The paper describes having a telescoping section on each end to accommodate thermal expansion, but the previous commenter is stating that it is likely this telescoping method will not be sufficient, as the variation in length of the structure will cause significant bending of the pylons themselves. Between the ends, there is no mechanism to allow for thermal expansion, which will either cause massive axial stress on the steel tube, bending of the pylons, or both. The shock absobers will not have sufficient travel to address this, (100s of meters); axial bearings on the tube may be required.
Each 30M long tube expands 15mm over 40C of temperature change. You could imagine a 15mm expansion joint at each connection, which is roughly what oil/gas pipelines do, but the paper explicitly says tubes will be welded together for straightness.
