Dye-Dilution Flow Loop (2D)

Plan view, space-time heat map and Stewart-Hamilton analysis of a dye bolus in a laboratory flow loop. Not a patient model.

Switch to 3D

Readouts

Presets

Loop

Left: top view of the loop (Blender render) with random-walk dye particles. Right: concentration along the loop (across) against time (downwards). The green line is the sensor position.

Independent numerical checks

The science

The indicator-dilution (Stewart-Hamilton) method measures flow without a flow meter: inject a known mass m of tracer upstream, record concentration c(t) downstream. Conservation of the tracer gives the flow: Q = m / ∫ c(t) dt (first pass only).

Advection-diffusion: dc/dt + u dc/dx = D d²c/dx², Pe = uL/D. For a pulse on an infinite line the fixed-position integral is exactly m/Q for any D. A closed loop is a periodic sum of images: c(x,t) = (m/V) Σk G(x + kL, t). The heat map is this function: the bolus is a diagonal band that spreads as √t and wraps around the loop; after several circulation times the colour becomes uniform at m/V (well mixed).

The semi-log graph shows why the tail must be cut: the true curve (orange) is contaminated by recirculation, so the estimate integrates to 30 % of the peak on the down-slope and extrapolates with the exponential fitted between 70 % and 30 % (the straight dashed line).

Simplifications, stated honestly. The loop is a straight-pipe, plug-flow-plus-dispersion model with one speed. Real circulation has branching, a valve-driven pulsatile flow, lung and organ mixing volumes and a changing dye concentration profile, none of which are modelled. Dye dose, loop volume and Peclet range are illustrative. The sensor curve is built from a few thousand random-walk particles (visible noise is real sampling noise). This is a physics teaching model of the measurement principle, not a diagnostic tool, and nothing here gives medical advice or makes any healing claim.

Top-view sprite: Blender (Cycles) render of the 3D page's bench model.

Sensor curve: simulated (cyan) vs analytic (orange), mg/L vs s
Semi-log: ln c vs s with the 70 % to 30 % tail fit (dashed white)
Made with 🟠 Blender