Hi Everyone, I am using CFX 11 package. I want to inject dye only for 2 seconds from domain inlet(already predicted flow field, I turned off the fluid solver from expert parameters, so it is just transport equation of additional variable)and get total concentration curve at the outlet.

Here is the two seconds injection of dye. "DyeMassFlow = (step ((2[s]-t)/ (1[s])))*(1[kg])/(2[s])"

Here is the additional variable. "ADDITIONAL VARIABLE: Dye

Option = Definition

Tensor Type = SCALAR

Units = [kg/s]

Variable Type = Volumetric"

And here is the total concentration at the outlet

"TotalDye = sum(Dye)@Outlet" The problem is I am injecting 1 kg dye as a total, however the total I am geting at the outlet is far more than the value I am injecting.

Can anyone tell me what I am doing wrong? Thanks a lot Regards, Mike

Hi Mike,

Your scalar units shoud be [kg/m^3] for volumetric or [kg/kg] for specific. Also, the step() function will return 0.5 when the operand is zero, so you may want to write it as (step ((1.9999[s]-t)/ (1[s])))*(1[kg])/(2[s]) so the value at 2 [s] is exactly zero.

If you apply this at the inlet, the dye mass flow rate will be a function of the inlet flow rate (because the flow is transporting it). If you solve it as a specific variable [kg/kg], the dye flow rate will be equal to the dye mass fraction (it's value) times the mass flow rate through the inlet. If you set it up as a volumetric AV, it will be the inlet volume flow rate times the volume fraction (i.e. (mass flow/density)*volume fraction).

Alternatively, you could specify the dye as a source injection at the inlet, as opposed to a value. In this case you would use the function as you have it now and the local concentration will depend on the flow rate.

-CycLone

Thank you so much,CycLone.

Hi,

Also the "sum" function is not the correct way of getting the amount of AV in a volume or across an interface. The sum function just does a simple sum at all nodes. You should use one of the integrating functions, such as volumeInt() or areaInt(). You might also want to use a mass flow function massflowInt() which integrates against mass flow across an interface rather than the area integral function.

Regards, Glenn

Thanks a lot, Glenn.