The computation of the chemical reactions and of the species transport can be executed in the coupled and decoupled mode with the fluid dynamics model.

The coupled mode in the backward coupling sense is activated, if
at least one of two conditions is satisfied: either the
checkbox ** effect on gas convection**
activated or endothermic or exothermic reactions are
available. If the chemical model is coupled, then the execution
of the chemical model can start either by selecting of the
actual variable group **Chemistry** or
**Temperature** or
**Convection** or
**turbulent_Convection** and pressing
button.
In all cases the chemical model will be computed alternately with the
treatment of the global heat transport. The convection will be
iteratively computed each time after the chemical model is
updated for all variable groups excepted
**Temperature**.
The decoupled chemical model is computed after the variables
group **Chemistry** was selected and the
item **Computation** >
**Start Computation** is selected or also by clicking on
button.

Before the computation in the decoupled mode begin, first the divergence free solution for the fluid flow in the computational domain of the chemical model is checked. The balance of the mass fluxes is computed. If the mass conservation in the fluid flow satisfies to the conditions of the chemical model, then the chemical model assembly and execution start immediately.

The total mass flux balance error in all fluid
domains where the chemistry should be executed, should be
less than 1.e-10. Otherwise the program first is going to improve
the mass balance. A wrong mass balance leads to the
artificial effects as flow sources and sinks. These
numerical effects are dangerous because they can be
interpreted as effects of chemical reactions. On the other
hand sometimes it is very difficult to get a deep
convergence for the fluid flow. A typical
situation is the gas flow in the complex environment by high
Re and Gr numbers. Therefore a special hint is implemented
in *CrysMAS*.

For solution of the divergence free flow field problem the
SIMPLEC algorithm is utilized. The momentum equations are
assembled once and the coefficients for the pressure
correction equation are stored. Than the pressure correction
equation is assembled block wise. The matrix equations for the
unknown pressure correction are copied then from each numerical
block into the global matrix. The boundary conditions for the
pressure correction at the symmetry axis and at the outflow
boundary are accounted in the the global matrix equation
implicitly. The assembled pressure correction equation is
passed to the selected solver in the dialog window
**Computation** ->
** Numerical Parameters** ->
** Forward** dialog tab. Usually
the GSSV direct solver is applied. The pressure correction
equation is solved with precision allowed by the computer
accuracy
(remaining error about 1.e-15). Then the mass fluxes are
corrected according to the SIMPLEC procedure. The remaining
mass fluxes balance error corresponds to the precision of the
pressure correction solution and has the same order of
magnitude.

The corrected mass fluxes distribution corresponds to a slightly changed velocity field. The magnitude of the change is dependent on the residual error of the mass balance before the chemical model starts. Only mass fluxes will be changed by the correction, the velocity distribution remains not changed.

Only one step of the pressure correction procedure is required and the pressure correction equation should be solved once thanks to its linearity. As a result overall fluxes correction procedure takes only few seconds even for a large size 2D fluid dynamics problem.