OK, you drove me back to look again at the P&B App Note. Here is my shot at understanding what these guys are talking about:
You all know that current lags AC voltage by 90 degrees in a pure inductance under steady-state conditions. Therefore, under steady-state conditions, flux in the inductor core is at maximum negative value at the time the voltage is passing through zero. Since the basic characteristic of an inductor is that current (and, therefore, core flux) cannot change instantaneously, but changes gradually in response to voltage changes (how gradually depends upon the value of inductance), the sudden application of peak voltage does not, as implied by Sceadwian, result in current limited only by winding resistance. The current at the instant of abruptly applying peak voltage, in fact, starts to ramp up exponentially from its starting point of zero current and flux and reaches its normal peak value just as it would under steady-state conditions.
However, when you abruptly apply AC voltage at the zero-voltage crossing point the core flux is at zero, not at maximum negative value as it would be under steady-state conditions. Therefore, during the first quarter cycle, when flux would normally be changing from max negative to zero under steady state conditions, it is now changing from zero to some (mathematically predictable) positive value. In lay terms, the flux is getting a "head start" toward max positive flux compared to the steady-state condition.
Now, as the flux and current continue to increase (because the applied voltage is still positive up to end of the positive half-cycle), the flux will reach a higher value than under steady-state conditions because it had the "head start." If the core saturates due to this greater than normal flux, THEN you lose the current limiting effect of the inductance and the current may be limited only by the winding resistance. This is the cause of the up to 40X multiplication of steady-state current that the P&B App Note discusses.
So it sounds like what we really need is a peak voltage-switching relay rather than a zero-crossing switching relay for inductive loads. Of course, such a relay would shock loads that are resistive or capacitive.
awright