Old and new values of a field

Franko · February 19, 2015, 18:07

Hi, I have a problem with using old value of a field.

I have A and B - dimensionless volScalarFields
I also have

Code:

dimensionedScalar c ("c", dimensionSet(0, 0, 1, 0, 0, 0 ,0), 1.0);

In the beginning, I know

and

,

I want to solve two equations below in OpenFOAM (n - number of time layer):

1) $A_i^{n+1} = 2*B_i^n$ ;

2) $\frac{A_i^{n+1}B_i^{n+1} - A_i^nB_i^n}{\tau} + B_i^{n+1} = 0$

(I want to find $A_i^{n+1}$ from the first equation and then $B_i^{n+1}$ from the second one)

I used this code with Euler scheme:

Code:

A = 2*B;

fvScalarMatrix BEqn
(
c*fvm::ddt(A,B) + fvm::Sp(1,B)
);
BEqn.solve();

The problem here with row

Code:

A = 2*B

This row let me find $A_i^{n+1}$ but it also deleted old

values because OpenFOAM took the second equation as

$\frac{A_i^{n+1}B_i^{n+1} - A_i^{n+1}B_i^n}{\tau} + B_i^{n+1} = 0$

And I got the wrong decision!

What code will solve $A_i^{n+1} = 2*B_i^n$ equation and won't delete old

values?
Thank you very much for any help ...

Franko · February 24, 2015, 07:42

So, what will be the best way to solve the equations without any derivatives in OpenFOAM?
For example look at the equation

(A and B - volScalarFields).

Do you use this code to solve it:

Code:

A=2*B;

??
Is that correct code?
Please help!

ssss · February 27, 2015, 05:06

Are you going to use always that kind of temporal discretization? If the answer is yes, you can access runTime delta T:

Code:

dimensionedScalar mydeltaT1("mydeltaT",dimensionSet(0, 0, 1, 0, 0, 0, 0),scalar(runTime.deltaTValue()));

And then solve manually the set of equations.

February 19, 2015, 18:07	Old and new values of a field	#1
Franko New Member Join Date: Sep 2014 Posts: 15 Rep Power: 12	Hi, I have a problem with using old value of a field. I have A and B - dimensionless volScalarFields I also have Code: dimensionedScalar c ("c", dimensionSet(0, 0, 1, 0, 0, 0 ,0), 1.0); In the beginning, I know and , I want to solve two equations below in OpenFOAM (n - number of time layer): 1) $A_i^{n+1} = 2B_i^n$ ; 2) $\frac{A_i^{n+1}B_i^{n+1} - A_i^nB_i^n}{\tau} + B_i^{n+1} = 0$ (I want to find $A_i^{n+1}$ from the first equation and then $B_i^{n+1}$ from the second one) I used this code with Euler scheme: Code: A = 2B; fvScalarMatrix BEqn ( cfvm::ddt(A,B) + fvm::Sp(1,B) ); BEqn.solve(); The problem here with row Code: A = 2B This row let me find $A_i^{n+1}$ but it also deleted old values because OpenFOAM took the second equation as $\frac{A_i^{n+1}B_i^{n+1} - A_i^{n+1}B_i^n}{\tau} + B_i^{n+1} = 0$ And I got the wrong decision! What code will solve $A_i^{n+1} = 2*B_i^n$ equation and won't delete old values? Thank you very much for any help ...

February 24, 2015, 07:42	simple equation	#2
Franko New Member Join Date: Sep 2014 Posts: 15 Rep Power: 12	So, what will be the best way to solve the equations without any derivatives in OpenFOAM? For example look at the equation (A and B - volScalarFields). Do you use this code to solve it: Code: A=2*B; ?? Is that correct code? Please help!

February 27, 2015, 05:06		#3
ssss Senior Member anonymous Join Date: Aug 2014 Posts: 205 Rep Power: 13	Are you going to use always that kind of temporal discretization? If the answer is yes, you can access runTime delta T: Code: dimensionedScalar mydeltaT1("mydeltaT",dimensionSet(0, 0, 1, 0, 0, 0, 0),scalar(runTime.deltaTValue())); And then solve manually the set of equations.