Many times we don´t know why to use the Continuity Convergence Accelerator. This article tries mainly to give some understanding how the Continuity Convergence Accelerator works.
Details
There are simulations where the Coupled Flow model shows poor convergence, with regards to mass balance. This often leads to poor convergence and in some instances can lead to divergence of the simulation. For simulations involving multi-phase flow, this mass imbalance can lead to mass loss or unwanted mass generation. Improper mass imbalance can also lead to non-physical temperatures and hot spots in the simulation.
To understand the role of Continuity Convergence Accelerator, one has to understand what problem it is trying to address.
The Coupled Flow model upon applying discretization, assembles a block coupled matrix for momentum and continuity equation together. Thus, the linear system which is resulting, is then solved to some predefined tolerance (typically 0.1), which means that the linear system is considered converged when the residuals during solution process are below this tolerance, times the initial residuals. Or in other words, for a tolerance of 0.1, if residuals are below 0.1 of initial residuals, the system is considered converged.
The problem arises when the final residuals for momentum and continuity system, individually are not below the specified tolerances, even if the global system is deemed converged. To restate, this means that even if global coupled system is converged, the momentum and continuity might not have sufficiently converged.
The lack of convergence in continuity results in above mentioned problems.
Continuity Convergence Accelerator, solves this problem by solving for continuity equation after the global system has converged. This leads to tighter convergences in continuity and a more robust solution.