Dispersed Bubble Flow Simulation via Continuum-Stochastic Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for simulating dispersed bubble flow, especially with millions of small bubbles, are impractical due to high computational burden and system stability issues, as they require direct computation of interactions among vast collections of particles, which is not feasible with existing technologies.
Innovation Solution
A method that represents gas and liquid volumes using a level set surface on a grid mesh, models gas-liquid dynamics with incompressible Navier-Stokes equations, converts level set fields to fraction fields using Heaviside functions, and models bubble-bubble interactions as discrete random walks with Schlick's phase function, separating computation into continuum and stochastic solvers to efficiently simulate dispersed bubble flow.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If direct computation of interactions among vast collections of particles is used, then simulation accuracy is improved, but computational burden increases
Solution Approach 1:
The patent segments the simulation approach into two distinct methods: a continuum solver for computing average bubble motion and a stochastic solver for computing subgrid interactions. This segmentation allows each solver to handle specific aspects of the simulation, improving overall computational efficiency while maintaining accuracy through the complementary nature of the two approaches.
Solution Approach 2:
The patent introduces an intermediary representation where bubbles are modeled as massless particles in a continuum framework. This intermediary approach allows the system to capture essential bubble behavior without requiring direct computation of all particle interactions, thus reducing computational burden while preserving simulation accuracy.
2Measurement precision
If direct computation of interactions among vast collections of particles is used, then simulation accuracy is improved, but system stability deteriorates
Solution Approach 1:
By segmenting the simulation into continuum and stochastic solvers, the patent avoids the stability issues inherent in direct particle interaction computations. The continuum solver provides a stable framework for average motion, while the stochastic solver handles subgrid interactions in a controlled manner, preventing system instability.
3Productivity
If continuum solver is used for average bubble motion, then computational burden is reduced, but visual complexity may be lost
Solution Approach 1:
The patent merges the continuum solver and stochastic solver into a unified simulation framework. The continuum solver computes average bubble motion efficiently, while the stochastic solver adds visual complexity by computing subgrid interactions. This combination preserves visual complexity while maintaining computational efficiency through the complementary strengths of both solvers.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach allows for the realistic and efficient simulation of dispersed bubble flow, approximating average bubble motion as a continuum and computing subgrid interactions stochastically, reducing computational burden and maintaining visual complexity, enabling simulation of complex scenes with millions of microscale bubbles in less than 16 seconds per frame on a single PC.
Implementation Method 1
capturing and tracking the gas-liquid interface, which defines signed-distance function φ such that |φ|=1 for all domains and the interface is defined at φ=0
Implementation Method 2
converting a level set field to a fraction field using a Heaviside function; encoding volume information of media into the fraction field, which is a spatial average of each medium as a continuum
Implementation Method 3
modeling a gas-liquid dynamics by assuming viscous incompressible two-phase flow, wherein the gas-liquid dynamics is governed by an incompressible Navier-Stokes equation and a mass conservation equation
Implementation Method 4
modeling a gas-liquid dynamics by assuming viscous incompressible two-phase flow
Implementation Method 5
modeling bubble-bubble interactions as a discrete random walk and determining a direction of a bubble using a Schlick phase function
Implementation Method 6
defining a probability function over space where the scattering of a bubble might occur as s(x)=νρgas[1−f(x)]|u(x)|2
Implementation Method 7
advecting bubble particles according to velocity field
Data Source
AI summary
A simple and efficient method for simulating dispersed bubble flow is provided. Instead of modeling the complex hydrodynamics of numerous small bubbles explicitly, the method approximates the average motion of these bubbles using a continuum multiphase solver. The subgrid interactions among bubbles are computed using a new stochastic solver. Using the proposed scheme, complex scenes with millions of bubbles can be simulated efficiently.