Feature/mod battery cycle only #84

lightningclaw001 · 2021-11-01T01:32:46Z

-added pure cycling features in a modular format
-added tests for full and half cells to test cycling

@d-cogswell @loostrum the new cycling tests are passing on my computer, but not remotely. Any thoughts?

loostrum · 2021-11-01T10:27:47Z

I'm not sure what's wrong with test030, that one's fine on my PC as well. Perhaps something with the noise? I remember that was an issue earlier for some tests.
test029 is also failing for me, specifically it's complaining that the shapes of the data in the reference output and new output aren't the same: I see ValueError: operands could not be broadcast together with shapes (1,40) (1,38) on the Github page. I ran it on my PC too, and I added a line to print which variable is giving problems. I get Fail from ValueError in variable phi_lyte_c: operands could not be broadcast together with shapes (47,1) (45,1) . Interestingly this is not the same exact error, but it's also a problem with mismatching shapes with an offset of 2.
I'm surprised it's working for you. Could it be you updated the reference solution for test029, but didn't commit it to the repository yet?

lightningclaw001 · 2021-11-02T04:46:03Z

I'm not sure what's wrong with test030, that one's fine on my PC as well. Perhaps something with the noise? I remember that was an issue earlier for some tests. test029 is also failing for me, specifically it's complaining that the shapes of the data in the reference output and new output aren't the same: I see ValueError: operands could not be broadcast together with shapes (1,40) (1,38) on the Github page. I ran it on my PC too, and I added a line to print which variable is giving problems. I get Fail from ValueError in variable phi_lyte_c: operands could not be broadcast together with shapes (47,1) (45,1) . Interestingly this is not the same exact error, but it's also a problem with mismatching shapes with an offset of 2. I'm surprised it's working for you. Could it be you updated the reference solution for test029, but didn't commit it to the repository yet?

@loostrum interesting that it isn't working for you. I did push the new solutions. I don't think there is a test030, are you sure you're on the right branch? I did previously get these errors, but I realized that they were correlated with the phase separating models and that if I switched to the OCV models all my tests were passing. I figured the issue was something else and that I would tell Dan in person (and was likely that it was correlated with the noise). Are you sure you're on the latest commit of this branch? Are the config files in test025 and test029 from LIONSIMBA files?

loostrum · 2021-11-03T15:41:55Z

I was wrong about test030, the output actually listed Dirs030 which I assumed belonged to a test030, but that isn't the case. I double-checked that I'm on the correct branch. Indeed the config files are from LIONSIMBA. I ran the tests on both my Mac and on Linux in a docker container. On linux, only test025 fails. On mac, test025 as well as test029 fail.

On linux, the test025 error is a ValueError about shapes, but it's not the same variable that's failing every time which is a bit worrisome. It only shows the first failure, but unless the order in which the variables are checked is random I'd really expect the same failure every time. I've seen two different errors so far: a (1,47) vs (1,45) shape mismatch in partTrodecvol0part0_cbar, and a (47,11) vs (45,11) mismatch in partTrodecvol0part0_c. Do you know where the 45/47 number comes from? It seems that the reference solution has 47, while the newly calculated solution has 45.

lightningclaw001 · 2021-11-03T20:39:03Z

I was wrong about test030, the output actually listed Dirs030 which I assumed belonged to a test030, but that isn't the case. I double-checked that I'm on the correct branch. Indeed the config files are from LIONSIMBA. I ran the tests on both my Mac and on Linux in a docker container. On linux, only test025 fails. On mac, test025 as well as test029 fail.

On linux, the test025 error is a ValueError about shapes, but it's not the same variable that's failing every time which is a bit worrisome. It only shows the first failure, but unless the order in which the variables are checked is random I'd really expect the same failure every time. I've seen two different errors so far: a (1,47) vs (1,45) shape mismatch in partTrodecvol0part0_cbar, and a (47,11) vs (45,11) mismatch in partTrodecvol0part0_c. Do you know where the 45/47 number comes from? It seems that the reference solution has 47, while the newly calculated solution has 45.

