Erik Kostson
Ansys Employee

Not sure, but the below approach might help.
Perhaps use a thick block on top of the beam to transfer the load with onto the beam (via a frictionless or no separation contact say - frictionless transfers only compression to the beam and no separation transfers both tension and compression of the sinusoidal force). Use then a remote displacement to translate that block on top of the beam and a force that can contain a sinusoidal variation - also use transient dynamic if that is important (the below schematic setup is just for reference just to show a possible way of doing this)
Or even easier and better is to take away the force and use an enforced sinusoidal displacement (just tune the magnitude to get the correct force) since that will be much easier to get to converge (sinusoidal force will be an issue if we have frictionless contact, force will though work ok with no separation contact).

All the best