Jacques Moss s5616052 CFGAA Assignment:

Report for my Assignment:

Please use FluidSimQT as that is the final save with added QT. However FluidSim is quicker efficiency wise.

Initial Idea:

I wanted to create an SPH fluid simulation that applies the physics off of water and how it works in real life. Using a particle simulation to replicate and mimic those movements and algorithms.

Video:

This video is the most updated video on how the fluid simulation works with the QT controls:

https://youtu.be/TiQEhSULAbs

How the fluid simulation looks with the optimum recommended values: https://youtu.be/GjH93HdGTUc

Iterations trying to work through QT problems: https://youtu.be/XdgDH1SLI6Q https://youtu.be/DP7xY7tK5vg

FluidSim (Non-QT version; with faster and more accurate fluid movement but no UI controls; all changes to variables have to be manual.): https://youtu.be/Kcz7_Tj5SKw

How to use:

Make sure you have all of the files in FluidSimQT and make sure they all organised as it is in the github repository.
Run "clion ." in the terminal within FluidSimQT.
Within Clion make sure to go to your build in the top right corner and link your copyshaders to the build using cmake target.
Press ctrl + f5/ build and run using the buttons in the top right corner.

What my Code does:

Uses ParticleSOA by Jon Macey to create a particle system.
Calculates density, pressure and viscosity and applys to the particles based off acceleration and closeness of other particles.
Creates and draws a boundary box that repels particles when they collide and dampens the force.
Uses a vertex and fragment shader in order to determine, normalize the density and colour those high density areas more red.
Created a QT GUI using BlankQTNgl and qt creqator. Applying yOffset, Pressure Multiplier and Viscosity Strength based off of input values from the user.

Variables and Usage:

pressureMultiplier in physics.h effects how particles are seperated based of the amount of particles (density) within a certain radius(smoothing kernel.) It can be changed to make the particles more repulsive agaisnt eachother.
viscosityStrength in physics.h effects the movement of the particles and how they flow in a more similiar direction when the viscosity is higher.
particleSpacing effects the distant between particles when initialized and therefore the amount of particles in the intialized cube at the start. The cube will stay the same size having the same dimensions unless changed manually which it can be in the initializeParticles function. However the spacing between particles can be changed. The particle spacing currently directly effects the smoothingKernel.
m_width, m_height and m_depth are the variables in BoundingBox.h that can change the physical dimensions of the m_boxVAO and how it is drawn and resolves collisions.**The passing through of these variables from QT is currently commented out as the m_BoxVAO isn't able to be redrawn properly a second time yet.
yOffset found in initializeParticles effects how high the cube is in the y-axis when initialzed.
There is alot of other variables but for user access these are the best variables to mess around with if you don't have any knowledge about SPH fluid simulations.

ALL variables work properly and can be changed manually in the code before running to effect the conditions. Only the drawing of the bounding box VAO when the initialisation button is pressed is not occurring properly as of yet; however if the values are changed before running the code the bounding box is drawn properly with those dimensions.

Structure of Directory:

src: main.cpp, NGLScene.cpp, NGLSceneMouseControls.cpp, Physics.cpp, Emitter.cpp, BoundingBox.cpp, MainWindow.cpp
include:MainWindow.h, NGLScene.h, NGLSceneMouseControls.h, Physics.h, Emitter.h, BoundingBox.h
CMakeList.txt
ui: MainWindow.ui
shaders: ParticleFragment.glsl, ParticleVertex.glsl

Evaluation and Potential Improvements:

I ran into an issue where my smoothing kernel value was directly effecting my density value which shouldn't be happening. Also my particles were slowly merging with the ground at y=0. I managed to temporarily fix the particles combining at y=0 by adding a minimum repulsive force from the floor. I had other errors that I ran into that I was able to completely and safely resolve but for these problems I had to work around them. All my QT values were eventually able to pass through to my variables in each respective header and adapt in real time the fluid simulation, but unfortunately I wasn't able to change the bounding box VAO and how it was drawn based of these variables, so I commented out the passing throiugh of the dimension variables. Overall if I had more time I know what I would adapt and try to fix to make my fluid simulation running more efficiently and better.

Uncomment this code in NGLScene.cpp if you want the inputted dimensions in the QT width, height and depth inputs to affect the collision resolving. However the drawn red box will not match this new collision boundary.

I would've liked to include a spatial hashing system in order to speed up calculations as less not all densities would be calculated and only the densities within the smoothing kernel.
I would've liked if my bounding box values fropm QT were able to affect the drawing off the bounding box VAO.
I wanted to allow the user to effect the fluid with repulsion and attraction mechanics around the mouse cursor.

Annotated Bibliography and Research:

Reference list

Lague, S., 2023. Coding Adventure: Simulating Fluids [online]. www.youtube.com. Available from: https://www.youtube.com/watch?v=rSKMYc1CQHE [Accessed 15 May 2025].
- A very useful youtube video which takes you through a simulation creation on unity of a 2D particle system, with the included maths and a great visual description of how that maths applies to the screen. This was one of my primary sources of research and I worked closely with this video in order to build my system.
Macey, J., 2025. Jon Macey’s WebPages [online]. Jon Macey’s WebPages. Available from: https://nccastaff.bournemouth.ac.uk/jmacey/ [Accessed 15 May 2025].
- All my professor Jon Macey's working resources and lessons with links to how he builds his code and the theory behind hsi code, QT and other resources.
NCCA, 2024. GitHub - NCCA/BlankQtNGL [online]. GitHub. Available from: https://github.com/NCCA/BlankQtNGL [Accessed 15 May 2025].
- A blankQTNGL template which I used as boiler plate code so that I could easily create my GUI in QT and then add my running code from clion and copy it over to this GUI code.
NCCA, 2025. GitHub - NCCA/labcode-jmacey-2: labcode-jmacey-2 created by GitHub Classroom [online]. GitHub. Available from: https://github.com/NCCA/labcode-jmacey-2 [Accessed 20 Apr 2025].
- I used Particle SOA a template for Jon Macey's particle system which I made in class with him and then built off of this code the rest of my fluid simulation code. Manipulating and changing how the particles are drawn to the screen, coloured and updated thrloughout this scene and adapting the code to work better for me.
Weaver, T. and Xiao, Z., 2016. Fluid Simulation by the Smoothed Particle Hydrodynamics Method: A Survey [online]. Available from: https://eprints.bournemouth.ac.uk/23384/1/2016%20Fluid%20simulation.pdf [Accessed 9 Apr 2025].
- This is a master's paper from a Bournemouth Universiy alumni which helped me understand, in tangent with the Lague Youtube video, what algorithms and equations are needed to create and effect my particles making it move like a fluid. This paper is very in-depth about the mathematical background behind how fluid simulations work and how to break down the overall maths into funtions and steps in order to create the overall project.

NCCA/programming-assignment-DeathsIncarnate

ErrorLooks like something went wrong!

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Languages

Error
Looks like something went wrong!

Packages