Error Estimation - CosmosWorks

[djw2k3]

Hi,

Have been using CMOS for sometime now. Doing some post processing analysis I want to provide an accurate measure of error. The .out files list some statistics including total strain error - is this really a good error representation?

Could someone confirm to put my mind at rest, how the differnces in strain energy (hence error)occur?

Thanks in advance

Dave

 

[manam]

I am guessing you want to know how this error in the strain energy is estimated ? Almost all estimates of error in the strain enery arised from the "residual force" and the "jumps in the tractions across the boundaries of the elements". The residual force arises due to the fact that the finite element solution does not satisfy static equilibrium within the element. And the jumps in the tractions across the boundaries of the elements in most elements can be visualized easily..