Simulation Error: Bad allocation

[Johannes-hub22]

Hi guys,

I am struggling to run a dummy degradation cycle on the O'Kane (Chen) dataset. After running 660 cycles I get the following error:
"Simulation error: bad allocation"

The code is in the text file attached. Please let me know if you think there is anything faulty from my side. Any help appreciated.

I have added another 2 cycles before the error cycle:

2022-10-13 09:47:34.559 - [NOTICE] callbacks.on_step_start(182): Cycle 658/1000, step 1/6: Discharge at 1C until 2.5 V
At t = 0.0141337 and h = 6.23111e-14, the corrector convergence failed repeatedly or with |h| = hmin.
2022-10-13 09:49:56.351 - [NOTICE] callbacks.on_step_start(182): Cycle 658/1000, step 2/6: Hold at 2.5 V until 50mA
2022-10-13 09:50:34.104 - [NOTICE] callbacks.on_step_start(182): Cycle 658/1000, step 3/6: Rest for 300 seconds
2022-10-13 09:50:55.228 - [NOTICE] callbacks.on_step_start(182): Cycle 658/1000, step 4/6: Charge at 0.3 C until 4.2 V
2022-10-13 09:56:53.276 - [NOTICE] callbacks.on_step_start(182): Cycle 658/1000, step 5/6: Hold at 4.2 V until 50 mA
2022-10-13 09:58:26.805 - [NOTICE] callbacks.on_step_start(182): Cycle 658/1000, step 6/6: Rest for 300 seconds
2022-10-13 10:00:52.763 - [NOTICE] callbacks.on_cycle_end(196): Capacity is now 4.994 Ah (originally 5.091 Ah, will stop at 1.018 Ah)
2022-10-13 10:00:52.786 - [NOTICE] callbacks.on_cycle_start(174): Cycle 659/1000 (1 day, 17 hours, 45 minutes, 12 seconds elapsed) --------------------
2022-10-13 10:00:52.819 - [NOTICE] callbacks.on_step_start(182): Cycle 659/1000, step 1/6: Discharge at 1C until 2.5 V
2022-10-13 10:03:30.196 - [NOTICE] callbacks.on_step_start(182): Cycle 659/1000, step 2/6: Hold at 2.5 V until 50mA
2022-10-13 10:04:09.080 - [NOTICE] callbacks.on_step_start(182): Cycle 659/1000, step 3/6: Rest for 300 seconds
2022-10-13 10:04:33.092 - [NOTICE] callbacks.on_step_start(182): Cycle 659/1000, step 4/6: Charge at 0.3 C until 4.2 V
2022-10-13 10:11:17.568 - [NOTICE] callbacks.on_step_start(182): Cycle 659/1000, step 5/6: Hold at 4.2 V until 50 mA
2022-10-13 10:13:11.722 - [NOTICE] callbacks.on_step_start(182): Cycle 659/1000, step 6/6: Rest for 300 seconds
2022-10-13 10:15:49.792 - [NOTICE] callbacks.on_cycle_end(196): Capacity is now 4.994 Ah (originally 5.091 Ah, will stop at 1.018 Ah)
2022-10-13 10:15:49.843 - [NOTICE] callbacks.on_cycle_start(174): Cycle 660/1000 (1 day, 18 hours, 0 minutes, 9 seconds elapsed) --------------------
2022-10-13 10:15:49.892 - [NOTICE] callbacks.on_step_start(182): Cycle 660/1000, step 1/6: Discharge at 1C until 2.5 V
2022-10-13 10:17:00.111 - [ERROR] callbacks.on_experiment_error(226): Simulation error: bad allocation
2022-10-13 10:17:00.474 - [NOTICE] callbacks.on_cycle_end(196): Capacity is now 4.994 Ah (originally 5.091 Ah, will stop at 1.018 Ah)
2022-10-13 10:17:02.099 - [NOTICE] callbacks.on_experiment_end(222): Finish experiment simulation, took 1 day, 18 hours, 1 minute, 19 seconds
Sim_error_Bad_allocation.txt

[rtimms]

When you run a simulation it is storing the whole result in RAM, so I'm guessing it's a memory issue. You could stop after a certain number of cycles, write the data you are interested in to disc, then resume the simulation using the final state as your initial condition (e.g. see this example).

You can also choose to only store the full solution at certain cycles, and store summary variables for the others, see this notebook.

[Johannes-hub22]

@rtimms : Thanks! I will give this a try.

[valentinsulzer]

If your simulation is taking 2 days you should really be saving intermediate steps anyway