There's an issue in convergence after adding the convergence scheme for Anderson acceleration, the simulation keeps getting stuck after the 4th iteration.
It doesn't give any errors, and just keeps going until kernel interrupt/restart.
Also the dilution factor shows up negative on the 4th iteration.
To Reproduce
This bug can be reproduced using #2737.
Go to the "convergence_test.ipynb", and execute the 2nd cell. (The .yml file is named tardis_anderson.yml)
Description
To Reproduce
Screenshots
System