What is being solved physically?

The equation **−Δu=f** can model electrostatic potential from charge density, steady heat conduction with sources, or a membrane deflection under load. The interpretation changes, but the elliptic operator and boundary-condition ideas are the same.

Why are Neumann boundaries “natural” in FEM?

After integrating by parts, flux boundary terms appear on the right-hand side as **∫Γ_N g v ds**. Dirichlet values must be imposed by pinning degrees of freedom, while Neumann data enters directly as load contributions.

Is this a production-quality FEM code?

No. It is a teaching model on a structured triangular mesh with simple quadrature and a basic CG solver. Real FEM software adds mesh generation, higher-order elements, robust preconditioning, error estimators, and careful treatment of complex geometries.

Finite Element Poisson Solver (2D)

Triangular P1 finite elements for -Δu=f on a square: assemble stiffness matrices, pin Dirichlet nodes or add natural Neumann flux, solve by CG, and view potential as a colored membrane.

About this model

This page is a compact finite-element Poisson solver for −Δu = f on the unit square. The domain is split into a regular triangular mesh and each triangle uses linear P1 basis functions. For every element, the local stiffness matrix is assembled from the gradients of the basis functions, Kᵉ_ij = (b_i b_j + c_i c_j)/(4A), then accumulated into a sparse global system. Source terms are integrated by a centroid rule. In pure Dirichlet mode all boundary nodes are pinned to u=0, like a membrane clamped to a frame. In mixed mode, the left/right sides are Dirichlet while top/bottom edges contribute natural Neumann flux terms to the right-hand side. The resulting symmetric positive-definite system is solved by conjugate gradients. The left view colors the potential on the triangular mesh; the right view lifts the same scalar field into a “deformable membrane” surface.

Who it's for: Students in numerical PDEs, finite elements, electrostatics, and computational mechanics learning weak forms, triangular P1 elements, and boundary-condition handling.

Key terms

Finite element method
Poisson equation
Triangular mesh
P1 element
Weak form
Dirichlet boundary
Neumann boundary
Conjugate gradients

How it works

A compact 2D finite-element Poisson solver: assemble linear triangular P1 elements for -Δu=f, apply Dirichlet or natural Neumann boundaries, solve the sparse system by conjugate gradients, and view the potential as a colored membrane.

Key equations

Weak form: find u such that ∫Ω ∇u·∇v dΩ = ∫Ω f v dΩ + ∫Γ_N g v ds. P1 triangles give local stiffness K_ij^e = (b_i b_j + c_i c_j)/(4A); Dirichlet values are pinned.

Frequently asked questions

What is being solved physically?: The equation −Δu=f can model electrostatic potential from charge density, steady heat conduction with sources, or a membrane deflection under load. The interpretation changes, but the elliptic operator and boundary-condition ideas are the same.
Why are Neumann boundaries “natural” in FEM?: After integrating by parts, flux boundary terms appear on the right-hand side as ∫Γ_N g v ds. Dirichlet values must be imposed by pinning degrees of freedom, while Neumann data enters directly as load contributions.
Is this a production-quality FEM code?: No. It is a teaching model on a structured triangular mesh with simple quadrature and a basic CG solver. Real FEM software adds mesh generation, higher-order elements, robust preconditioning, error estimators, and careful treatment of complex geometries.

Continue with similar topics in this category — or all 85 in Math Visualization.

View category →

NewUniversity / research

Heat Equation: Finite Differences

Launch Simulator

1D heat equation u_t = αu_xx with explicit FTCS and implicit backward Euler; tune CFL r = αΔt/Δx², watch explicit blow-up for r > 1/2, and compare numerical diffusion.

NewUniversity / research

Conjugate Gradient Solver

Launch Simulator

SPD system Ax=b as quadratic minimization: contour geometry, CG vs steepest descent path, residual norm, and condition number.

NewUniversity / research

MinD / MinE Oscillation (E. coli Rod)

Launch Simulator

1D reaction–diffusion: membrane MinD u, fast MinE v; pole-to-pole oscillation; division plane at time-averaged MinD minimum.

NewUniversity / research

Finite-Volume Advection-Diffusion 2D

Launch Simulator

Conservative scalar transport on a 2D grid: face fluxes, upwind vs central interpolation, Peclet number, CFL, and numerical diffusion.

NewUniversity / research

Lattice Boltzmann D2Q9 Flow

Launch Simulator

Interactive D2Q9 BGK solver: lid-driven cavity or flow past a cylinder, vorticity colors, Reynolds-number control, and bounce-back walls.

NewUniversity / research

Keller–Segel Chemotaxis

Launch Simulator

∂ₜn = Dₙ∇²n − χ∇·(n∇c), ∂ₜc = D_c∇²c + αn − βc; bacteria follow attractant; collapse at high χ; 96² grid.

Finite Element Poisson Solver (2D)

About this model

How it works

Key equations

Frequently asked questions

Related simulators

Heat Equation: Finite Differences

Conjugate Gradient Solver

MinD / MinE Oscillation (E. coli Rod)

Finite-Volume Advection-Diffusion 2D

Lattice Boltzmann D2Q9 Flow

Keller–Segel Chemotaxis

Finite Element Poisson Solver (2D)

About this model

How it works

Key equations

Frequently asked questions