ABSTRACT
The first 3D methods were based on nodal elements but, in the late 1980s, it was observed that the precision of the formulations associated to the vector equations was, in high-frequency cases, poor and led to” spurious modes creating, obviously, difficulties for the engineering projects in that area. The main reason for such difficulties seems to come from the fact that nodal elements are appropriate to describe scalar quantities. In J. P. Webb presents an interesting discussion about possible difficulties originated by describing vector fields by nodal elements. It also considers the “awkward and unnatural” way to impose boundary conditions when using vectors defined by nodal elements. During the late 1970s and the 1980s, A. Bossavit and J. C. Verite proposed the use of “edge elements” to solve magnetic problems. Using a vector formulation with nodal elements required describing the vector potential components and therefore the number of unknowns is approximately three times the number of nodes.
