Hi,

I am presently undertaking a study to quantify arc flash hazards in our 11 kV and 400 V sub-networks. This is mainly focused on the danger to contractors working on panels, transformer kiosks, etc.

For this purpose I am attempting to use formula #8 given in the paper "Improved method for arc flash hazard analysis" as I have heard this gives more accurate results than the NFPA 70E equations.

However, I am unsure of two parameters. Firstly, the reflected energy term. The paper says this is calculated using radiative view factors based on the dimensions of the box. In this case the dimensions would be about 400W x 600H x 300D (mm). Can anyone tell me how to calculate this?

Secondly, I obtained the arc energy by halving the fault MVA (as suggested in 70E) and multiplying this by the arc duration. Will this yield accurate enough results?

Thanks.

Wilkins Formula

I have never heard of this method, maybe someone else here has.

Alan

I havent heard of it either but am interested to learn about it.

My understanding of this paper was to develop a time domain model to help better represent current limitation. Mike Lang (from Shawmut) was also one of the authors, I know Mike from the 1584 committee - very sharp person. Since the 1584 equations are valid up to 15 kV, why not use them since they are widely accepted?

In the paper, they were very vague (as you already know) about the reflectivity they refer to as alpha. I think they were more concerned with developing a different current limiting model by taking into acount the dynamics of current limitation and its effects on the plasma in an enclosure.

I'll see Mike at the next 1584 meeting in February, perhaps I can bring this up then.

For those of you who haven't heard of it, the paper is online here:
http://us.ferrazshawmut.com/resources/media/articles/ICPS-wilkins011904_web.pdf

The Wilkins model is a very promising model that seeks to include more of the physical aspects of the arc behavior into account. It uses both the physical parameters, and is corrected for the statistical analysis. In this way, he is able to achieve better results over a wider data range then the 1584 equations.

This model also looks to be the route that the joint IEEE/NFPA taskforce is taking, based on a paper written by Tammy Gannon and John Matthews in the July/Aug 2008 edition of IEEE Industry Applications Magazine.

I've been a fan of this method for some time, not only because it claims to produce better results, but because it introduces a physics aspect back into the equations that were missing in the 1584 statistical analysis.

tss26 - I would recommend going ahead and using the 1584 equations for now, as they are the industry accepted method, but look to revisit your analysis in a couple of years with the Wilkins model, or some variation of same. Good luck!

There are several equations given in NFPA 70E and they give very different results. I believe the empirical equations (D.8.3(c) and D.8.5(a)) give the most accurate results; the others seem a bit dubious. I would like another equation to validate the figures given by the above equations, if possible.



The paper states: "The only unknown is the reflectivity alpha. By varying alpha and computing the correlation between the predictions of (8) and the test data, the optimum value of alpha was found to be 0.56"

The notation seems to suggest a single effective view factor (a function of alpha) that takes into account single and multiple reflections involving all five sides of the open cabinet. Yet the paper gives no clue as to how this may be computed. It only suggests that the reflected component makes up approximately half of the incident radiation in a typical case.

I haven't bothered with any time domain analysis. I just wanted something an equation to give me a second (well, actually third) opinion on the incident energy since the other equations gave such different results.

To date, the 1584 equations are the best accepted method. I say that because, as I stated, I think that the Wilkin's model is an improvement, and will be looked to for future equations. To date, this method has very little industry acceptance. You should still probably stick to the 1584 equations until something better comes out. The Wilkins model also involves more data analysis and input into the equations. No software program that I know of (except Wilkin's own...) to date uses these equations. You will need to know X/R at the fault point, as well as the box dimensions.

The Wilkins model gives similar results to the 1584 equations, with respect to arc energy, largely because he does normalize his equations against the 1584 test results, using the a and k factors.

You should also get a hold of the second paper, "Simple Improved Equations for Arc Flash Hazard Analysis."