Conditional expression: UFL condition as boolean

Question

Hi all, I would like to define a function (call it Func) of my trial function 'u' in a way that depends on the value of a different function (call it epsilon) of the trial function u.
My (naive?) attempt gives an error (ufl.log.UFLException: UFL conditions cannot be evaluated as bool in a Python context.)

Here is a sketch of my attempt:

class F(Expression):
    def eval(self, value, x):
        if tr(epsilon) >0:
            value[0] = tr(epsilon)**2 
        else:
            value[0] = -tr(epsilon)**2 + tr(epsilon)**2

Vv = VectorFunctionSpace(mesh, "Lagrange", 1)
u = TrialFunction(Vv) 
epsilon = sym(nabla_grad(u))
Func = F()

Can someone tell me how this should actually be done? Do I need the class structure?
Thank you so much in advance!

Answer 1

This functionality is already built in,

F = conditional(lt(0,epsilon),tr(epsilon)**2,-tr(epsilon)**2 + tr(epsilon)**2)

I doubt you will be able to solve the system with the built-in non-linear solver, i.e. you have to use a fixed point method. You could also make the system more smooth by transitioning continuously between the two expressions...