Let's assume for a moment, that all merits generated in powerplay are generated by rating 5 pledges, who are only making merits until they have enough to maintain rating 5. Let's assume that this means they'll do 6,000 merits each per week. This will give us a reasonable lower bound for the number of players in powerplay. For support (preparations, expansions and fortifications) you can even break it down into the number of pledges per power.
This gives us the following distribution for cycle 14:
Support
Edmund Mahon - 60
Arissa Lavigny-Duval - 510
Aisling Duval - 85
Denton Patreus - 50
Zemina Torval - 28
Felicia Winters - 45
Zachary Hudson - 208
Li Yong-Rui - 35
Pranav Antal - 44
Archon Delaine - 47
Major Power Allegiance:
Alliance - 60
Empire - 673
Federation - 254
Independents - 125
In terms of offence (opposition and undermining) it isn't possible to do a reasonable distribution assumption, as the only reference point (support) will change dramatically as powers cycle in and out of turmoil.
But for cycle 14, the number of rating 5 offensive players supported is 1,201
Combined this means that we can reasonably assume that there were an absolute minimum of 2,302 active powerplay participants for cycle 14.