Fix Heisenbug in ode-brusselator notebook

[Da-Be-Ru]

It seems that some testing CI runs crash somewhat irregularly with issues concerning NaN assemblies in the notebook tests (here's also a past such occurrence). I have as yet been unable to discern a pattern and these seem to show up and sometimes just vanish again.

That's also why this specific notebook is currently not being run in the test set (see #188).

Can we perhaps figure this out?

[j-fu]

I think Brusselator, though it is fun, is not a good example for testing due to the fact that the solution can be heavily influenced by small changes in the initial value. So for demonstrating how to work for reaction-diffusion problems, some example with unique stationary solution would be better.