Zan Ahmad · Computational Cardiology

Conservation, viscous flow and transport

Begin with a control-volume budget, derive pressure-driven motion between stationary walls, and use that velocity to transport a passive tracer. The calculation separates a reduced momentum equation from the subsequent transport problem.

Velocity from the pressure gradient

The steady fully developed channel equation has a quadratic solution. The finite-difference solution reproduces its nodal values to roundoff; integrating those values introduces a separate quadrature error.

Axial channel speed is zero at both walls and reaches 12.5 millimeters per second at the middle of a two millimeter gap.
A planar channel benchmark with stationary walls and a constant pressure gradient.

Transport through a shear flow

A short inlet pulse moves faster near the center of the channel. Diffusion spreads the normalized concentration across the shear. Every interior numerical face contributes equally and oppositely to its neighboring cells.

Normalized tracer concentration after approximately one second, plotted against axial position from zero to twenty millimeters and gap position from zero to two millimeters.
The axes have different scales. Concentration is normalized by the imposed inlet pulse level.

Reproduce and examine the calculation

With NumPy, SciPy and Matplotlib installed, run python channel_transport.py --output results. The notes specify the boundary conditions, sufficient time-step restriction, experiments and exercises. Compare changes in physical diffusivity with changes in mesh resolution.

Experiments and exercises · Return to the momentum lecture