Finite Element Analysis
Finite Element Analysis
Finite Element Analysis is a numerical procedure that only approximates exact solutions
of a system, not at any point of the system, but only at discrete points or at the nodes. This is due
to the fact that actual systems are actually complex geometrically and exact solutions cannot be
obtained. Finite Element Analysis, from how it is called, is an analysis of finite elements. More
specifically, it is an analysis of a whole system, a structure, for instance, by breaking down the
entire structure into smaller entities or bodies, or what we call “finite elements”, interconnected
together at specific points called nodes and/or by surfaces and/or by boundary lines. This
breakdown is called discretization. Nodes are the points where values from the analysis are given
at and the values between the nodes are obtained through interpolations. Finite elements can be
one-dimensional (line), two-dimensional (plane), or three-dimensional (solid). There can be
millions of finite elements in a structure. When all of these finite elements are combined
together, they produce the shape of the structure. In the same way, computations made on these
finite elements, when combined together, yield the outcome of the entire structure. Since this
analysis approximates the behavior of a structure, certain parameters are involved in the process.
These parameters are of two sets. The first set of parameters are those which give the structure its
natural behavior such as its modulus of elasticity and thermal conductivity. These actually are
the material and geometric properties of the system or the elements. The second set of
parameters are those which give disturbances to the system such as external forces and moments.
So, how does it work? It works just like an input-output system in a matrix. In FEA softwares
such as ANSYS or ABACUS, the aforementioned sets of parameters must first be known in
order to run the software because these parameters should be inputted. It is like inputting these
parameters in a matrix, just like what we do when manually solving deformations of beams or
frames subjected to loadings in Structural Theory using matrix method. Once these parameters
are inputted, the first thing that the analysis would do is to solve for the nodal displacements,
which eventually would express the stresses and strains of the elements. It would then show how
the system performs and behaves given the inputted parameters.
We, engineers, specifically civil engineers, use Finite Element Analysis, in order to have
a picture of what the actual physical situation would look like using software models as
representations of the actual thing. This is very important for us, especially when we are
designing. For instance, we should be able to predict and assess how the building we are
designing would behave when we are to use a certain support, a certain material, a certain
dimension, and the like, and when we subject this to a certain loading. Since the analysis is on
elements of the structure, we would also be able to point out specific weaknesses of our design.
Moreover, we would be able to assess if our design is safe or bound to failure. We can use FEA
to do simulations until we get the right and safe design, instead of using physical prototypes or
models to represent actual structures. Instead of assessing the elements of a structure one by one
manually, beam after beam, for example, FEA softwares are programmed to assess the elements
of the structure in just a few seconds. Aside from it would take you years to assess an entire
structure, it is also definitely not a guarantee that you would be able to do it accurately. Thus,
obviously, FEA would make our work easier, faster, and even more accurate.