An intermediate representation system for multi-physics problems.

Define Topology, Attach Metadata, Solve.
Keep consistent representation across physics.

---

Important

This README contains the architectural and conceptual vision of PyFEA. Version 0.1 is intended to be released on December 11, 2026.

Overview

PyFEA is a solver-adaptor engine that functions as an intermediate representation system for computational engineering. It creates a consistent representation across domains because continuous problems should use continuous tooling.

Important

Objectives:

• Allow for the same methodology across domains: define, attach, and solve.
• Allow for solver-adaptors across planar, axisymmetric, and full 3D solutions using CSG.
• Support integration with solvers across finite element, lumped parameters, and SPICE models.
• Restrict all inputs and outputs to dimensionally consistent units.

What is a Solver Adaptor?

An abstract boundary between a solver and PyFEA, it allows PyFEA to orchestrate the problem while the solver computes the solution.

For example, if you wanted to simulate an axial flux motor, it would require a 3D magnetostatic solver and perhaps a circuit solver for the triple half-bridge driver.

SPICE Circuit Solver
        ↓
3D Magnetostatic Solver
        ↓
Mechanical Integrator
         ↺ (Feeds back over time state)

Note

↺ signifies that the Mechanical Integrator feeds back into the Circuit Solver at each time step.

This is much easier than writing one large solver for axial flux motors. However, this isn't the only benefit. The main benefit is that a new arbitrary problem becomes a single custom adaptor away from solving.

For example, if you wanted to simulate an Atmospheric ion engine, it would require a 3D electromagnetic solver and a 3D fluid dynamics solver. But what about ionization? This is where a custom solver-adaptor comes in — you can write your own ionization solver and the pipeline is complete.

3D Electromagnetic Solver
          ↓
3D Fluid Dynamics Solver
          ↓
Custom Ionization Solver
          ↺

High-Level Architecture

PyFEA has a series of foundational dependencies that allow the engine itself to stay streamlined.

UIV/UT (DSL) → PicoUnits (Runtime Analysis) → PicoMaterials (Material Library) → PyFEA (Solver-Adaptors)

Unit-Informed Values (.uiv) is the custom domain-specific language for parameter and material files. PicoUnits interprets the .uiv file format and performs runtime dimensional analysis. Using .uiv and PicoUnits, PicoMaterials stores material data and passes material assumptions to PyFEA, which orchestrates the solver adaptors to solve the problem and returns the assumption tree.

Installation

Until release, this only installs the overview page and related files:

pip install pyfea

Documentation

Important

Internal Documentation refers to engineering logs, problem-solving notes, and unpolished application notes. For polished documentation, refer to External Documentation.

All internal documentation can be found within this repo's issues.