You may be right about the pulses needing to be narrow and strong in
order for most noise blankers to work on them. I do not think you are
correct about them needing to have a regular repetition rate. For
example the noise blanker in the Omni VI has a detector to trigger the
generation of a blanking pulse, and a switch in the IF signal path
controlled by the blanking pulse, which performs the blanking. There is
nothing in the Omni VI noise blanker that requires repetitive noise
pulses in order to work. It can blank a single noise pulse as long as
the pulse is strong enough to trigger the blanking pulse.
There are noise blanker circuits that are designed to work on repetitive
pulse type noise, particularly 60 and 120 Hz repetition rate noise
pulses produced by AC power lines (or 50 and 100 Hz in the other
hemisphere) These circuits are typically synced to the the AC power
source using a sample directly from the mains rather than from the
receiver IF, and can blank the receiver even when the noise pulse would
not be strong enough to trigger a blanking circuit such as in the Omni VI.
There are other types of noise blankers which use a whole separate
receiver, working at a different frequency from the one the main
receiver is tuned to, to sense noise pulses and trigger the blanking of
the main receiver. These also do not require repetitive noise pulses to
work.
NO ANALOG NOISE BLANKER HAS EVER WORKED ON ATMOSPHERIC NOISE. The
noise characteristics its made for are ignition noise and line noise
that have a regular repetition rate but are made of isolated narrow
but VERY STRONG pulses. Unless you have a thunderstorm close by, you
don't get similar pulses from the static crashes and the noise blanker
doesn't see big enough or narrow enough pulses to work to cut them out.