fluid simulation fundamentals

laval

Before simulating a lava lamp, learn how a computer represents a fluid. A sustained thermal flow needs an energy source and an energy sink. The heater adds temperature while ambient cooling continuously removes it.

\n\n

\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n

\n\n

Click a cell to inspect its coordinates.

—selected cell

—array index

—cell value

0update steps

\n\n

section 1

A fluid needs an address

We divide a rectangular region into cells. Each cell has an x and y coordinate, just like a pixel.

neighbor blend per step25%

Each cell closes this percentage of the difference between itself and its neighbors.

distance per step0.75 cells

How far backward each cell looks along its velocity vector.

pressure iterations60

More iterations approximate the pressure solution more closely.

mean |divergence|: —

wall leakage

outward-pointing wall cells: —

buoyancy strength0.50

How much upward velocity one step adds in hotter cells.

heater per step0.12cooling per step2%

The heater adds temperature at the bottom; cooling removes this fraction from every cell.

temperature diffusion15%wall cooling8%

Conduction exchanges heat with neighbors; wall cooling removes heat only at the glass boundary.

const field = createScalarField(28, 18);\n\nfield.at(x, y);

What to noticeA two-dimensional cell can live in a one-dimensional array.

Deliberately not includedThere is no motion, pressure, viscosity, or fluid physics yet.

\n\n

The path from this grid to a lava lamp

checkpoint 01Grid + scalar field

02Velocity vectors

03Advection

04Diffusion

05Divergence

06Pressure projection

07Boundaries

08Heat + buoyancy

09Heat transport gap

10Coupled thermal step

11Heating + cooling

12Conduction + wall cooling

13Wax + liquid phases

Current claim: this demonstrates a grid, a scalar field, and neighbor-based smoothing. It is not yet a fluid simulation.

Diff · v145