Annex F in the 2012 Edition of NFPA 70E was greatly expanded and provides a detailed method for performing a risk assessment including risk estimation and risk evaluation.

Have you used the approach in Annex F of NFPA 70E to perform a risk assessment?

• Yes
• No

Far too many issues with the procedure as presented. It is far away from the methodologies and scoring that others (ANSI recognized or IEC) use. I'm not clear at all what the underlying basis for it is. It looks like something that somebody just threw together. for instance most of the human performance methods (HEART, etc.) come up with wildly different scores.

In addition as I've posted in the past, I have huge problems with the idea of reducing PPE. The IEEE 1584 procedure calculates an incident energy. Unless you are reducing this, you can't realistically reduce PPE. It comes down to either the exposure is below your risk tolerance (likelihood of occurrence is acceptably low) in which case no PPE required, or not in which case PPE sufficient for the hazard is required.

I don't see any philosophical problem of reducing PPE for unlikely events. The calculated IE at a location is not an absolute, deterministic value; there is some probability involved. The working distance is not always what you estimate so there is some probability that it is larger with lower IE (it could also be smaller with higher IE). The IE formulas are based on regression analyses, with inherent variability. There could be reduction of IE from panels, doors, etc, shielding the worker.

If an event has a low probability of occurring, there could be a lower PPE that protects for the same percent of the time as a higher PPE for a more likely event.

If the IE was higher than the PPE rating, there would still be some protection. It could be considered acceptable to have more severe injuries for a very unlikely event. There are also probabilities involved in the protection with a particular PPE. 8 cal/cm² rated material does not fail exactly at 8.001 cal/cm²; there is a range of protection.

It is not as great as expected. I've run the numbers estimating fatality vs. second degree burns and posted the results. If you simply project the distance out so that it covers a circle, the edge will necessarily have to be 1.2 cal/cm^2. If we pick an arbitrary area for the circle, we can back-calculate the radius. The human body has an average surface area of 1.46 m^2 so if we have some idea of what percentage of the body we have burns over AND assume that it is not so substantial that we can assume a flat surface then we can estimate the size of the circle. Based on light reading about concepts like the rule of 9's, 10% seems reasonable. Then if we know the assumed working distance then we can use Pythagorean theorem to calculate the distance to the edge of the circle, and then using the denormalization formula in IEEE 1584, arrive at the incident energy at the center of the circle.

Whew! That's a mouthful and I've posted the results previously. With a 455 mm working distance (600 V MCC, the lowest working distance given in IEEE 1584), the resulting incident energy given 2nd degree burn coverage over 10% of the human body is 1.46 cal/cm^2.

In addition, the estimated ASTM value for cloth samples is not exactly a "range" as expected. Carefully reading the test method says that the cloth samples are all rated either "pass" or "fail". A number of samples near the rated value are statistically fitted to a sigmoidal curve. But if you carefully study the actual raw results, a value roughly 0.5 to 1.0 cal/cm^2 above the calculated threshold results in 100% passing whereas the sigmoidal curve chosen only approaches 100% asymptotically.

For these reasons, although there may indeed be a range of protection, the range is so small that since PPE reductions are normally not in the range of tenths of a cal/cm^2, there is effectively no difference between "no" and "reduced" PPE.

As I've also posted previously though, there is ample evidence to suggest that whatever one's risk tolerance, many circumstances have such low probability of injury that they would be acceptable risks. Two cases that the 70E Technical Committee have rated as such are working on 120 V control circuits, and opening and closing disconnects with the doors closed and latched. Having done similar independent analysis myself (because my employer has their own established risk assessment procedure), I've come to similar conclusions.

There are already a number of published standards for risk assessment procedures that are already well vetted such as IEC 61508, IEC 61511, The RIA standard, and the ANSI B11.TR3 standard, to name a few.