What is fluid simulation?

The core idea

Real fluids push, swirl, and mix. A simulation stores two basic things, velocity and density, across a grid or a set of particles. Velocity says how fast and in what direction each point moves. Density says how much dye, smoke, or liquid sits at that point.

Each step the solver updates those values using rules based on fluid dynamics, the branch of physics built on the Navier-Stokes equations. Real engineering software solves those equations with care for accuracy. Real-time tools use faster shortcuts so the screen can update dozens of times per second while you interact.

The result still looks like a fluid because the shortcuts keep the parts people notice: swirling eddies, trails that stretch and fold, and motion that responds to force in a believable way.

A short history

Fluid simulation started as an engineering tool. Early computational fluid dynamics, usually shortened to CFD, grew from aerospace and weather research in the mid 20th century, when solving flow equations by hand became too slow for real problems.

Real-time fluid simulation is newer. In 1999 researcher Jos Stam published a method called stable fluids that let a flow field update every frame without blowing up numerically, even with a large time step. Games and later browsers adopted that method because it traded some precision for speed and stability. The live simulation on this site builds on that same lineage of techniques.

Two main methods

Most fluid simulations use one of two broad approaches. Each one solves the same underlying physics with a different data structure.

Grid based methods, also called Eulerian methods, store velocity and density on a fixed grid of cells and let fluid flow between cells. Particle based methods, also called Lagrangian methods, track many small particles that carry their own velocity and move through space. Smoothed particle hydrodynamics, often shortened to SPH, is a well known particle method used for splashy liquids.

• Grid methods: fast on a GPU, good for smoke, gas, and dye
• Particle methods: good for splashes, droplets, and liquid surfaces
• Hybrid methods: mix both to get stable gas motion and sharp liquid surfaces

Common uses

People use fluid simulation for very different jobs. The right tool for each job depends on how much accuracy you need and how much time you can spend waiting on a result.

• Visual effects and games, for water, smoke, fire, and magic style motion
• Product and architectural visuals, to preview airflow or water features
• Teaching, to show how pressure and swirl behave without a chalkboard
• Design exploration, to try an idea before committing to a full build
• Engineering analysis with specialist CFD software, for aerospace, automotive, HVAC, and pump design
• Weather and climate research, where fluid simulation models air and ocean currents at a huge scale

Where a browser WebGL tool fits

A browser fluid simulation trades some accuracy for instant feedback. On the live fluid simulation on this site, you paint with your mouse or finger and watch dye move through a field that keeps itself roughly incompressible using pressure projection. That matches play, demos, and light learning. It does not replace lab grade CFD.

Fluidsimulation.net focuses on that instant path. The whole simulation runs on your device inside the tab, using WebGL to push the math onto the GPU. Modes like water, fire, ink, and calm change the feel through different settings on the same solver, not through a new install or a new page load, and the step by step loop behind that GPU motion is covered in how WebGL fluid simulation works.

What a browser simulation does not do

A real-time WebGL fluid is honest about its limits. It does not output measured numbers you could use to size a pipe, certify an aircraft part, or pass a safety review. It uses a coarse grid and a fixed, small number of solver iterations per frame so it can keep up with your screen refresh rate. Engineering grade accuracy belongs to a different class of tool. See CFD vs browser fluid simulation.