A colleague said once before "Explosion protection is sometimes done esoteric". He meant that they are overtraveling the goal by lightyears. And on the other side they neglect functional safety... in combination: FAIL
In most cases you're beyond the upper explosion boundary (ignition results in a fire) and sometimes you don't even reach the lower explosion boundary.
But I can tell you a common way I determine it when I have to assess it: Zone 2 = in case of a failure or within a extremly low period of time regularly with a maxium, summed up amount of 30 Minutes a year. Now it depends on how long we think there would be an atmosohere. 1 second maybe? --> 1800 times a year Or 1 minute? --> 30 times a year.
And in that short amound of time there have to be an ignition source. So ther're two probablities which we must multiply.
And you're right! Nobody can tell us the real probability. The data for statistics depend on various other factors. Like wind. Technical equipment. Temperature.
I don't trust any statistic which isn't manipulated by myself