Hello,
From an older topic I know that the internal radii seems to vary between 3.6 and 3.8mm.  The outer radii varies between 4.8 and 5.2mm.  That has been measured by hand. Is this information still true today (any process improvement?). 
What about the tolerance of the 20 mm of the side of the tube ? Is it +/- 0.1 mm ? In my application the tube shall slide inside the groove of roller which looks like this one <https://www.aliexpress.com/item/10-56-31mm-belt-groove-H-type-injection-molding-machine-safety-door-roller-square-groove-D10/32816622292.html>. The groove profile shall fit the tube side profile as close as possible. 

First of all, your diagram isn't  in balance, as the leftward force isn't equal to the rightward
besides that, if the diagram is accurate, this might work, as 220N isn't that great force for a 20mm tube. But usually there is some sort of radial forces, especially when the axial force is dynamic (you are speaking of a force peak).
The problem is, that pultruded tubes have no fibres in a direction to take the sheer forces under radial load. As long as you are shure that there is no load perpendicular to your tube axis, this might work indeed.