankly, yes a little, but any decent sized arcing fault also puts roughly 2 PSI of pressure on the doors, so unless the doors are designed to withstand this much pressure (almost none are), you probably don't have to worry about this. Just treat the "doors open" HAZARD the same as "DOORS CLOSED" and don't worry about it. E-Hazard does or at least used to have a video on their web site where they tested various concepts of "standing to one side". It was pretty clear that for any appreciable arcing fault, the door latches and hinges do not undergo plastic deformation due to the stress on them. Instead you get brittle fracture failure and the door pretty much flies in a nice, straight flight path directly away from the arc flash, taking anything in front of it (remember...2 PSI) along for the ride.



In numerous responses from the 70E Committee on countless proposals, the Committee has stated over and over again that just walking by is not an appreciable arc flash risk. The only reason that none of these proposals ever made it into the text is because of the inherent difficulty of scoping...specifying under what circumstances that this statement is true. For instance "just walking by" would be a much different situation if it is a production worker walking right through the middle of an ongoing troubleshooting effort with all the doors open, meters out, hands in boxes, with a clear issue of water ingress into the equipment.

Carefully read the definition of "arc flash hazard" in the definitions in 70E, especially the notes. No matter what the HAZARD is, the LIKELIHOOD is very small. ESFI ([url="http://www.esfi.org"]www.esfi.org[/url]) has carefully analyzed about the last 10-15 years of data and published results that show that on average, there is about 2 electrocution fatalities per 100,000 workers per year and about 1 arc flash hazard per 100,000 workers per year. That puts the odds at 10^-5 on a "per worker" basis. When we then factor in roughly 2,000 working hours per year divided by 8760 hours in a year, arc flashes are rare, about 1 in a million, ON AVERAGE. Incident rates for a lot of equipment such as disconnect switches is much lower but the incident rate for certain tasks such as inserting/withdrawing gear from an energized cell, or messing around with energized wiring where the worker's skill and reliability control the outcome, or poorly maintained equipment, especially circuit breakers, clearly tilts the odds in favor of the occurrence of an arc flash. In the case of what is typically done for maintenance on drives, you are probably closer to the "1 in a million" odds, a statistic which is generally considered acceptable in most circles. However you should perform a risk assessment (determine both likelihood and hazard) before going with that number. Note that the 1st draft of the 2015 edition of 70E already specifically requires an arc flash risk assessment, not just a hazard assessment, and this language is highly unlikely to revert back to the old requirement of only a hazard assessment.

Paul,

Thanks for your comments.

The reduced 8.24 cal/cm2 does take into account the distance from the arc flash source inside the panel to a workers face and chest while standing in front of the enclosure i,e, 18" from the door.

Yes, I realized after posting that one would have to be within the Arc Flash Boundary to read the label that no hazard would exist just in approaching the enclosure.

If the equipment was in operation, would interaction include a contactor operating remotely? Also, we do often have touchscreens on our enclosures, so it is possible to interact with the controls with the doors closed.

I do realize that the liklihood of an arc flash incident is statistically remote, but when they do occur the results can be very bad. I have been much closer to a couple arc flash incidents than I really cared to be and escaped unscathed.

In one case it was just dumb luck in my case though. I have always tried to practice safe techniques like never standing directly in front of a breaker, disconnect, variable speed control, or contactor when it is closed or started. I have tried to be a voice of caution to field service engineers and repair technicians.

There is a thought in some circles that we need to move away from controls directly on starter buckets, especially when you can be at up to around 21 times full load current with energy efficient IEC rated motors during inrush conditions. All it takes is an arcing fault from say a loose connection. But this gets into speculation and conjecture. IEEE 493 provides failure rate data some of which indicates failure mode such as arcing fault, and other data which shows for instance failure under the operating condition such as during opening or closing. What is missing is BOTH. I suspect that for instance all disconnects are most likely to fail during opening with an arcing fault. But I suspect that the vast majority of contactor failures are going to occur during motor starting but the two tables are not merged. So we have to often be conservative and assume that all arcing faults only occur at the most inopportune time. If this is done you will quickly notice that most equipment arcing fault rates in that data set are around 1 in 100,000, give or take an order of magnitude, under average conditions. This also indicates that the data seems to be close to what ESFI data indicates, so use of IEEE 493 data appears to be a valid approach. Without collecting data on a large scale and quantifying specific failure details leading to arc flashes in a meaningful way, this is probably as close as we will ever get, because with failure rates approaching 1 per 1, 000, 000 per year, you need roughly 10 times that (10 million samples per year) to develop specific, meaningful, and accurate data. By the time that kind of data is collected, technology will evolve past it. For instance breakers are starting to come out with magnetic actuators that do not contain any springs and only one sliding bearing and only one moving part compared to 2 or more sliding joints, several rotating joints, and roughly 20-30 parts total. Also lubes have moved from Red grease with a requirement to exercise annually to fluro based greases with 5-8 year exercise times. It is getting to the point where breaker maintenance may move to test every 8-10 years and replace at 25-30 years, for a total of 3 breaker service jobs over its entire operating life, compared to designs of 20 years ago which required annual maintenance to maintain them properly.

I guess I didn't answer the thought. I have moved away from placing personnel in front of equipment when practical. Again, this is all a rare event but decreasing exposures especially when equipment is changing states (energizing or deenergizing) is a cheap way to decrease exposure whenever practical. On cheap MV mining 600 A vaccuum breakers, costing around $4, 000 US, motor operators run around $200. That is what it costs to move the operator out of line of fire since we already run microprocessor protection relays with plenty of IO.

I don't disagree with you, but inverters are different from starters. An inverter enclosure will have a breaker and a contactor, but the inverter is between the motor and the contactor. There certainly is opportunity from arc-flash with an inverter, but it really depends on where the source is.

As long as there are no loose connectors, if there is a short across the motor leads, an inverter will generally trip immediately when you start, indicating an I2t or instantaneous overcurrent fault condition. It is entirely a different matter on the input side to the inverter because the inverter has no control over shorts that occur prior to its control. And if there is something loose that drops down across the BUS while it is running, you will get an arc flash as well.

I have seen inverters come in for repair that didn't trip until the loose connections melted the terminal strip and solder to the power board, and still never flashed over.

On the other hand, I have also seen inverters explode spectacularly from things ranging from fly-back diodes installed backwards to squirrels building a nest inside the inverter during an extended downtime.


But, I agree with your premise - I don't like to see someone operating controls directly in front of an inverter, either.

One nit to pick here. Virtually every inverter manufacturer I use these days requires inverters to NOT be operated from a contactor. I have also seen excessive failures from doing so. Almost all require or recommend some sort of high speed fusing on the smaller ones where it is not built in. Even then as long as the drive is set up to do current limiting and stall/overcurrent tripping, the only failure mode resulting in blown fuses is a dead short device failure in the front end, which is not very common.

I understand what you are saying. Just slapping power to any drive that is off, having the controls power up and the contactor close in an uncontrolled manner is just asking for problems (Doing this with an armature contactor on a DC drive is a good way to blow up SCRs).

Our inverters and DC Drives have contactors that disconnect the power bridge from the incoming power. The inverter (or drive) is powered by control power prior to energizing the contactor, and the output to the motor is not enabled until after the contactor is verified as closed.