Okay so I did it, it's a bit sloppy (like this forum post is going to be, because I accidentally deleted the previous one), but the results seem consistent, so without further ado:
Engine size doesn't matter.
It's linear.
See the following graph (mean explosion damage vs. engine power):
The equation is:
mean explosion damage = 0.12 * EP + 0.81
And the standard deviation is likewise linear (I can pull up a graph if you guys want) and the equation is as follows:
stddev explosion damage = 0.07 * EP - 0.11
Also, as the previous post says, military and commercial engines might be different (I can test if people are curious), but I did all military engines.
I can also put the actual data up here if you guys want, but the means were all calculated from many explosions (between 30 and 336), most significantly more than 30.