Discussion Forum

Dear all,

I am now working on implementing the magnetostrictive model on COMSOL, for which I encountered a converging problem and I'm looking for your help.

The constitutive equations I am interested in is strongly coupled between magnetic and elastic fields, as shown below:
T_{ij} = c_{ijkl}*S_{kl} - e_{kij}*H_k
B_i = e_{ijk}*S_{jk} + mu_{ij}*H_j
where T and S are stress and strain; B and H are magnetic flux density and magnetic field intensity.
c is the stiffness constants, e the strain to magnetic field coupling constants, and mu permeability.
[c.f. Journal of Applied Physics 104, 074104 (2008) ]


Basically, my model consists of a cantilevered beam surrounded by an air box which is a cylinder (in 3D).
I added "Solid Mechanics" and "Magnetic and Electric Fields" as physics interfaces.
The boundary conditions are:
fixed constraint on one end of the beam for mechanical interface.
magnetic field along the beam-length direction on the curved surface of the cylinder, and perfect magnetic conductor on the both ends (flat) of the cylinder.

The mechanical problem is only solved on the beam domain.
To enforce the constitutive equations, I added :
1° non-zero initial stress under the linear elastic material feature of the solid mechanics interface;
2° a second Ampère's law and Current conservation feature on the beam domain and chose Remanent flux density under the Magnetic field section. I then added non-zero Br components as a function of the e constants and mechanical strain (solid.el11 for example).

For the solver, I tried almost all possible combinations (direct, iterative, multigrid ...). The direct method is not appropriate because of memory insufficiency. For iterative methods, none converged, which seemed very strange to me.

Can you help me with that, please? B.t.w., I've already read the nonlinear magnetostrictive model in your model library which does not help much in my case.
Looking forward to hearing from you soon.

Thanks!