You can define an alternative coordinate system, you can use normal components nx, ny, nz, and other things. Perhaps you can upload the file?

Thank you for answering @René Christensen

I read a lot in the forums and discussions about this issue, as what you said.

But I am not sure about the results and I don't know how exactly I should transform the results from cartesian to cylindrical.

Here is my file

The thing is that the radial stress in a cylinder which is under torsion loading is always zero. But I don't obtain this result.