Problem with groovyBC ---urgent
 

Hi Foamers,
I am solving transient heat conduction problem. For the boundary condition I have used the heat flux boundary condition as follows---

dT/dx = 2.92e5*exp(-10e4*pos().x*pos().x) (just the gaussian beam at the upper surface )
But what the problem is for me now, I am getting the temperature greater than the published paper that I have decided for validation. I have checked the grid , and all sorts of stuffs. I have also done other simulations but for all the problem I am getting greater value than the other software(COMSOL), for example the peak temperature for the COMSOL I am getting 950K and for OpenFOAM I am getting 1260K.
I struggling with the problem..............
any idea????
or any one who has worked with the gaussian beam in OpenFOAM pls help me out of this problem....................

No idea.

The gradient groovyBC calculates gets written (the refGradient-field). The first step would be to check if this is the gradient you expect from your reference simulation.

BTW: what I noticed in your expression is 10e4 ... which is an OK, but rather unorthodox way of writing 10^5. Sure this is not a typo?

Is your expression really the thermal gradient, or is it the heat input? Latter has to be divided by the thermal conductivity before you plug it into groovyBC.

@gschaider: I have checked the gradient with replayTransientBC... at each position '(pos().x)' it's giving the value correct. But I am little bit confused about the gradient calculation: as I am giving the value at every face centers(boundary) and its calculate the gradient, and between two face centers there is vertex for which the value is not imposed, and this question arises to my mind as I am using 'exp' function. Please let me know whether I am wrong/right?? And if I am correct how can I imposed it??
@akidess: it's a heat input and I had divided it with thermal conductivity. To my best knowledge, you may have worked with the Gaussian Heat Source, could you please tell me how you did this in openFoam?

Indeed I have used Gaussian distributions in groovyBC as well. I have a gradientExpression which is the heat input divided by the thermal conductivity, a fractionExpression of 0. The heat input is a function of the form q(x) = a*exp(-b*x^2). There is nothing obviously wrong in what you have posted.

For a volume field the gradient is only imposed on the faces (and for point-fields there is no gradient BTW)