SYMPTOM:
In some chemistry calculations, it is possible to use the 'Equilibrium' reaction option despite it being physically invalid. The CFX-Solver treatment for this changed in CFX-11.0, viz.:
1) In CFX-10.0 and earlier versions of CFX-5, and also in CFX-4, the recipe for 'Equilibrium' is applied regardless of it being physical invalid. However, this has significant drawbacks:
- The backward rate has wrong physical dimensions! The dimension are wrong by factor [m^3n mol^-n] whenre 'n' is a number of order O(1).
- Forward and backward rates to not balance in chemical equilibrium, i.e. the mixture composition does not converge towards the chemical/thermodynamic equilibrium, but will approach some other composition.
2) Since CFX-11.0 a correction to the 'Equilibrium' option is applied such that the backward rate has the correct dimensions and that the forward/backward rates balance when in equilibrium. The drawbacks in this approach are:
- The stiff chemistry solver may not converge in many cases, and there is no obvious way to fix this.
- The reaction mechanisms may have been developed assuming approach 1), because this is what everybody else is doing.
Unfortunately, users may expect treatment 1) to operate, despite its drawbacks.
PROBLEM:
This is due to an improvement to the software conflicting with user expectation.
|
WORKAROUND: No workaround is currently available. FIXED IN: ANSYS CFX Release 12.0. In CFX-12.0, users will be able to use an expert parameter to choose from the above options, with the default being that the code will stop if the equilibrium reaction is non-physical. |
||
|