-
-
Notifications
You must be signed in to change notification settings - Fork 667
Description
Description
We now have three options for modelling open-circuit potential hysteresis in PyBaMM: 1) "current sigmoid"; 2) "Wycisk"; and 3) "Axen", all of which have slightly different implementations.
Both "Wycisk" and "Axen" are single-state hysteresis models which contain a single ODE for a hysteresis state variable, h
, of the form dh/dt = K(1-sgn(i))h
, but again with slightly different implementations. The current sigmoid model is a zero-state model (the OCP just jumps between branches depending on the sign of the applied current), and doesn't track the variable h
We propose we unify these models by (see #4332 (comment)):
-
Making sure all hysteresis models use the state variable
-1<h<1
(in the current sigmoid modelh
will just jump from -1 to 1) -
Add a base one-state model that solves a general one-state equation:
for
- Allow a scalar input or arbitrary functional input for
$\gamma$ , or allow$\gamma$ to be computed using the specific Wycisk formula with parameters$K$ ,$n$ and the OCP:
where the evaluation at a reference point (implicitly at
- Allow general (particle-wise) initialisation of the hysteresis state, not necessarily = 0.
Unifying the models will also allow us to more reasonably include hysteresis heating (see https://github.com/About-Energy-OpenSource/AEPyBaMM/blob/main/src/aepybamm/pybamm_tools.py#L91).
Motivation
No response
Possible Implementation
No response
Additional context
No response