Hi everyone,

for a mesh convergence study on a non-linear problem, I want to use the coarse mesh solution as a good initial guess for a Newton method after mesh refinement. However, for order>1, the prolongation appears to modify the solution, (so that the GridFunction on the fine space is the prolonged P1 function from the coarse mesh). I may have misunderstood something, so is this behaviour expected?

Best wishes,

Henry

```
from ngsolve import H1, Mesh, GridFunction, x
from netgen.geom2d import unit_square
from ngsolve.webgui import Draw
mesh = Mesh(unit_square.GenerateMesh(maxh=0.2))
fes = H1(mesh, order=2)
gfu = GridFunction(fes, nested=True) # nested=True to do the prolongation
gfu.Set(x * (1 - x))
Draw(gfu)
mesh.Refine()
fes.Update()
gfu.Update()
Draw(gfu)
```