You probably should divide all those number of days estimates by at least 2 or 3, because in many (most?) places the times when power failures are most likely a

You probably should divide all those number of days estimates by at least 2 or 3, because in many (most?) places the times when power failures are most likely are also the times when households need the most electricity. For example here in western Washington outages are most likely during winter, which is also when we are using the most electricity. My daily winter use is 3.5-4 times my daily summer use, perhaps even more on those years when we get a week of near record cold.