Project Study

A browser-based dam-break simulation reframed as a cleaner technical case study.

Originally developed in 2023 for an advanced graphics course, this revision keeps the strongest idea from the work, particle-based fluid motion with accelerated neighborhood queries, and presents it with a calmer, more inspectable interface.

JavaScript + Canvas Particle fluid solver Spatial hashing Interactive controls

Scope

Rework an older assignment into something that reads as simulation work rather than a course submission.

Method

Keep the dam-break setup, preserve local spatial hashing, and rebuild the presentation around inspection and control.

Outcome

A standalone browser demo that lets viewers compare presets, tune solver parameters, and disturb the flow directly.

Overview

This piece sits at the intersection of simulation and visual computing: a fluid block is released behind a wall, particles accelerate under gravity, local density is corrected through nearby interactions, and the entire scene runs directly in the browser. The portfolio version focuses less on assignment framing and more on what makes the build interesting: real-time behavior, spatial partitioning, and interactive inspection.

What Changed In The Portfolio Version

  • The project now carries a descriptive title rather than an assignment label.
  • The demo has been rebuilt as a small standalone artifact with its own interface, presets, and live metrics.
  • The presentation highlights the engineering decisions that matter most to a reviewer: particle count, solver substeps, collision behavior, and neighborhood acceleration.
  • The interaction model is stronger: viewers can pause, reset, switch density presets, and drag through the water to disturb the simulation.

Technical Notes

  • Neighborhood queries are accelerated with a spatial hash so the solver only considers nearby particles rather than every particle pair.
  • The scene is intentionally bounded and stylized: a retaining wall, a small house volume, and a tuned world scale make the motion easy to read.
  • The portfolio pass prioritizes clarity and responsiveness over strict archival fidelity to the original 2023 submission. It is a cleaner reinterpretation of the same core idea.

Next Steps

  • Add a second comparison mode that visualizes density or velocity as a heat map.
  • Record a short benchmark table that compares each preset’s particle count against average frame time.
  • Extend the scene with alternate obstacle layouts so the project reads more like a small simulation lab.