cap4000: My w3 equation is written positive upward, which means if it gives a negative result (which it does), then the force is downward.
By the way, using a rigid-body mechanics solution, such as in my previous post, is reasonable only if the beam is relatively stiff compared to the embedment material. If, instead, the beam is relatively flexible compared to the embedment stiffness, the load will transfer more rapidly to the embedment material, such that embedment beyond a certain distance becomes ineffective. In other words, if the beam is relatively flexible, then one can't ignore elasticity and compliance, and the problem becomes more complicated, in which case you could use, e.g., finite element and surface contact to solve the problem.