PyBaMM cannot get solution when changing the active material volume fraction

[XuboGU]

Hi!

I found the PyBaMM cannot get the solution when changing the ‘Positive electrode active material volume fraction’ or 'Negativ electrode active material volume fraction'.

Taking Ai2020 parameters as examples. For the positive electrode, when multiplying the initial PE active material volume fraction by a factor < 0.91, we can not get the solution.
For the case of changing NE, it can be more complex: multiplying by 0.4 or 0.3 does not give solutions, while 0.39/0.38 etc works.

The pybamm version in this test is 25.1.1.

Here are the codes. Thanks!!

import matplotlib.pyplot as plt
import pybamm

params_list = pybamm.ParameterValues("Ai2020")
# print active material volume fraction of PE and NE
print(params_list['Negative electrode active material volume fraction'])
print(params_list['Positive electrode active material volume fraction'])
eps_ne = params_list['Negative electrode active material volume fraction']
eps_pe = params_list['Positive electrode active material volume fraction']

def chemomechamodel_spm(input_val):
    model = pybamm.lithium_ion.SPM(
        options = {
            "particle": "Fickian diffusion", 
            "cell geometry": "arbitrary", 
        }
    )

    experiment = pybamm.Experiment(
        [
            "Charge at C/3 until 4.2 V",
            # "Hold at 4.2 V until C/20",
        ]
    )

    # re-import the paramter for double security
    param = pybamm.ParameterValues(params_list)

    # parameter update
    param.update({
        'Negative electrode active material volume fraction': input_val[0],
        'Positive electrode active material volume fraction': input_val[1],
    }, check_already_exists=False)

    # update the mesh
    var = pybamm.standard_spatial_vars
    var_pts = {
    var.x_n: 20,
    var.x_s: 20,
    var.x_p: 20,
    var.r_n: 11,
    var.r_p: 11,
    }
    # define the simulation
    sim = pybamm.Simulation(
            model,
            var_pts=var_pts,
            parameter_values=param,
            experiment=experiment   
            # solver=pybamm.CasadiSolver(mode="fast")
            )

    sol_ = sim.solve(initial_soc=0.0)   # 

    ## collect the output
    t_steps = sol_['Time [h]'].entries
    terminal_V = sol_["Terminal voltage [V]"].entries

    # visualize the results
    fig, ax = plt.subplots(figsize=(5, 4))
    ax.plot(t_steps, terminal_V, label="Terminal voltage")
    ax.set_xlabel("Time [h]")
    ax.set_ylabel("Terminal voltage [V]")

    return sol_

factors = [1.0, 0.92]  # work
# factors = [1.0, 0.91]  # not work
# factors = [0.4, 1.0] # not work
# factors = [0.39, 1.0] # work
inputs_eps = [eps_ne * factors[0], eps_pe * factors[1]]   # multiplier * inital value
print(inputs_eps)
sol = chemomechamodel_spm(inputs_eps)

If not work, the error is like:

SolverError: Could not find acceptable solution: solver terminated successfully, but maximum solution error (0.030825501155488766) above tolerance (1e-06)

[github-actions[bot]]

Hi! We have now moved our discussions to Discourse. Please post your question there.

[valentinsulzer]

The new IDA KLU Solver (pybamm.IDAKLUSolver()) should converge more reliably (and be much faster)

[XuboGU]

Thanks for your reply, Valentin:).
But it seems like the issue does not come from the Solver, cuz after I change the Solver and test, the results remain the same.