While using Total Load at Point you select a surface and a point. Through this option one can indicate to the software to apply a load on the selected surface which is statically equivalent to the resultant load at the specified point.
The point does not have to be with in the model. It does not have to have material or any geometry associated with it.
There are many options for achieving the loading you describe. You can use Kinematic Coupling Constraints, Rigid Body Constraints, Equation Constraints, ... (there may be more).
I like to use Equation Constraints. In this case you define two node sets on the surface you want to apply a concentrated force on. One node set is a single node (driver node) and the other is every other node on that surface except for the driver node (slave nodes). Now you can define an Equation Constraint so that the load application direction D.O.F. of the driver node is equivalent to all the same D's.O.F of the slave nodes. Then simply apply a concentrated force to the driver node and you're set.
CM,
Thanks for the suggestions. I will try to find out more about each of the options you have mentioned.
In my case the driver node location is at a place where there is no geometry (no material).
In your case you may want to use Rigid Body Constraints or Kinematic Coupling Constraints. With those methods, you can define a reference point at any location and then describe how the load applied to the reference point is transferred to the actual part. Kinematic Coupling Constraints give you a little more freedom as to which degrees of freedom are "linked".
If you have any questions as to the details once you get comfortable, I'd be interested to help as it helps me expand my understanding as well. As an aside, do you work primarily with metallic parts or something more exotic like composites, etc.?
CM,
Thanks for the suggestions. I will try them out. I work mostly with metallic parts but have done some simple work with composites using pro/mechanica in the past.