STAAD.pro Global Statics Equilibrium Problem

Hello,

I am having some problems with a model I am trying to run for deflection purposes.  The structure is a steel plate girder bridge with a composite concrete deck.  I originally had the full structure modeled using beam elements for the girders and plate elements for the concrete deck.  To make the connection between the girder and deck I used rigid master/slave body constraints.  To verify the model was working properly I first ran it with only the 8" thick concrete deck weight applied.  After the model was run I went to the post processing and looked at the statics equilibrium check and saw that the vertical load equilibrium was not being satisified for some reason and the sum of the reactions was about 15% less than the applied load.  I went back and checked all my inputs and structure geometry to make sure there was not any error's or my part and from what I see everything looks ok.

In an effort to make trying to figure this issue out easier, I simplified the model to one girder with it's tributary deck width.  Again, I ran the model with only the selfweight of the 8" concrete deck applied.  As with the full model I got the same results where the sum of reactions was about 15% less than the total applied load.  One thing I noticed with the model is that even though the applied selfweight load is in the -y direction, when the model is analyzed it results with a reaction force at one of the supports in the x-direction, which should not be the case.  I have been through this model several times now and cannot figure out why this is happening.  If anyone has had this problem before and know what may be causing, could you please let me know.  I am attaching my simplified model in which am having the static equilibrium problem.  Thanks.