Exploring Partial Multigrid Solutions for Efficient CFD Simulations

In summary, the conversation discusses the challenge of solving Navier Stokes equations for simulation, particularly in terms of making the velocity field divergence free and solving a sparse linear system of equations. The two methods commonly used for solving the resulting Poisson equation are Pre conditioned CG and Multigrid, but there is a concern about the resulting pressure gradient and the limited computing resources available. Possible solutions include solving the system only partially or inaccurately at certain steps, and borrowing ideas from frameless rendering to optimize calculations. The conversation also mentions Multigrid and Conjugate Gradient methods.
  • #1
curiousOne
21
1
Hey everyone,
When solving Navier Stokes equations for simulation, one usually has to make the velocity field divergence free and solve a sparse linear system of equations. The Poisson equation that results is usually solved using either a Pre conditioned CG method or a Multigrid method.
Because the resulting pressure gradient is essential to the quality of the next simulation step (which is itself limited by Courant Friedriech Levy) is there any way that the system could be solved only partially ?
Normally, the problem is solved by taking a full step forward ( < CFL) and solving the pressure then repeating the iteration, using the pressure to advect the velocity.
The need for speed is at odds against limited computing resources, so the time step is chosen closer to the CFL and the Poisson system is solved once for each step. This way the error is limited and the simulation remains accurate.
The result however is a sparse set of pressure gradients at each step. Could someone solve for the pressure only in some areas of high velocity, or vorticity, re-using previous results to complete the system before solving ?
Could the system be solved inaccurately (say half the multigrid steps) at the half step mark, providing twices as many frames of semi-accurate pressure gradients ?
I'm borrowing these ideas from the emerging world of frameless rendering, where optimizations have been made to identify areas that need re-calculation and those than can be re-used.
J.D.
 
Physics news on Phys.org
  • #2
Is anyone familiar with Multigrid or Conjugate Gradient methods ?
 

FAQ: Exploring Partial Multigrid Solutions for Efficient CFD Simulations

What is Partial Multigrid for CFD?

Partial Multigrid for CFD (Computational Fluid Dynamics) is a numerical method used to solve partial differential equations that arise in fluid flow problems. It combines elements of both multigrid methods and domain decomposition methods to achieve faster and more efficient solutions.

How does Partial Multigrid for CFD work?

Partial Multigrid for CFD works by dividing the computational domain into smaller subdomains and solving the problem on each subdomain using a multigrid method. The solutions from each subdomain are then combined using a domain decomposition method to obtain an accurate solution for the entire domain.

What are the advantages of using Partial Multigrid for CFD?

Partial Multigrid for CFD offers several advantages, including faster convergence rates, improved accuracy, and better scalability for large problems. It also allows for parallel computing, making it suitable for high-performance computing environments.

What types of problems can be solved using Partial Multigrid for CFD?

Partial Multigrid for CFD is commonly used to solve problems involving fluid flow, such as in aerodynamics, hydrodynamics, and heat transfer. It can also be applied to other types of problems that involve partial differential equations, such as in structural analysis and electromagnetic simulations.

Are there any limitations to using Partial Multigrid for CFD?

Partial Multigrid for CFD may not be suitable for all types of problems. It is most effective for problems with smooth solutions and may not work well for problems with discontinuities or sharp gradients. Additionally, it requires careful selection of parameters and may not always outperform other methods for certain types of problems.

Back
Top