Sorry about asking six more questions but you will need to test the design important parameters. If possible you should test the device itself.
1. Will high pressure fluid, (such as water), be confined inside the pipe by the seal, while being used to test the porosity and integrity of the weld? Will you use a visible dye or a fluorescent chemical? A pinhole in a weld will produce a fine jet of fluid that will cause physical injury with a subsequent infection to anyone who encounters it. If oil is used as the testing fluid the resultant infection will probably be gangrene. You should surround the weld while it is under pressure test with a steel clam shell to reduce the risk of injury to the inspectors.
2. Was your design with the 3mm radial clearance designed to pass the irregular internal surface of the finished weld after testing?
3. Seals come in different types. Some must slide while expanding. Others less often, must grip with static friction and lock in place under pressure. Am I correct that you want to measure static friction while the seal is locked in place?
4. There are ways of using mechanically expanded deep cups to static seal a pipeline. You appear to be considering only a cylindrical plug without any cup. Would you consider cup seals?
5. Will the weld testing device have one seal, behind the welded joint, with pressure applied from the open end of the pipeline, (where another seal is needed)? Or will the testing system have two seals, close on each side of the welded joint, with pressure applied through a small diameter, long tube? If you select two seals on the same tension rod, the force on the two seals will cancel and friction will be less important. It will then be a case of expanding the elastomer against the pipeline wall rather than relying on static friction to hold it in place.
6. Your 3mm radial clearance, (6mm gap on diameter), presents a significant but tractable design challenge. Hydraulics systems with significantly smaller gaps prevent seal extrusion by employing a 'back-up ring' or an 'anti-extrusion device' Google those and similar terms.
The experimental testing jig for a single axis UTM should be based on the design of the real world device.
I have a few ideas for design of the UTM jig, but we need to further restrict the possible device configurations.