More intelligent zone merging

[mcgratta]

We now have logic to compute the volumes of pressure zones at startup. In the MERGE_PRESSURE_ZONE routine we are looping over all wall cells looking to see when zones merge. We could have another array like CONNECTED_ZONES only a REAL array, maybe MERGE_AREA(N_ZONE,N_ZONE), where we sum the area that is connected. Then rather than having the user specify a relax time, we could use the pressure difference, area, volume, assume orifice flow, and compute a time for that merge.

[drjfloyd]

The initial mass flow rate is:

$\frac{dm}{dt}=C A \rho \sqrt{\frac{2 (P_H-P_L)}{\rho_H}}$

where H is the high pressure side and L the low. I did some hand calcs for different combinations of zone volume and initial pressure difference (assuming ambient temperature). The flow rate decreases in a near linear fashion, so we could estimate the merge time by dividing the initial flow rate by its derivative. To make things simpler could assume ambient temperature and density in each volume. The pressures difference cannot be more than ~5000 Pa to keep things low Mach so assuming the same temperature in each volume, then the densities are also essentially the same.