Pe degrade

[mzzhuo]

Description

(Draft for reviewing) Add the core-shell submodel for phase transition caused PE (NMC811) degradation.

Fixes # (issue)

Type of change

• [ ] New feature (non-breaking change which adds functionality)
• [ ] Bug fix (non-breaking change which fixes an issue)

Key checklist:

• [ ] All tests pass: $ python run-tests.py --all (or $ nox -s tests)

Further checks:

• [ ] Code is commented, particularly in hard-to-understand areas

[kratman]

@mzzhuo Are you still working on this?

[mzzhuo]

@mzzhuo Are you still working on this?

From my side, the submodel is done, not working on it now but waiting for the feedback from pybamm team. I think Feran and Rob are reviewing the code

[kratman]

@mzzhuo I was primarily asking since the CI has been failing for this PR. Can you update the branch to resolve some of the issues? That needs to be done before it is merged anyway

[mzzhuo]

Hi MZ, thank you so much for writing this code, which must have taken a huge amount of time! Robert Timms and I made some comments in specific areas of the code, but I also have some general comments that must be addressed before we can proceed any further.

Firstly, the model is still incompatible with the particle mechanics submodel as things stand. You can either accept this and add code to base_battery_model that prevents users from trying to use both, or alter the code for particle mechanics so that they are compatible.

Secondly, I don't understand why you need a separate set of "cyclable lithium" variables. If the lithium in the shell phase is dead with no posibility of recovery, then all you need to do is arange it so PyBaMM does not count the lithium in the shell when calculating "Total lithium in particle [mol]".

That said, both Ghosh et al. (2021) and Zhuo et al. (2023) are rigorous pieces of research that deserve to be included in PyBaMM. I also like how you've added a "positive electrode degradation" option, so that more models like yours can be added in the future.

Hi Simone, thank you so much for your reviewing. 1) yes, incompatible at the moment, we can prevent the use of both submodels now for simplicity. 2) what you said about "Total lithium" is what was defined and calculated in the paper with LLI_tot, see Eq. 16 and Fig. 4. I guess you wonder why we bother to define an extra variable called cyclable lithium. OK, when we were discussing the degradation model, Greg had an assumption based on mass conservation at the interface, that is, the total lithium may increase and not always decrease. This scenario is only possible with the definition of cyclable lithium. The results in Fig. 4 with negative LLI_cyc shows that the model (particularly the mass conservation at the interface) and its implementation are validated, this is what we should expect. Besides this, LLI_total is always positive, i.e., we always lose lithium, but this info. is not that relevant to lithium shutting between the two electrodes. The stochiometry change from 0 to 100% corresponds to the cyclable lithium shutting. Note that in Figs. 5, 7 and 9 the results are all about cyclable lithium so that we can see the subtle difference between different scenarios.