Suggestion: Adding a Note for Units in Documentation and Source code

[vikramroybarc]

Hi Erlap,

I wanted to share a suggestion regarding the documentation and source code. I believe it will significantly enhance the user-friendliness of the source code.

In the current implementation, certain variables like "KB" and "tolerance" depend on the unit system used for analysis. These variables are stored in different modules, mainly as global variables and user inputs. For a new user, it would be very helpful to include a note in both the documentation and the source code, explaining which variables need to be adjusted when changing units and what preferred units to use with this umat.

For instance, if a user is utilizing stress in MPa units, the present value of tolerance (1.d-8) implemented in user inputs is valid. However, if the user is employing stress in Pa units, adjusting the tolerance values to (1.d-2 or 1.d-1) may be more appropriate.

I became aware of these details while debugging the subroutine in VSCode. By incorporating these clarifications in the documentation and source code, we can minimize sources of error and significantly improve the code's user-friendliness.

Thank you for your attention to this matter.

Best regards,
Vikram Roy

[EralpDemir]

Thank you so much for your suggestion Vikram. Even though I tried to put the unit to the documentations, it is difficult to follow.

I will prepare a user-guide for this, but probably not soon. So, I keep this issue open.

Units of stress is MPa in the code, I would adopt the constitutive laws based on that otherwise all the units in the usermaterial.f such as elastic constants need to be revisited.

Dimension units are micrometers which is relevant to crystal plasticity and scales we are dealing with. The RVE generators also generate models with micrometer dimensions. I would also stick to this unit.

Best wishes,
Eralp

[Akashgiitk]

Dear sir ,
if i use stress unit in MPa ( N/mm² ) and it doesn’t match μm-based length units,
Should I rescale stress and material properties to N/μm² to maintain consistency?
Or is there any internal unit conversion in UMAT that handles this mismatch automatically?