Because of #5882 I had a look through the cookbook to try and see what the problem was.
These are unrelated improvements (unfortunately they do not fix the SIGFPE).
The only change in the setup is that the model had particles, it tracked the initial composition with these particles, but it did not interpolated them back onto the compositional field and instead solved the equation with the field method. I did not think this was intended, so I changed the particles to be active instead of passive, which also meant I had to change the generator to avoid empty cells. Also see #5882.
Because of #5882 I had a look through the cookbook to try and see what the problem was. These are unrelated improvements (unfortunately they do not fix the SIGFPE).
The only change in the setup is that the model had particles, it tracked the initial composition with these particles, but it did not interpolated them back onto the compositional field and instead solved the equation with the field method. I did not think this was intended, so I changed the particles to be active instead of passive, which also meant I had to change the generator to avoid empty cells. Also see #5882.