Stabilized Explicit Fluid Simulation on Irregular Grids

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Computational fluid dynamics simulations, particularly using the Lattice Boltzmann Method, face inefficiencies due to the Courant-Friedrichs-Lewy (CFL) constraint, which requires extremely small time-steps on explicit schemes when dealing with complex geometries and irregular grids, leading to increased computational costs and instability issues.

Innovation Solution

A modified heat flux calculation approach is introduced, which accounts for spatially averaged temperature gradients and material/geometric properties, allowing for stable numerical solutions regardless of element size, and reduces to explicit scheme implementation when feasible, thus overcoming CFL constraint limitations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If explicit time-marching schemes are used for numerical simulation, then implementation simplicity and computational efficiency per iteration are improved, but the time-step size must be extremely small to satisfy the CFL constraint on irregular grids

Engineering Contradiction:
Improveimplementation simplicityVSAvoidtime-step size
Core Design Contradiction:
Ease of manufactureVSLoss of time

Solution Approach 1:

The patent changes the parameter relationship by introducing a modified time-step constraint that depends on the ratio of face area to volume (A/V) rather than solely on the smallest spatial dimension. This allows the time-step to be determined by local geometric properties of each control volume, enabling larger time-steps in regions with larger volumes even when small faces exist, thereby resolving the contradiction between explicit scheme simplicity and the restrictive time-step requirement

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent applies local quality by allowing different time-step sizes for different control volumes based on their local A/V ratio. Instead of using a global time-step constrained by the smallest element in the entire domain, each control volume can use a locally optimized time-step, enabling regions with larger volumes to use larger time-steps while maintaining stability at interfaces with smaller faces

Inventive Principle:
Principle #3Local quality

2Manufacturing precision

If the spatial grid size is decreased to resolve complex geometries, then geometric accuracy is improved, but the time-step size must decrease by a factor of F^2 to maintain stability

Engineering Contradiction:
Improvegeometric accuracyVSAvoidtime-step size
Core Design Contradiction:
Manufacturing precisionVSLoss of time

Solution Approach 1:

The patent changes the scaling relationship between spatial resolution and time-step size by using the A/V ratio as the governing parameter. When the spatial grid is refined to improve geometric accuracy, the time-step decreases proportionally to the linear dimension (F) rather than the square of the linear dimension (F^2), significantly reducing the computational burden of fine geometric resolution

Inventive Principle:
Principle #35Parameter changes

3Loss of time

If implicit time-marching schemes are used to avoid CFL constraint, then large time-steps can be used, but computational cost increases due to solving large systems of matrix coefficients

Engineering Contradiction:
Improvetime-step sizeVSAvoidcomputational cost
Core Design Contradiction:
Loss of timeVSPower

Solution Approach 1:

The patent uses a cheap explicit scheme that requires minimal computational resources per iteration, accepting that many iterations may be needed, rather than using expensive implicit schemes that require solving large matrix systems. The modified A/V-based time-step constraint enables the cheap explicit scheme to achieve stability without the high computational cost of implicit methods

Inventive Principle:
Principle #27Cheap short-living objects (Disposable)

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 enables the use of a single time-step size based on time accuracy, reducing computational costs and maintaining spatio-temporal accuracy, while avoiding the need for small time-steps and complex grid generation, thus improving simulation efficiency and stability.

Implementation Method 1

The advection process involves modeling movement of particles from one location to another according to the particles microscopic velocities

Methodology Applied
Scientific EffectAdvection: Advection

Implementation Method 2

The collision process involves interactions among particles obeying conservation laws and to relax to an equilibrium

Methodology Applied
Scientific EffectCollision process:

Implementation Method 3

Numerical simulation of diffusion-dominated physical phenomena is common used for applications involving conductive heat transfer

Methodology Applied
Scientific EffectConductive heat transfer: Conduction (thermal)

Data Source

PatentEP3751444B1Computer simulation of physical fluids on irregular spatial grids with stabilized explicit numerical diffusion
Publication Date: 2024.08.14 DASSAULT SYSTEMS AMERICAS CORP
  • EP3751444B1 patent drawingFigure 1
  • EP3751444B1 patent drawingFigure 2
  • EP3751444B1 patent drawingFigure 3~4

AI summary

Computer implemented techniques for simulating a fluid flow about a surface of a solid, includes receiving a model of a simulation space including a Boltzmann lattice structure represented as a collection of voxels and a representation of a physical object, with the voxels having appropriate resolutions to account for surfaces of the physical object, simulating movement of particles in a volume of fluid, with the movement of the particles causing collisions among the particles, identifying by the computing system faces between two voxels where at least one of the faces violates a stability condition, computing a modified heat flux using a spatially averaged temperature gradient in the vicinity of the two voxels, where the at least one of the faces violates the stability condition and performing by the computing system, advection operations on the particles to subsequent voxels.