Which of non-linear modifications for the Shrödinger equation did You say about? For example, we can solve the well-known k|Ψ|²Ψ -containing equation. I think it is a good idea to start with eigenvalues/eigenfunctions problem solving (small 'k' limit). Both MatLab or SciLab are useful here. Then You can translate some of routine units into C or Fortran while GUI is still working.
May be it is possible to fine-tune the AbInit program for the purpose. It is a free software. I believe the tube geometry is not too hard to be described.
If You want to use a time-dependent model, well, it is another matter. You'll need a rather powerful machine. My colleagues and me solved 1+1 and 2+1 problems for some simple quantum systems since 2001. Note many of difficulties happened in the way are of gnoseological, fundamental type, not mathematical.