Learning Forum

[helen.durand]

Hello!

My understanding is the general procedure of FEA works something like this (at least for some types of problems): the equations in FEA are written in terms of the displacements (something like Ku=F, where K is the global stiffness matrix created though summing many local element stiffness matrices, and F is a nodal forces vector), and the displacements are first solved for.  Then the strain matrix (sometimes called the B matrix) is used to calculate strains from the displacement results, and matrix representing the relationship between stress and strain (called the D matrix in the theory reference) is used to calculate the stresses from the strains.

Does this seem like a correct high-level assessment of the FEA procedure?

Does the overall procedure change for transient or nonlinear analyses? 

Thank you!

[Ankush Choudhary]

Hi Helen,

Your understanding of FEA is correct. In the case of transients, we need to use time discretization techniques like Newmark algorithms. In the case of non-linear systems, we need to use an iterative solver like Newton-Raphson.

There are plenty of courses available on ANSYS Innovation Space related to these topics. These two courses below will help you understand the transient and non-linear solutions.

https://courses.ansys.com/index.php/courses/fea-linear-elastodynamics/

https://courses.ansys.com/index.php/courses/computational-resource-considerations-recommended/