[Daaman]

Hi,

I want to know what do +/- regions in the adjoint surface sensitivity map signify in terms of normal surface displacement. Would normal displacement of the positive regions, outwards, and negative region, inwards, result in an increase in the objective function? Or is it other way round?

I have attached the surface sensitivity map of Onera M6 test case with drag as the objective function and displacement in the +z direction. Looking at the extremities of the sensitivity map as well as the direction of the resultant vectors, drag increases when - the positive region is normally displaced inwards and the negative region, outwards. This is the opposite of what I had stated previously.

I am confused about the sign convention and would appreciate your inputs.

Thanks!

PS: I have used uniform vectors only for the purpose of visualizing the directions.

[pcg]

Hi,
If you were to move the surface along the direction pointed by the arrow the objective function should increase.

[Daaman]

Thanks Pedro! I was curious to see how it worked and therefore deformed the FFD box, and the results are in line with what you said.

I am trying to generate a similar surface sensitivity map for a propeller blade with thrust (Force-x) as the objective function. The direction of flow is in the +x direction and thus, the thrust is produced in the -x direction. The Cfx from the force breakdown file is negative as well. So, if I want to perform an adjoint sensitivity analysis for Force-x (Thrust) which is in the -x direction in my case, would it be appropriate to change the signs of all the resulting adjoint variables, sensitivities, and gradients from positive to negative?

[lalatuobu]

Originally Posted by Daaman

Hi,

I want to know what do +/- regions in the adjoint surface sensitivity map signify in terms of normal surface displacement. Would normal displacement of the positive regions, outwards, and negative region, inwards, result in an increase in the objective function? Or is it other way round?

I have attached the surface sensitivity map of Onera M6 test case with drag as the objective function and displacement in the +z direction. Looking at the extremities of the sensitivity map as well as the direction of the resultant vectors, drag increases when - the positive region is normally displaced inwards and the negative region, outwards. This is the opposite of what I had stated previously.

I am confused about the sign convention and would appreciate your inputs.

Thanks!

PS: I have used uniform vectors only for the purpose of visualizing the directions.

Could you post the cfg file for the case?