ABOVEBOUND / RESEARCH SYSTEMLOCAL COMPUTE
EXPERIMENT 002 / 3D PERPENDICULAR FLOW · REV 2

One flow. Connected through depth.

A 3D pressure correction couples all XY sheets. Water can move across the stack (W ≠ 0): one canvas shows in-plane arrows and perpendicular-flow signs, while another shows how that cross-plane motion changes with depth. The numerical solver runs locally.

01 / MAIN PHYSICAL SLICEAUTO SELECTING
CALCULATED VELOCITY / CROSS-PLANE FLOW
AMBER DOT = +Z · CYAN RING = −Z
W > 0: AMBER • · W < 0: CYAN ○
—PRESSURE PROJECTED 3D FIELD
02 / COUPLED NEIGHBOURING SLICESNOT SEPARATE UNIVERSES
9 CONNECTED XY PLANES
AMBER = +Z · CYAN = −Z · DARK = ~0
TAP A SLICE TO INSPECT
PHYSICAL DISTANCE SETS THE Z DERIVATIVESCONNECTED
t = 0.00FLUID AT REST∇·u = 0STARTING LOCAL WORKER
REQUIRED GEARS / AUTO-INFERRED FROM THE NUMBERS

The calculations pull their own dependencies.

Initializing…
What this model can and cannot claim

Implemented: non-dimensional 3D staggered-grid momentum (u, v, w), global pressure projection for incompressibility, impermeable walls, asymmetric bottom heating, buoyancy, 3D scalar advection and diffusion, pressure residual and mass-balance diagnostics. Every XY slice is a view into the coupled volume, not an independent simulation. Computation runs inside one local Web Worker.

Limitations: experimental coarse-grid scheme with approximate tangential no-slip, no interior membrane solid masks, no mesh-convergence study, no external benchmark certification, no compressibility. The adaptive pressure-correction gear chooses iterations; all physically coupled cells must still update. This is not a proof of the Navier–Stokes Millennium Problem.