mufem - Fast and Accurate Electromagnetic Simulations
mufem is a highly scalable finite element framework focused on electromagnetics, plasma, and multi-physics.
It provides a native Python interface, and is designed for automation, machine learning, AI-driven studies,
and large-scale repetitive simulation workflows.
Technology
- Using MFEM as backend for efficient discretization.
- Higher-order accuracy with higher-order geometry on unstructured meshes.
- Adaptive Mesh refinement with hanging nodes support.
- Native Python interface for problem specification and automation.
- Performance portability across heterogeneous CPU/GPU platforms.
- Cartesian meshes with implicit geometry workflows.
Models
Electromagnetics
Models electric machines including motors and generators with nonlinear magnetic materials, transient currents, and coupled electromechanical motion.
Simulates transformers, induction heating, and eddy-current losses in the frequency domain with accurate impedance and power loss prediction.
Time-Transient Maxwell
Solves full-wave transient electromagnetic propagation including pulsed excitation, switching dynamics, and wave–material interaction.
Computes frequency-domain electromagnetic fields for RF, microwave, resonance, scattering, antenna, and waveguide applications.
Structural
Solves heat conduction in solids including steady and transient regimes with temperature-dependent materials and boundary conditions.
Structural (Elasticity)
Computes linear elasticity for solid mechanics with scalable solvers for large meshes and parameter sweeps.
Plasma
Plasma
Model low-temperature plasmas with fluid and kinetic plasma models, including plasma chemistry, sheath physics, and plasma-surface interactions.