Overhead line insulation will generally be self-restoring. The BIL of such insulation is defined as the crest voltage of standard waveshape, at which a 10% probability of flashover exists. However, it is more common and appropriate to use the critical flashover voltage (CFO), at which a 50% probability of flashover exists.
Numerically, the BIL and CFO values differ by approximately 5% for lightning impulses. Because the S-curve describing the insulation strength is so steep (low value of s/CFO), it is common to take the CFO as single-valued line insulation strength.
That is, the probability of flashover is 0% for crest voltages less than the CFO, and 100% for crest voltages greater than the CFO. For switching surges, the s/CFO value is higher and the flashover probability should be treated as a function of voltage.
The line insulation strength is also a function of the wave front and tail. In system studies of arrester applications, these time-dependent effects should be modeled using a volt-time curve, a destructive effect model, or a leader progression model.
With self-restoring line insulation, it is not appropriate to use a fixed coordination current to calculate protective margins. Instead, the probability distribution of lightning stroke currents is applied to the line, and the probability of flashover is calculated.
Combined with local ground flash density, this will produce a lightning flashover rate per unit length per year. This flashover rate is the lightning performance metric for the line. For transmission lines, a typical target might be 1 flashover per 100 miles per year.
For distribution lines, a reasonable target may be somewhat higher. The effect of a flashover depends on protective relaying practices and the possibility of arc quenching by wood. With line arrester applications, the arrester failure rate should also be considered in the line performance.