I did some digging and it seems like the reason "extra steps" are being taken is probably a tolerance issue. In some of the simulations, the (ffrac-cutoff) hits exactly zero, while the cutoff that I set was (ffrac-cutoff) >= 0. Let me see if there's some way to make the condition more tolerant of these kinds of issues.

lightningclaw001 · 2021-11-04T04:20:43Z

@loostrum I'm going to email you a patch (since it is a little large since the test case solutions also changed), can you try testing that and see if it still fails for the tests? Thanks!

loostrum · 2021-11-17T09:33:31Z

HI @lightningclaw001, I tried applying the patch you sent me, but unfortunately git diff doesn't handle the difference in binary files, so it fails to apply the changes you made to the output_data.mat and input_dict files in the ref_outputs folder. Perhaps you could send me the entire repo as you have it now, or push it to a separate branch?

lightningclaw001 · 2021-11-17T16:52:28Z

@loostrum I pushed the commits into feature/mod_battery_cycle_only_test, can you try them here? Thanks!

loostrum · 2021-11-19T09:16:43Z

Thanks! Unfortunately tests 25 and 29 still fail for me on the new branch, still with the 2 mismatch in shapes.

lightningclaw001 · 2021-11-19T16:49:42Z

Thanks! Unfortunately tests 25 and 29 still fail for me on the new branch, still with the 2 mismatch in shapes.

Can you try applying this commit to that branch:
test_diff.txt
and let me know if it works? If not, can you send me the output_data.mat files that are failing?
Really appreciate it!

loostrum · 2021-12-14T15:06:40Z

I tried the patch, but it's not working. I'll send you the mat files in an email!

lightningclaw001 · 2021-12-22T01:58:49Z

@loostrum I realized that multiplying the cutoff (eg. x > cutoff turns into x1e4 > cutoff1e4) by 1e4 works by turning the numbers closer to integer values but this is a very sketchy solution... do you have any other suggestions as to what to do to avoid these floating point errors?

lightningclaw001 · 2021-12-27T17:52:42Z

also @loostrum what version of python were you using for the tests? I realized that mine fail for 3.9 but pass for 3.7 locally...

loostrum · 2022-01-10T13:18:26Z

I'm using python 3.6.
Not sure what we could do about the cutoff floating-point error, I'll think about it.

d-cogswell · 2022-01-18T19:48:39Z

I wonder if it has to do with the port class you added. Are all of the port variables being initialized? Residual equations that copy one variable to another isn't great either. Do you really need the port class?

lightningclaw001 · 2022-01-18T20:00:20Z

I wonder if it has to do with the port class you added. Are all of the port variables being initialized? Residual equations that copy one variable to another isn't great either. Do you really need the port class?

If we want to switch to a different module but also import the data a level above I think we need to keep the port class. Pretty annoying but the other option is to go back to a nonmodular format.

I also don't think we had issues with the ports in 3.7?

lightningclaw001 · 2022-01-18T20:01:00Z

I'm using python 3.6. Not sure what we could do about the cutoff floating-point error, I'll think about it.

So your 3.6 is also having issues with the cutoffs? Hm... weird...

d-cogswell · 2022-01-21T16:10:36Z

I made some progress removing the ports and residual equations that copy one variable to another. I think ports are really only for when different models need to change the same variable.

Tests now pass in different versions of Python on Linux, but fail on Windows. Will keep working on it.

d-cogswell · 2022-06-27T15:31:39Z

This last commit e4fc3f8 seems to have broken plotting for me?

d-cogswell · 2022-07-06T15:41:17Z

mpet/plot/plot_data.py

@@ -47,6 +49,9 @@ def show_data(indir, plot_type, print_flag, save_flag, data_only, vOut=None, pOu
    trodes = config["trodes"]
    # Pick out some useful calculated values
    limtrode = config["limtrode"]
+    tot_cycle = 1
+    if "totalCycle" in config:


This check causes a crash.

@d-cogswell I just removed the last commit--it's no longer a dict after the old commands so some of the commands I was using wasn't working.

