zbMATH — the first resource for mathematics

A Petrov-Galerkin mixed finite element method with exponential fitting. (English) Zbl 0840.65130
The author is interested in the semiconductor continuity equation on a rectangular region with Dirichlet boundary conditions. The numerical methods are tied to rectangular mesh elements. The discretization is constructed by first expressing the problem using a mixed formulation, then approximating the coefficients as constants in each mesh element, and, finally introducing bilinear forms based on a special quadrature. This quadrature is designed so that the case of constant coefficients and exponential solution is exactly reproduced. A priori \(O(h)\) error estimates are demonstrated, an a posteriori error estimator (based on higher-order quadrature) is presented, and numerical results are presented. The numerical results show better convergence behavior than the theoretical estimates.
The author presents three error estimates, the first pair of estimates indicates that the error in the solution is \(O(h)\), for \(h\) small enough. The third relaxes the condition on \(h\) in the error estimate for the solution by employing estimates of the discrete adjoint problem.
The author observes that the special exponential weighting reduces to an upwind difference method in the convection-dominated limit. Further, use of this weighting results in an \(M\)-matrix to be inverted for the unknown, after static condensation is used to eliminate its gradient.

65Z05 Applications to the sciences
35Q60 PDEs in connection with optics and electromagnetic theory
78A55 Technical applications of optics and electromagnetic theory
65N30 Finite element, Rayleigh-Ritz and Galerkin methods for boundary value problems involving PDEs
65N15 Error bounds for boundary value problems involving PDEs
Full Text: DOI
[1] Scharfetter, IEEE Transactions on Electron Devices ED-16 pp 64– (1969)
[2] Analysis and Simulation of Semiconductor Devices, Springer-Verlag, Wien, New York, 1984.
[3] The Stationary Semiconductor Device Equations, Springer-Verlag, Wien, New York, 1986.
[4] Fichtner, SIAM J. Sci. Stat. Comp. 4 pp 416– (1983)
[5] , and , ”Mixed exponential fitting schemes for current continuity equations,” in Proceedings of the Sixth International NASECODE Conference, Ed., Press Ltd., Dublin, 1989. · Zbl 0800.65026
[6] A Numerical Study of Stiff Two-Point Boundary Problems, Mathematical Centre Tracts, 80, Mathematical Centre, Amsterdam, 1977.
[7] Douglas, Math. Comp. 44 pp 39– (1985)
[8] Sobolev Spaces, Academic Press, New York, 1975. · Zbl 0314.46030
[9] and , Finite Element Methods for Navier-Stokes Equations, Springer-Verlag, Berlin, 1986. · Zbl 0585.65077
[10] Polak, Int. J. Numer. Meth. Eng. 24 pp 763– (1987)
[11] and , ”Mixed and Hybrid Finite Element Methods,” RR 737, INRIA, Rocquencourt, October 1987.
[12] and , ”A mixed finite element for 2-nd order elliptic problems,” in Mathematical Aspects of the Finite Element Method, Lecture Notes in Mathematics, Springer, 606, 292 (1977).
[13] Nedelec, Numer. Math. 35 pp 315– (1980)
[14] Ciarlet, Arch. Rational Mech. Anal. 46 pp 177– (1972)
[15] Special Matrices and Their Applications in Numerical Mathematics, Martinus Nijhoff, Dordrecht, 1986.
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.