search

AtChem / AtChem2

Atmospheric chemistry box-model for the MCM
MIT License
59 stars 23 forks source link

CVODE ERROR #498

Closed zyh9291 closed 1 year ago

zyh9291 commented 1 year ago

Hello, In a recent simulation, I only changed the species' concentrations with leaving the rest of parameter settings unchanged, but it reported a running error that was " [CVODE ERROR] CVode At t = 61200 and h = 2.36018e-13, the corrector convergence test failed repeatedly or with |h| = hmin. SUNDIALS_ERROR: FCVODE() returned ier = -4 Linear Solver returned ier = 2" I’m not sure if it's because of the data. I was wondering if there are some methods to solve this problem. Thank you for your attention to this matter. I would appreciate any help you can offer.

rs028 commented 1 year ago

Hi @zyh9291 I assume you are talking about the initial concentration. If you change back to the value that you had before do you still get the same error?

zyh9291 commented 1 year ago

Hello @rs028 There was a 'CVODE ERROR' after changing the constraint concentration, but it was possible to run successfully with the previous value.

rs028 commented 1 year ago

In this case I think it strongly suggests that the new value you are using is creating the problem. Which variable is it and what are the values (old and new)?

zyh9291 commented 1 year ago

I changed the constraint concentrations for 72 species(VOCs, NOx, CO...). May I take it that this set of constrained datas cannot be solved by CVODE?

rs028 commented 1 year ago

Possibly. I suggest you review those inputs and see if there are any errors or abnormal values. Perhaps you can change 10 species to the new value and leave the other 62 with the old values. If the model runs then those 10 variables are fine and you can change 10 more and so on. That should help narrow it down to the particular variable that is causing the problem.

zyh9291 commented 1 year ago

@rs028 Thanks for your advice and assistance.