CV.

d-cogswell · 2023-05-23T13:45:05Z

@lightningclaw001 I don't think we need ab6b66c, this was already fixed in v0.1.9. Also, @v1kko is looking into why the regression tests are suddenly taking so long.

d-cogswell · 2023-05-23T13:48:05Z

@lightningclaw001 what do you think of this?
https://github.com/d-cogswell/mpet/commits/feature/mod_battery_cycle_only

I did some cleaning.

lightningclaw001 · 2023-05-23T14:38:03Z

@lightningclaw001 I don't think we need ab6b66c, this was already fixed in v0.1.9. Also, @v1kko is looking into why the regression tests are suddenly taking so long.

ah yes, I remember now

lightningclaw001 · 2023-05-23T14:49:23Z

@lightningclaw001 what do you think of this? https://github.com/d-cogswell/mpet/commits/feature/mod_battery_cycle_only

I did some cleaning.

I think other than getting rid of the tramp for the increment cycle (which helps with abrupt steps because I've had problems with them before), I'm happy with the changes

d-cogswell · 2023-05-24T13:34:48Z

@lightningclaw001 what do you think of this? https://github.com/d-cogswell/mpet/commits/feature/mod_battery_cycle_only
I did some cleaning.

I think other than getting rid of the tramp for the increment cycle (which helps with abrupt steps because I've had problems with them before), I'm happy with the changes

Ok, I'll restore tramp. I think I have a better solution for handling transitions, but let's save that for later.

lightningclaw001 · 2023-05-24T14:30:45Z

@d-cogswell I'm happy with the changes now

lightningclaw001 added 5 commits October 22, 2021 11:04

added cycling module and connected to the rest of the code

cf3f405

added plots for cycling data

03a7fa9

removed sympy related stuff

622dfe1

added test cases for cycling

065011b

removed sympy utility (will add later in another branch)

9dd88c0

lightningclaw001 requested a review from d-cogswell November 1, 2021 01:32

lightningclaw001 changed the base branch from master to development November 1, 2021 01:32

changed test case start/end points

fd2c872

lightningclaw001 force-pushed the feature/mod_battery_cycle_only branch from 0646027 to fd2c872 Compare November 3, 2021 21:57

Merge branch 'development' into feature/mod_battery_cycle_only

7869d23

d-cogswell added 3 commits January 21, 2022 11:16

Removes ports from CCCVCPcycle.

5069033

Removes maccor_step_number residual equation.

e9be3b8

Removes unnecessary charge_discharge variable.

0d718a2

d-cogswell force-pushed the feature/mod_battery_cycle_only branch from 24d9716 to 387e22b Compare February 9, 2022 14:41

removed material type and switched to volumetric capacity

14f330d

lightningclaw001 force-pushed the feature/mod_battery_cycle_only branch from b8ef780 to 14f330d Compare February 10, 2022 17:47

d-cogswell reviewed Jul 6, 2022

View reviewed changes

lightningclaw001 force-pushed the feature/mod_battery_cycle_only branch from e4fc3f8 to 14f330d Compare July 6, 2022 17:21

Merge branch 'development' into feature/mod_battery_cycle_only

70326ab

d-cogswell force-pushed the feature/mod_battery_cycle_only branch from 70326ab to 14f330d Compare September 13, 2022 18:10

d-cogswell and others added 2 commits September 13, 2022 14:18

Remove logPad from reference tests.

ba6996a

changed config parser so that repeated cycle states can be added

5aa29c7

lightningclaw001 force-pushed the feature/mod_battery_cycle_only branch 2 times, most recently from e14e431 to 0c7f8ed Compare December 16, 2022 17:20

Base automatically changed from development to master January 27, 2023 20:03

Bug fix for mpetplot. Can no longer assume profileType is either CC or

085c908

CV.

