Finite Element Methods

Finite element methods recast a PDE in weak (variational) form and approximate its solution by piecewise-polynomial functions on a mesh of simple elements, yielding accurate solutions on complex geometries.

Definition

The finite element method is a numerical technique that approximates the solution of a PDE by projecting its weak formulation onto a finite-dimensional space of piecewise-polynomial functions defined over a mesh, reducing the problem to a system of algebraic equations.

Scope

This topic covers weak formulations and Sobolev-space settings, the Galerkin method and Cea's lemma, construction of finite element spaces on triangulations, assembly of the stiffness matrix and load vector, a priori error estimates, and a posteriori estimates that drive adaptive mesh refinement.

Core questions

• How does the weak formulation broaden the class of admissible solutions and underpin the method?
• How does Galerkin projection, via Cea's lemma, relate the discrete error to the best approximation?
• How are finite element spaces constructed and the global system assembled from local element contributions?
• How do a priori and a posteriori error estimates quantify accuracy and guide mesh adaptation?

Key theories

Weak formulation and Lax-Milgram

Multiplying the PDE by test functions and integrating recasts it as a variational problem in a Sobolev space; the Lax-Milgram theorem guarantees a unique weak solution when the associated bilinear form is bounded and coercive, providing the rigorous foundation for the method.

Galerkin orthogonality and Cea's lemma

The finite element solution satisfies Galerkin orthogonality, and Cea's lemma bounds its error by a constant times the best approximation error in the finite element space, reducing convergence analysis to the approximation power of the chosen elements.

A posteriori estimation and adaptivity

Computable a posteriori error estimators bound the actual error using only the discrete solution and data, enabling adaptive algorithms that refine the mesh where the error is largest to achieve a target accuracy efficiently.

Mechanisms

The domain is partitioned into elements (triangles, tetrahedra, or quadrilaterals), and on each element the solution is represented by polynomial basis functions whose supports overlap only on shared faces, giving locally supported global basis functions. Substituting these into the weak form produces a sparse linear system: the stiffness matrix from the bilinear form and the load vector from the data, both assembled element by element. Solving the system yields the coefficients of the approximate solution. A priori estimates relate the error to the mesh size and polynomial degree, while a posteriori estimators steer adaptive refinement.

Clinical relevance

The finite element method is the dominant simulation technology in structural and solid mechanics, heat transfer, electromagnetics, and biomechanics, and is widely used in fluid dynamics; its ability to handle complex geometries, varied material properties, and adaptive refinement makes it the backbone of most commercial engineering analysis software.

History

The method arose from structural engineering in the 1950s and was given a variational mathematical foundation drawing on Courant's earlier work; the rigorous approximation theory was developed by Ciarlet, Babuska, and others in the 1970s, turning the finite element method into both a practical tool and a deep area of numerical analysis.

Key figures

• Richard Courant
• Olgierd Zienkiewicz
• Philippe Ciarlet
• Susanne C. Brenner

Seminal works

• brenner2008
• ern2004

Frequently asked questions

Why reformulate the PDE in weak form?

The weak form lowers the differentiability required of the solution and casts the problem in a Hilbert-space setting where existence, uniqueness, and approximation can be analyzed rigorously, and it naturally accommodates piecewise-polynomial approximations on complex meshes.

What makes finite elements good for complex geometries?

Because the domain is broken into small, simply shaped elements that can be sized and oriented to fit the boundary, finite element meshes conform to intricate shapes far more easily than the regular grids that finite difference methods require.