Gradient of an Interpolated Trial/Test Function

Dear community,

I am trying to add to the bilinear form the row-gradient of a vector-version Crouzeix-Raviart element after its interpolation onto the RT0 element. My code looks like this:

[code]W0 = HDiv(mesh, order=0, RT=True, dirichlet=dirichBDs)

V_cr1 = FESpace(‘nonconforming’, mesh, dirichlet=dirichBDs)
V_cr2 = FESpace(‘nonconforming’, mesh, dirichlet=dirichBDs)
fes_cr = V_cr1 * V_cr2
(ux_cr, uy_cr), (vx_cr, vy_cr) = fes_cr.TnT()
u_cr = CF((ux_cr, uy_cr))
v_cr = CF((vx_cr, vy_cr))
a_cr = BilinearForm(fes_cr)
a_cr += -Grad(Interpolate(v_cr, W0)) * wind * Interpolate(u_cr, W0) * dx[/code]

But I can not get the desired results, and after debugging we found it should be due to the “Grad(Interpolate())” is not returning the row gradient of the interpolated RT0 function. May I ask if there is any method to overcome this? Thanks in advance for any insight into this!