电磁场数值分析18.ppt

* * -18- Introduction to the finite element method (FEM) (II) ---- the variational method Xie, Yongjun （谢拥军） Department of Microwave Telecommunication Engineering Introduction The variational method is a different way of applying the the finite element method. The method has been employed extensively in electromagnetic formulations in the past, but recently Galerkin’s method has most often been employed. With the variational method, instead of solving directly the governing equations, we minimize the “enengy functional” (or, more often, an “energy related functional” ) corresponding to the governing equations. Needless to say, this can only be done if the energy functional is known. Energy functional (I) Using a very simple and not entirely rigorous analogy, the mathematical idea behind this method cab be described by the following example. Suppose that the equation “ ax+b=0” must be solved; instead of solving it directly, we can minimize the function (18.1) Assuming that this function has a minimum, this minimum corresponds to the point where df/dx=0, or (18.2) which is the equation we wish to solve. Energy functional (II) In physical terms, the state for which the energy functional is minimum, corresponds to an equilibrium state of any physical system possessing potential energy. In the example chosen here, which is described by the equation , the energy functional in a dielectric medium is (18.3) The discretized functional for the 1D problem, with N elements (or segments), is (18.4) To minimize this function, we need to find the minimum with respect to each unknown potential value in the solution domain FEM (I) (18.5) which results in k equations. Therefore, we need to evaluate for all elements. For a generic element, the first integral of Eq. (18.4) is (18.6) With the adopted approximation for V(x), the electric field intensity is constant in the element and we can write (18.7) FEM (II) From Eq.(16.28), (18.8) and (18.9) For