[qutadah.r]

Hello, I have a question regarding how openfoam solves equations for simulation. If someone can guide me through my thought process and tell me if its right... Please point me out if I have mistakes in my understanding.



1- we have a computational mesh or domain which we know with its vertices or nodes.


2- Discretize the NS equations using the FVM approach and make a local equation for each cell in the mesh using one of the discretization methods for each term in the equation. All will be cell-centered based.



3- Interpolation methods is used for the diffusion term and only that term to find the face values because we need them in the equation.



4- we need to make a global matrix and solve it iteratively using one of the iterative solvers, where we under-relax specific terms if they increase instability in solving our matrix. Where does preconditioning come into all of this?


*constructing a global matrix of 3 million cells in a steady state incompressible flow means we have 4 unknowns at each cell, which means 12 million unknowns in the equation?



5- We use one of the Algorithms to solve the pressure/velocity coupled momentum and continuity equations, and find the respective fields for each.


6- We solve using Initial and boundary conditions in our 0/ folder, however these values are also cell-centered.


7- nonorthogonal correctors are used to determine how many times the pressure equation is solved in a pressure/velocity coupling algorithm like the simple/piso.



8- we keep solving until we get a converged solution based on our residuals.


Thanks!!