I'm confused that in my sloving problem there is a gas-moisture mixture with combined thermal and concentration force in my momentum equation. However, when I validate my problem by using a published paper which is a double diffusive cavity, I can't get the exact profiles whatever streamlin, isothermal line, and iso-concentration in aiding or opposing buoyany flow. Therefore, I begin check the pure thermal buoyancy driven flow in a square cavity, and a pure concentration byoyancy driven flow. However, in a pure thermal buoyancy driven flow, I can get a similar velocity and temperature distribution. But, in a pure concentration difference flow within a square cavity, there is no flow driven in my cavity.

Therefore, how can I turn on or add my buoyancy force due to concentration difference to get a combined gas-moisture problem when the Boussinesq approximation is valid.

As I recall the Buossinesq buoyancy force provided as an option in PHOENICS is a function of temperature difference only. Are you using this option? Or have you implemented your own buoyancy source term as a function of both temperature and concentration difference? It seems to me that you must do the latter either by use of the INFORM facility or by use of GROUND coding.