[solalito]

Hi,

I'm trying to implement a combustion model as a UDF.


Say I define a variable xi at every active node of the domain. How do I compute d(xi)dx_i?
From examining the 'include/csi_libraries.h' file, I've found this function:
calc_grad_global_scalar(double *grad_var, int kk, PRECISION *global_var, int ndim, int idim));

What exactly are ndim and idim? The only time this function is used in 'example_src/user_turbulence_model.c'. "ndim" seems to be the size of the array passed, but I'm not sure what idim is for.

Also, given the temperature, pressure and mass composition of a cell, can I retrieve the reaction rate?

Thank you!

[Dan.Maciejewski]

For CONVERGE 2.4 (and 2.3), you've identified the correct function to calculate the gradient: SHARED_calc_grad_global_scalar_node.
This function calculates the gradient of a scalar or vector. The gradient operator raises the dimensionality: gradient of a scalar is a vector; the gradient of a vector is a tensor. The last two arguments relate to the dimension of the input variable and which component of the result is of interest.

1st arg: pointer to an array to hold the computed gradient value
2nd arg: cell index
3rd arg: pointer to variable who's gradient you want to calculate
4rd arg: dimension of the variable you want to calculate
5th arg: index of the first component of the gradient you want to store.

To calculate the gradient of a scalar, ndim = 1 (one value per cell), and idim = 0, since you want to store the first three components of the gradient in your new variable; these 3 components are stored in a 1D array, so the first three elements are the 3 components of the gradient of the first cell, next three elements are for the 2nd cell, etc.

If taking the gradient of a vector, ndim = 3, and idim may be non-zero. Please contact support@convergecfd.com if taking the gradient of a vector (or tensor).

[solalito]

Thank you for your help!

Originally Posted by Dan.Maciejewski

For CONVERGE 2.4 (and 2.3), you've identified the correct function to calculate the gradient: SHARED_calc_grad_global_scalar_node.
This function calculates the gradient of a scalar or vector. The gradient operator raises the dimensionality: gradient of a scalar is a vector; the gradient of a vector is a tensor. The last two arguments relate to the dimension of the input variable and which component of the result is of interest.

1st arg: pointer to an array to hold the computed gradient value
2nd arg: cell index
3rd arg: pointer to variable who's gradient you want to calculate
4rd arg: dimension of the variable you want to calculate
5th arg: index of the first component of the gradient you want to store.

To calculate the gradient of a scalar, ndim = 1 (one value per cell), and idim = 0, since you want to store the first three components of the gradient in your new variable; these 3 components are stored in a 1D array, so the first three elements are the 3 components of the gradient of the first cell, next three elements are for the 2nd cell, etc.

If taking the gradient of a vector, ndim = 3, and idim may be non-zero. Please contact support@convergecfd.com if taking the gradient of a vector (or tensor).