It's highly unlikely that one can build a directional anchor which allowed NO movement. And any movement, no matter how small, would exhibit frictional loadings.

When dealing with friction, one must constantly keep in mind that friction force vs. displacement curves exhibit a "sawtooth" form - the force builds up until it overcomes the coefficient of static friction, then it moves (with a resistive force derived from the coefficient of dynamic friction) until the dynamic friction force is sufficient to stop the movement. Once the movement stops, the static friction coefficient becomes relevant again.

Having said all that, one would expect that the clearances on a directional anchor would be "small enough" to be better modeled by ignoring friction. Fortunately, the analyst can do a simple, practical check to see if his analysis is vaild under this assumption - he can compare the various sliding restraint forces observed with (mu * weight load); if the restraint forces are larger than the potential friction loads, then friction is not relevant to the load case. When the sliding restraint forces are smaller than the potential friction loads, it might be a good idea to add friction forces to the analysis at that restraint node.

And, as I have cautioned others before, NEVER file an analysis report wher you use friction to help you in any way.

Friction is NEVER your friend!