Jim,
It is normal that Type997 is slow; it is solving a finite difference problem for the entire ground field. The model will run particularly slowly at the beginning of your simulation but will increase in solving speed to some extent as it gets farther in (and as the changes in the ground field from one timestep to the next diminish). You can speed the model up to some extent by increasing the size of your smallest soil node and increasing the value of the node size multiplier. The point of Type997, though it to provide a rigorous and stable solution. Type952 models a single buried pipe rather than a network of pipes. We use it (or its twin-conductor cousin) to model the runouts to and from the ground field.
You have to be a bit careful if if your ground field is actually
buried air ducts rather than pipes containing a liquid. Both
Type952 and 997 make the assumption that the thermal properties
(density, specific heat, etc.) are constant over the working
temperature range, which isn't always a great assumption with air.
This is particularly true if there is a chance that the water
vapor in the air condenses as it goes through the ground field.
There are models available (Transsolar's Type460 for example) that
are designed to model underground hypocausts such as the one it
sounds like you are modeling. If indeed air is your working fluid,
I would recommend using one of those models in place of either 997
or 952.
kind regards,
David