d-cogswell force-pushed the feature/mod_battery_cycle_only branch from 0c7f8ed to 085c908 Compare May 3, 2023 22:04

d-cogswell changed the base branch from master to development May 22, 2023 16:32

d-cogswell mentioned this pull request May 22, 2023

Discharge current proportional to cathode thickness #120

Open

d-cogswell force-pushed the development branch from 1c31554 to 8b4c351 Compare May 24, 2023 00:42

d-cogswell force-pushed the feature/mod_battery_cycle_only branch from ab6b66c to abb1a29 Compare May 24, 2023 13:37

Merge branch 'development' into feature/mod_battery_cycle_only

24f3815

d-cogswell force-pushed the feature/mod_battery_cycle_only branch from abb1a29 to 24f3815 Compare May 24, 2023 13:46

removed maccor cycling steps options in input

cef503f

d-cogswell merged commit 5a91c98 into development May 25, 2023

d-cogswell deleted the feature/mod_battery_cycle_only branch May 25, 2023 19:58

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Feature/mod battery cycle only #84

Feature/mod battery cycle only #84

lightningclaw001 commented Nov 1, 2021 •

edited

Loading

loostrum commented Nov 1, 2021

lightningclaw001 commented Nov 2, 2021

loostrum commented Nov 3, 2021

lightningclaw001 commented Nov 3, 2021

lightningclaw001 commented Nov 4, 2021

loostrum commented Nov 17, 2021 •

edited

Loading

lightningclaw001 commented Nov 17, 2021

loostrum commented Nov 19, 2021

lightningclaw001 commented Nov 19, 2021 •

edited

Loading

loostrum commented Dec 14, 2021

lightningclaw001 commented Dec 22, 2021

lightningclaw001 commented Dec 27, 2021 •

edited

Loading

loostrum commented Jan 10, 2022

d-cogswell commented Jan 18, 2022

lightningclaw001 commented Jan 18, 2022 •

edited

Loading

lightningclaw001 commented Jan 18, 2022

d-cogswell commented Jan 21, 2022

d-cogswell commented Jun 27, 2022

d-cogswell Jul 6, 2022

lightningclaw001 Jul 6, 2022

d-cogswell commented May 23, 2023

d-cogswell commented May 23, 2023

lightningclaw001 commented May 23, 2023

lightningclaw001 commented May 23, 2023

d-cogswell commented May 24, 2023

lightningclaw001 commented May 24, 2023

Feature/mod battery cycle only #84

Feature/mod battery cycle only #84

Conversation

lightningclaw001 commented Nov 1, 2021 • edited Loading

loostrum commented Nov 1, 2021

lightningclaw001 commented Nov 2, 2021

loostrum commented Nov 3, 2021

lightningclaw001 commented Nov 3, 2021

lightningclaw001 commented Nov 4, 2021

loostrum commented Nov 17, 2021 • edited Loading

lightningclaw001 commented Nov 17, 2021

loostrum commented Nov 19, 2021

lightningclaw001 commented Nov 19, 2021 • edited Loading

loostrum commented Dec 14, 2021

lightningclaw001 commented Dec 22, 2021

lightningclaw001 commented Dec 27, 2021 • edited Loading

loostrum commented Jan 10, 2022

d-cogswell commented Jan 18, 2022

lightningclaw001 commented Jan 18, 2022 • edited Loading

lightningclaw001 commented Jan 18, 2022

d-cogswell commented Jan 21, 2022

d-cogswell commented Jun 27, 2022

d-cogswell Jul 6, 2022

Choose a reason for hiding this comment

lightningclaw001 Jul 6, 2022

Choose a reason for hiding this comment

d-cogswell commented May 23, 2023

d-cogswell commented May 23, 2023

lightningclaw001 commented May 23, 2023

lightningclaw001 commented May 23, 2023

d-cogswell commented May 24, 2023

lightningclaw001 commented May 24, 2023

lightningclaw001 commented Nov 1, 2021 •

edited

Loading

loostrum commented Nov 17, 2021 •

edited

Loading

lightningclaw001 commented Nov 19, 2021 •

edited

Loading

lightningclaw001 commented Dec 27, 2021 •

edited

Loading

lightningclaw001 commented Jan 18, 2022 •

edited

Loading