Effect of Arc Electrode Geometry and Distance on Cotton Shirt Ignition

The results of experimental arc testing of a nontreated cotton T-shirt with three different geometrical arc electrode arrangements are presented and discussed. Different geometries of arc electrodes result in different electromagnetic interactions of an arc current with current flowing in arc electrodes, which, in turn, cause very different arc behaviors. The ignition threshold of a T-shirt (arc energy attributed to 50% probability of ignition) was determined using the following: 1) apparatus for mannequin arc test methods ASTM F2621 and IEC 61482-1-1 (Method B) with vertical butted arc electrodes in open air (open arc); 2) apparatus for arc-in-a-box (box) test method IEC 61482-1-2 with vertical butted arc electrodes placed in a box; and 3) nonstandard end-of-line (ejected arc) apparatus with parallel horizontal arc electrodes in open air. Test parameters for the ignition threshold included the following: 1) 4-, 8-, and 16-kA arc current levels for each test apparatus; 2) same distance of 18 in from arc electrodes to the tested T-shirt for all apparatuses; and 3) same arc gap (distance between arc electrodes) except the arc in a box. A strong dependence of the ignition threshold on the electromagnetics was demonstrated by the tests. The T-shirt ignition for the ejected arc demonstrated a longer arc flash boundary distance than that calculated by NFPA 70E or IEEE 1584 calculations, raising a need for strong caution in using non-fire retardant (FR) cotton in arc flash conditions. It is suggested that the electrode configuration should be considered in risk analysis. Reported test results could provide a framework in developing arc flash boundary for ejected arc applications. It is shown that, unlike most cases where the IEEE 1584 calculations and NFPA 70E tables were observed to be conservative, in the ejected arc, calculation results should be used with caution as they are less than conservative.