Skip to content

modelica

A Modelica toolchain written only in Python. It reads Modelica source, flattens it to a DAE, simulates it, and draws it. It exchanges models through FMI 3.0 and SSP 2.0.

pip install modelica          # the core: parse, model structure, diagrams. No dependencies.
pip install "modelica[sim]"   # adds numpy and scipy, and therefore simulation

Pre-alpha

This is a full rewrite of a prototype from 2010, and it is not finished. The API is not stable. The layers below are being built from the bottom up. Python 3.12 or later.

The layers

Module What it does
modelica.lang Modelica source to tokens to an AST
modelica.build Python classes to the same AST values
modelica.ir AST to a flat model, then matching, BLT sorting and index reduction
modelica.sim Flat model to an ODE or DAE problem, with solvers and events
modelica.diagram Any stage to a picture: SVG, Graphviz or Mermaid
modelica.fmi FMI 3.0 — import, simulate and export an FMU
modelica.ssp SSP 2.0 — system composition and a co-simulation master

Start with Architecture for the data model and the build order.

See a model first

A Modelica model is a graph, and the fastest way to understand one is to look at it.

from modelica import diagram

diagram.of("examples/models/RLCCircuit.mo", "examples/library/Electrical.mo")

graph LR
  nsource["source<br/>ConstantVoltage"]
  nr["r<br/>Resistor"]
  nl["l<br/>Inductor"]
  nc["c<br/>Capacitor"]
  nground["ground<br/>Ground"]
  n0(( ))
  nsource ---|p p| nr
  nr ---|n p| nl
  nl ---|n p| nc
  nc ---|n| n0
  nsource ---|n| n0
  nground ---|p| n0

The three connectors that meet at the bottom are one junction, and not three separate lines. That is what the flattener makes of them: one sum of currents, taken once.

Drawing a model shows the other two pictures — equations against variables, and blocks in the order the solver runs them.

Notebooks

Every notebook here is executed when this site is built. What you read is what the current code does. The .ipynb pages carry figures and sliders, and they are meant to be opened.

Page What it shows
Simulating a model Source to AST to flat model to system to code to trajectory, stopping at each stage
Drawing a model The three pictures of one model, and what each one is for
Turning the knobs A parameter study as a for loop over copies of the flat model, then the same thing on sliders
The shape of the equations Incidence matrices, what alias elimination removes, and the BLT staircase
Where the smoothness breaks Bounce events, the step size collapsing into them, and the pendulum changing state sets
Roasting a batch A model of a real machine, checked against what the machine recorded
Enumerating the discrete half Model checking of the event iteration, state selection, FMI modes and the co-simulation master

The .ipynb pages read their models from examples/, so you need a checkout to run them:

git clone https://gitlab.com/jorgeecardona/pymodelica && cd pymodelica
pip install "modelica[sim]" matplotlib ipywidgets
jupyter lab docs/notebooks/

Standards

How these pages are written

They follow ASD-STE100 Simplified Technical English. The style page lists the rules and the two places this project departs from them.

Trademarks

Modelica® is a registered trademark of the Modelica Association. FMI and SSP are also its trademarks. These pages use the names to identify the language this software reads and the standards it implements.

This project is not affiliated with the Modelica Association, and the Association does not sponsor or endorse it. It is an independent open-source implementation. The language, its specification and the Modelica Standard Library are at modelica.org.