I see few problems in the original scheme.
1. If I consider 15cfm fresh air per student then the fresh air is around 525cfm. You may require about a square feet leak area to push this for a room pressurisation of 5Pa. Pushing through doors may not be a possibility because you may have to have 60sq.ft to 100sq.ft doors depending upon the door construction.
2. As you are mixing the air in the room, i.e after the recirculating air being treated, the mixing may happen anywhere from the coil off temperature to the room condition. If the recirculated air is at a temperature less than that of dew point of ambient air, you will have condensation. So, moisture removal from the ambient air using the window A/C should be avoided and I deviate here from AbbyNormal's suggestion.
3. Homogenous mixing of fresh air with treated air is a problem.
If this an inevitable situation, I would use either of the two methods.
First one is to find out the grain difference by required humidity control(only room latent load) and then dropping both fresh air and recirculating air temperatures to a common ADP and then using combined cfm to plot ESHF line.
Second one is to draw ESHF line based upon only recirculating air flowrate and then cooling fresh air only to room condition.
Both these methods require treating fresh air separately.
AbbyNormal said:
Subtract how much moisture is being displaced through the exfiltration, theoretically this displaced air is at the design condition.
As the leaking air is at room condition, I think, moisture load on the A/C will not be reduced. You can only reduce moisture load on the A/C if the leaking air contains more moisture than the room condition.