VUMAT for Nitinol --- why VUMAT and not UMAT?

Dear all,

 I'm new in imechanica. I need some informations about VUMAT for Nitinol. I would like to implement in fortran a subroutine for the behaviour of shape memory alloy.I looked for some subroutines already implemented but they have been developed in the implicit way.

It's true that with the VUMAT(explicit) I could have some problems of stability and convergence.Could someone suggest me any reason to make this model in the VUMAT?I know that in VUMAT I can't redifine the time increment so before running the analysis It's important to pay attention on Dt.

Is it better for dynamics analysis or for coupling with CFD software like FLUENT?