Learning Forum

[Diane Salman]

Hello,

I am running a static structural simulation on a 2D model of an upper airway in which spring connections represent the passive behavior of muscles and I am applying a negative pressure on the airway walls. I have noticed that when the units change in "Mechanical" and I solve the model again, the results I obtain are different. From what I can see, the issue seems to be that the spring connections are not working properly anymore. 

For more clarity, when I created the geometry in DesignModeler, I used "millimeters". In Mechanical, when the units are in "millimeters" the solution is correct but when they change to "meters" the solution gives wrong results. Is there a reason for this to happen? Is it expected?  

Best, 

Diane 

[peteroznewman]

Hello Diane,

Did you use any Command Objects to define spring constants? 

Command objects don’t have units, so if you defined a spring constant in a command object, it could be given a value of 10 for the spring stiffness. When solving in mm, that means 10 N/mm, but when solving in meters that spring stiffness is 10 N/m.

Whenever I create a Command Object that includes an input parameter that has units, I use the Analysis Setting to set the Solver Unit System to Manual and set the Unit to what I want. That way the Unit setting in Mechanical doesn’t change what the solver does.

Best regards,
Peter

[Mike Rife]

Hi Diane

An "Usual Suspect" would be stiffness terms, especially contact stiffness, that are being rounded off.  On 'small' models that use MKS the stiffness' can get to be => 1E16 which can/will be rounded in double precision math.  When changing unit systems down to match the natural bounds of the model, the stiffness' would then be much less than 1E16.

If your model has contact check the contact stiffness (in solution output) when using MKS units.  There may also be an warning about these. 

What results are changing?  By how much?

If no contact, turn on the connections (Solution Information option) and see if there is a difference in the scoped region between the two unit systems.

Mike