I think the wording is a little backwards from an engineering design perspective. It's a solution looking for a problem that might exist, or it might not.

I.E. it's not a "restrained joint" versus "unrestrained joint" so much as it's just "joint." Whether or not your joints need to be restrained may vary from manufacturer to manufacturer or product to product.

If the joint is as good as the pipe, per manufacturer specifications and guidelines, then you need do nothing special for the joint itself. Treat it as normal pipe. In this case, I would consider the "unrestrained" joint to be as good as a welded joint for metallic piping.

If the joint is a slip joint, then each joint effectively acts as an expansion joint with only axial movement. Total axial loading on an anchor will be primarily F=P*A. As such, it needs to be restrained, i.e. a "restrained" joint.

You can complicate things further by considering expansion/contraction of the pipe itself due to pressure and thermal effects, as well as friction within the slip joints, and friction on the external supports, but you might not easily quantify that.

Whether or not you wish to try to design slip joints in CAESAR is up to you, but I would be tempted to only consider the sustained case, under the premise that thermal effects are taken up by the slip joints, and not concern myself with the slip joints.

A properly restrained joint design will consider:
1. Anchors spaced close enough such that piping isn't capable of slipping out of its joints.
2.Anchors spaced far enough around bends such that the F=PA doesn't turn into sufficient bending stresses to overload piping/fittings.

Based on the verbiage, this looks like a client specification. Look at the manufacturer specification.