[Goenitz]

Hi

I would like to use CFX variables like temperature, density, mass fraction, thermal conductivity.

For example, expression is Mass flux=100*T*Random (suppose) T being the temperature at that surface, line or point.
So can I write in CEL expression for fluid flux like that at wall (as boundary condition) as it is?
e.g.
LIBRARY:
CEL:
EXPRESSIONS:
MassFlux=100*T*Random (or 100*temperature*Random)
END
END
END

LIBRARY:
CEL:
EXPRESSIONS:
Random=0.00007 [kg m^-2 s^-1 K^-1]
END
END
END
and CFX will incorporate temperature into this expression calculated by its solver?
Thanx

[ghorrocks]

As long as the units check out CFX will use it. But why didn't you just try this in CFX and find out for yourself?

[Goenitz]

Originally Posted by ghorrocks

As long as the units check out CFX will use it. But why didn't you just try this in CFX and find out for yourself?

I tried and Total temperature keep diverging... w/o this boundary condition, simulation works fine. However, it works till 10 iteration so I can see that expression takes T values and increase/decease mass flux at that boundary. But as solution is not converged, so I cannot say definitely.

[Goenitz]

Originally Posted by ghorrocks

As long as the units check out CFX will use it. But why didn't you just try this in CFX and find out for yourself?

Anyway, I am more interested if actually it takes temperature value by this simple expression.

I took inspiration from tutorial 16 'reacting flow in mixing tube'. It had expressions like this.
But I have less faith as when I did tutorial 21 'Combustion in can...', I changed reaction rates to zero but it didn't effect on combustion process.

[ghorrocks]

Getting CFX to accept an expression and use it is a different thing to it being numerically stable and you being able to get a converged solution. It is hard to see a physical process where the flow rate is proportional to the temperature, and non-physical processes often lead to simulations which do not converge.

So the question is whether the mass flow rate condition you propose is appropriate.

[Goenitz]

Originally Posted by ghorrocks

Getting CFX to accept an expression and use it is a different thing to it being numerically stable and you being able to get a converged solution. It is hard to see a physical process where the flow rate is proportional to the temperature, and non-physical processes often lead to simulations which do not converge.

So the question is whether the mass flow rate condition you propose is appropriate.

yes, the mass flow condition wasn't appropriate so solution was diverging.
I want to set a boundary condition of Temperature on wall. it is ∂T/∂y=-(∑RHi)/λ. (R reaction rate, H enthalpy, λ thermal conductivity, w mass fraction, i specie)
λ=∑wi*λi (I have CH4,CO2, H2O,... as species) So i defined the following.

ThermCond=CH4.mf*CH4.cond+CO2.mf*CO2.cond+CO.mf*CO .cond+H2.mf*H2.cond+H2O.mf*H2O.cond+N2.mf*N2.cond

1. is that correct way as CFX is not recognising it.
2. How can I assign expression ∂T/∂y on a fixed wall temperature option? is that legit?

[Opaque]

It seems you are trying to impose a heat flux proportional to the reaction rate at the boundary, i.e.

q = - k dT/dn @boundary = Sum over i (H_i * R_i)

There is no need for the thermal conductivity of the mixture since CFX computes it if needed, neither the temperature gradient.

The only time both of those quantities are needed (K and dT/dn) is when backing out the boundary temperature once the equations are solved.

On the boundary of interest, select Heat Flux in, set the Heat Flux parameter to an expression equal to the summation on the RHS of the equation above.

The solution may tend to diverge if you use a large timescale, so reduce it until it converges. If you are successful at it, let us know.

[Goenitz]

Originally Posted by Opaque

If you are successful at it, let us know.

Hi,

No Success yet as there are other problems. But even forcing Flux as your suggested expression, doesn't work.

My expression is:
LIBRARY:

CEL:

EXPRESSIONS:

Arr= 390 [mol kg^-1 s^-1 Pa^-0.47]

Eact = 43200 [J mol^-1]

kconst=Arr*((CH4.p)^0.46)*((H2O.p)^0.01)

Rate= kconst* e^(-Eact/R/T)

END

END

END

Will the above expression capture T and partial pressure or I have to incorporate AreaAve(CH4.p)@Surface?

Another things is CH4.p for partial pressure is not recognised by CFX so I use CH4.mf*p for time being. Because Dalton Law says CH4.p=CH4.mol/mol_total*p_total.

Going through reference guide, I can say that when user assigns p,T, density etc then these are bulk values.

P.S I assigned Sum over i(H_i*R_i) (Sum over=areaAve=sum) but could not succeed. Actually it can operate one variable e.g. sum(p), sum(T) etc
Error is:

The function 'sum' referenced by parameter 'Heat Flux in' in object '/FLOW:Flow Analysis 1/DOMAIN:Fluid1/BOUNDARY:SinkF1/BOUNDARY CONDITIONS/HEAT TRANSFER' has an invalid argument, 'new'. Only arguments that consist of a single recognised variable name are supported by the solver.

[Goenitz]

Originally Posted by Opaque

If you are successful at it, let us know.

Though I am not successful converting gradients into flux , but the code is working.