Finite elements versus finite volumes. Is there really a choice?

Published in Non linear computational mechanics. State of the art. P. Wriggers and R. Wagner (eds.), Springer Verlag, 1991

Abstract

The finite volume method appears to be a particular case of finite elements with a non Galerkin weighting. It is course less accurate for self adjoint problems but has some computationally useful features for first order equations involving only surface integrals. For certain problems this is a substational economy and leads to computationally useful approximations.