Rotational Power Loss in Commercial Silicon-Iron Laminations

Finite difference and finite element methods are investigated as applied to Laplace and Poisson linear equations in two dimensions. The comparative analysis concentrates on first-order square, rectangular, triangular and polar algorithms which are characteristic to each method. The results are found to be independent of the method used to establish the matrix equations but dependent upon the choice of grid systems employed to discretize the problem. It is found that for special cases the solutions are also independent of the type and structure of the grid systems used.

Comparison of First-Order Finite Difference and Finite Element Algorithms for the Analysis of Magnetic Fields. Part I: Theoretical Analysis

The finite difference method, as developed at the University of Colorado [1,2,3,4] for the numerical solutions of the nonlinear counterparts of Laplace's, Poisson's, and the diffusion partial differential equations, yields calculated solutions that agree well with experimental results on several turboalternators. Computed no-load characteristics and local flux distributions are compared with measurements. In addition, it is shown that the Carter coefficients depend not only on slot and tooth dimensions but also on the curvature of stator and rotor. Subsequently, calculated excitation currents, synchronous reactances, and some eddy current distributions are compared with measurements.

Comparison of Iterative Solutions of the Finite Difference Method with Measurements as Applied to Poisson's and the Diffusion Equations

A program for the generation of a polycentric grid which fosters the application of numerical methods of computing magnetic fields in electrical power engineering is described. To achieve a still faster convergence of the finite difference method developed at the University of Colorado [1,2,3], a new complementary dynamic acceleration with an iteration degree larger than one is introduced. Estimates of truncation errors based on first-and second-order approximations of the potentials show that the error approaches zero asymptotically with increasing number of grid points.

Computer Generated Polucentric Grid Design and a Novel Dynamic Acceleration of Convergence for the Iterative Solution of Magnetic Fields Based on the Finite Difference Method

In the subtransient analysis of salient pole alternators the reactances X11d and X11q are generally assumed to be constants. In saturated alternators these reactances are in reality functions of the pre-transient load and saturationix This gaper is an extension of former papers by Fuchs, et it dealing with steady state, and transient reactances. The methods used are the same as in [1] and [2] with the addition of damper windings effects. Notations not explained here are those of Concordia4 and of Fuchs3. The analysis is applied to a G. E. alternator with the following name-plate data: 82.5 MVA, 16.5 kV, wye-connected, 40 poles and 60 Hz, described in [1].

Nonlinear Salient Pole Alternator Subtrannsient Reactances and Damper Winding Currents

The Colorado method for the solution of the non-linear form of Laplace's, Poisson's, and the diffusion partial differential equations is explained. Various boundary conditions can be satisfied. The transformation of the partial differential equation into a large set of finite difference equations is given. The discretization is based on a grid system consisting of two sets of orthogonal grid lines. The resulting meshes are nonuniform. Successive line overrelaxation method is used for the solution of the nonlinear equations in two steps. For the improvement of convergence, two methods of acceleration of convergence are described.

E. Erdelyi

Papers

Fields in electrical devices containing soft nonlinear magnetic materials