Derivation of the Boundary Integral Representation of the Solution to Laplace’s Equation
Similar to when deriving the weak form, multiply Laplace’s equation with a test function υ and integrate by parts:
Repeat with the roles of u and υ reversed:
Now subtract the first equation from the second equation and collect terms:
This relationship is known as Green’s second identity.
Now, use the fundamental solution by setting
with integration over x:
Using the property of the fundamental solution gives, for the first term in the volume integral:
For the second term of the volume integral, since u is assumed to be a solution to Laplace’s equation −∇ · (∇u) = 0:
The resulting relationship is known as the representation of u in terms of boundary integrals:
for y inside Ω
The value of u over the entire domain Ω is thus determined completely by the values of u and the flux n · ∇u on the boundary. The integrals are not singular anywhere inside Ω since the integral for each y is taken over x on the boundary where x ≠ y. Furthermore, note that this relationship is not an equation but merely a representation of u when we already know the solution and the flux on the boundary. Indeed, this representation is used to reconstruct the solution anywhere inside Ω once the field and flux are solved for and known on the boundary.
The representation of u can be written more compactly as:
in Ω
where nu = n · ∇u.
The integral operator
is called the single-layer volume potential.
The integral operator
is called the double-layer volume potential.