Adjoint solution with PointSource as source term (f) for Poisson equation

[Sparrow_Captain]

BEMPP can solve inverse problems ? Or just the forward ?

Thank you for the suggestion, really helps.

[dokken]

@Sparrow_Captain

Ok i figured it out. But when you change the point sources positions, you don’t really get the adjoint for each point source, right ? you get the 2nd plot I showed you where it looks like a single ellipsoid between the two point sources.

There is no single adjoint for each point source. They make a combined point source contribution.
Recall that what you are solving for the adjoint is a poisson equation where the solution 2*(u-u_d) is a source term, it is not a direct mapping.

[Sparrow_Captain]

Yeah you are right. Thank you again for your entire help !

PS : You saw my question about BEMPP ?

Thank you !

[dokken]

@Sparrow_Captain

PS : You saw my question about BEMPP ?

As I am not the author or developer of BEMPP I do not know alot about all capabilities. @mscroggs would be in a better position to comment.