We have a case of 24 header (12 bays) air cooler with 2 inlet nozzles (8”) and one outlet nozzle (6”) per header (split headers). The headers have thrust blocks in between. The end header displacement was estimated to be around 115 mm (along the header direction) in the initial stages of the project and the vendor agreed to maintain this gap.

Problem: Due to some installation error or by a last minute design mess up by the vendor the actual installed gap in all headers (including end header) is only 45 mm. If I try to input this 45 mm max displacement limit in the calculation I have expansion stress failure at extreme piping header tees. In this situation what could be the possible solutions? The ones we are exploring include….

Pl. note all the piping and air cooler is installed and hydro tested. This problem was found just weeks ahead of startup.

1)Cut all the thrust blocks and make the header move as individual units. Still to clear expansion stress I need minimum 60 mm gap at end headers.
2)Remove the guide (tube bundle direction restraint) on the headers and make them free along the tube bundle direction. This is giving good advantage.
3)Modify support locations on the piping (relocating axial stop etc). This option needs too many support modifications.

Options 2 and 3 was checked without removing thrust blocks fully but by just removing the bolting b/w thrust blocks so that they have 20 mm gap to slide like a hinge.

We have 3 times API 661 nozzle allowable, so the nozzle load is not an issue. Another point here is there is a huge difference b/w ope temp and design temp. 187 C and 270 C. The 270 C case may occur occasionally but we don’t have any further info on duration etc. The failure is in design case only.

Respected forum members kindly opine.