Simulate the circulation in your browser

This page runs the project's real Python engine (physiosim) with Pyodide, entirely on your device. Choose a model, change a few parameters, press Run and watch pressure, volume and flow evolve over the last three heartbeats.

Loading Python engine…

First visit downloads about 30 MB (Python, numpy, scipy, physiosim) from this site; your browser caches it afterwards.

1 · Model and parameters

Model Phase 1 — six compartments M1 — four chambers (passive atria)

Simulated beats

2 · Animated circulation

Shape area ∝ volume, fill colour = pressure. Valves are drawn open when their recorded flow is > 0. Moving dashes show flow direction and speed (∝ flow). Select a shape or an edge for numbers.

Scale each compartment to its own maximum (makes small chambers easier to see; not comparable between compartments)

Speed

Time within the final three beats

t = –

Compartment Edge

Volume (V): – mL
Pressure (P): – mmHg
Flow (Q): – mL/s
Edge state: –

All values at this time

3 · Plots

Final three beats. The vertical line is the animation time; click or tap a time plot to move it.

Pressures vs time

Volumes vs time

Valve flows vs time

Pressure–volume loops

4 · Summary metrics (computed by the engine)

All engine metrics

Metric	Value	Unit	Engine key

Parameter values used in this run

Parameter	Value

Agreement with the native engine

The same code path (runner.py) was run twice: natively (CPython 3.11, numpy 2.4.6, scipy 1.17.1) and inside this page's Pyodide (CPython 3.14.2, numpy 2.4.6, scipy 1.18.0), for the default scenarios of both models. Different numpy/scipy builds mean the results are not bit-identical; the differences below are the measured ones.

Loading comparison…

How this works

Pyodide runs the unmodified physiosim package (commit e3d560e) in your browser, in a Web Worker so the page stays responsive. Only the model and simulation modules are imported.
Nothing is sent to a server. Python, numpy, scipy and the physiosim wheel are served from this site; there are no cookies, storage or analytics.
Parameter values are exploratory design choices, not fitted to data. Phase 1 uses Parameters(); M1 uses the synthetic test fixture. Neither represents a patient or normal physiology.
The solver is the repository's event-aware integrator (ideal-diode valves; valve events located exactly). A fixed number of beats does not prove a steady state — see the beat-to-beat change reported with the metrics.
The models omit atrial contraction, AV delay, vessel inertance, valve regurgitation, pericardial and septal coupling, and any control or drug models.