memory lack

[kenan991]

Hello All
I am using pybamm.lithium_ion.DFN() and implementing the experiment on my data set, I mean I have data set(3000 values) containing power values and in every iteration, the experiment is dis/charge depending on the sign of the value but after 1000 iterations I am getting a memory lack (to know I have 8 Giga RAM) I tried to solve this problem by getting lose of unwanted previous values which saved in the arrays in the simulation class I mean (all_ts, all_ys, all_models, all_inputs, t_event, y_event) but I still get the same memory lack.
that is what I mean by simulation class
https://pybamm.readthedocs.io/en/latest/source/solvers/solution.html?highlight=solution#pybamm.Solution

please, is there a way to solve my problem?

Best
Kenan

[valentinsulzer]

What do you mean by "iteration" here? Cycles? Can you provide a minimum working example?

[kenan991]

yeah I mean cycles

this part of my code where I implement the battery using pybamm

class Battery: # check the power drwan/feeded to the grid

def __init__(self,  Discharge capacity=0):

    #self.last_solution = None
    self.parameter_values = pybamm.ParameterValues(chemistry=pybamm.parameter_sets.Chen2020)
    self.model = pybamm.lithium_ion.DFN()

    # TODO use these to initi
    self.Discharge capacity = Discharge capacity

def update(self, power):
    global last_solution
    #time_passed = self.env.now - self.last_update
    if power > 0:
        experiment = pybamm.Experiment([f"charge at {0.1 * power} W for 15 s"])
    elif power < 0:
        experiment = pybamm.Experiment([f"discharge at {-0.1 * power} W for 15 s"])
    else:
        return 0

    sim = pybamm.Simulation(self.model, parameter_values=self.parameter_values, experiment=experiment)
    #sim.solve(starting_solution=self.last_solution)
    sim.solve(starting_solution=last_solution)
    last_solution = sim.solution
    #self.last_solution = sim.solution
    d = last_solution['Discharge capacity [A.h]']
    self.Discharge capacity =d.entries[-1]
    sim.plot(['Current [A]','Terminal voltage [V]','Discharge capacity [A.h]'])

battery=Battery()
for i in range(0, len( delta_energy )): ##len(production_df)=2880
grid_energy = battery.update(delta_energy)

so it works well until i=1000 I mean after making 1000 cycle and then the simulation is slowing up and the memory gets full
and could not finish the simulation for the whole data set

[valentinsulzer]

Check out some of the examples in https://github.com/pybamm-team/PyBaMM/blob/develop/examples/notebooks/simulating-long-experiments.ipynb, especially save_at_cycles