Some quick comments on the model:
There is no benefit in having the footings so finely meshed. I suggest greatly simplifying the footings and the associated soil elements, but use 8 node elements rather than 4 node, since these can model bending behaviour much better.
The elements between the footings and the structure should be concrete rather than soil, I presume.
You have plane strain elements for the soil, which is appropriate for a 2D analysis, but remember you are analysing a unit width of soil, with a structure on top representing a much greater width, so you will need to modify properties accordingly.
You have restrained the nodes around each footing, but there are no restraints to the soil boundaries, so in effect you have a structure somehow suspended in space with a block of soil hanging off it. Remove the restraints to the footing, and add restraints to the soil boundaries that are in contact with more soil (not the top surface).
I'm not a seismic expert, but I don't think the node damping at the base is necessary. You can always add it later to see what effect it has.
Did you download the Strand7 web-note on seismic analysis; it looks like it should cover most of what you need.
Doug Jenkins
Interactive Design Services