Assign and mixed space

[cjham]

Hello,

I am trying to solve a nonlinear problem which has two scalar variables. I have set up a mixed function space
V = FunctionSpace(mesh, "CG", 1)
W = V * V

psi, alpha = Function(W)
u, v = TestFunction(W)

I need to assign an initial condition
ic = Function(W)
ic = Constant(0.0)

psi.assign(ic)
alpha.assign(ic)

But I get an error: 'Indexed' object has no attribute 'assign'. Please can you help?

[Ig-dolci]

Try with

psi = Function(W).sub(0)
alpha = Function(W).sub(1)

[JHopeCollins]

With this approach, psi and alpha would be from different Functions so you wouldn't be able to solve for the whole mixed function in one solve.

[JHopeCollins]

You'll need to give a name to the whole Function, then assign to each subfunction separately:

w = Function(W)
psi, alpha = w.subfunctions
psi.assign(...)
alpha.assign(...)

[cjham]

Many thanks for your answer. However, I'm still having difficulties. It would be good to get a very basic example working that I could use as a template. If we take the poisson_mixed example and solve it as if it was a nonlinear problem would that be a good idea? I have started this below. How would the initial conditions be set up in this? Thank you.

from firedrake import *
mesh = UnitSquareMesh(32, 32)

BDM = FunctionSpace(mesh, "BDM", 1)
DG = FunctionSpace(mesh, "DG", 0)
W = BDM * DG

sigma, u = Function(W)
tau, v = TestFunctions(W)

x, y = SpatialCoordinate(mesh)
f = Function(DG).interpolate(
    10*exp(-(pow(x - 0.5, 2) + pow(y - 0.5, 2)) / 0.02))

a = (dot(sigma, tau) + div(tau)*u + div(sigma)*v)*dx- f*v*dx

bc0 = DirichletBC(W.sub(0), as_vector([0.0, -sin(5*x)]), 3)
bc1 = DirichletBC(W.sub(0), as_vector([0.0, sin(5*x)]), 4)

w = Function(W)

solve(a==0, w , bcs=[bc0, bc1], solver_parameters=
{"snes_max_it":200,
    "snes_atol":1.0e-9,
    "snes_rtol":0.0,
    "snes_monitor": None,
    "snes_linesearch_monitor": None,
    "snes_linesearch_type": "l2",
    "snes_linesearch_maxstep": 1.0,
    "snes_linesearch_damping": 0.5,
    "mat_type": "aij",
    "ksp_type": "preonly",
    "pc_type": "lu",
    "pc_factor_mat_solver_type": "mumps"})

sigma, u = w.subfunctions

VTKFile("poisson_mixed.pvd").write(u)

[JHopeCollins]

You are still using sigma, u = Function(W) to set up the form, so you don't have a handle to the full Function. You need to pass this handle to solve so that it knows what to solve for. Here you are passing solve a Function that does not appear in the form.

Have you looked at any of the nonlinear examples? You can look over this one for a scalar nonlinear example: https://www.firedrakeproject.org/demos/burgers.py.html and this one for a mixed nonlinear example showing how to create a Function and use it in a nonlinear form: https://www.firedrakeproject.org/demos/navier_stokes.py.html

[cjham]

Thank you for your help.

[colinjcotter]

The Camassa-Holm equation example is good for this too. From: Josh Hope-Collins ***@***.***> Date: Friday, 10 January 2025 at 13:47 To: firedrakeproject/firedrake ***@***.***> Cc: Subscribed ***@***.***> Subject: Re: [firedrakeproject/firedrake] Assign and mixed space (Discussion #3953) You are still using sigma, u = Function(W) to set up the form, so you don't have a handle to the full Function. You need to pass this handle to solve so that it knows what to solve for. Here you are passing solve a Function that does not appear in the form. Have you looked at any of the nonlinear examples? You can look over this one for a scalar nonlinear example: https://www.firedrakeproject.org/demos/burgers.py.html and this one for a mixed nonlinear example showing how to create a Function and use it in a nonlinear form: https://www.firedrakeproject.org/demos/navier_stokes.py.html — Reply to this email directly, view it on GitHub<#3953 (reply in thread)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ABOSV4UKYOQWCWPGIE2JYLD2J7FPDAVCNFSM6AAAAABUV3KJQ2VHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTCNZZHAZTONA>. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[cjham]

Great, thank you I think I have it working now.