Chapter

Optimisation des reseaux electriques de grande taille

DOI: 10.1007/BFb0120065

ABSTRACT Cet article décrit une méthode de programmation quadratique successive permettant de résoudre le problème de la répartition
optimale des puissances dans un réseau électrique. Cette méthode converge globalement et superlinéairement. Des techniques
ont été mises en oeuvre pour traiter des réseaux de grande taille. Le programme quadratique de chaque itération est résolu
en utilisant un algorithme de gradient réduit avec factorisation LU des matrices de base et conjugaison des directions réduites. La méthode a été appliquée au problème particulier de l’optimisation
du plan de tension à puissances actives fixées. Le réseau étudié possède 838 noeuds et 1330 lignes ou transformateurs.

1 Follower
 · 
280 Views
  • [Show abstract] [Hide abstract]
    ABSTRACT: We study the convergence of a sequential quadratic programming algorithm for the nonlinear programming problem. The Hessian of the quadratic program is the sum of an approximation of the Lagrangian and of a multiple of the identity that allows us to penalize the displacement. Assuming only that the direction is a stationary point of the current quadratic program we study the local convergence properties without strict complementarity. In particular, we use a very weak condition on the approximation of the Hessian to the Lagrangian. We obtain some global and superlinearly convergent algorithm under weak hypotheses. As a particular case we formulate an extension of Newton’s method that is quadratically convergent to a point satisfying a strong sufficient second-order condition.
    SIAM Journal on Optimization 11/1995; 5(4). DOI:10.1137/0805038 · 2.11 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper is the second of a series dedicated to the mathematical study of very high voltage power networks, based on an asymptotic analysis, in which the square root of one of the small parameters is the inverse of the reference voltage of the network. We call this scheme the very high voltage approximation. We show that the active and reactive parts of the limit problem are decoupled. We give an analytic expansion for the solution of the problem, and prove the convergence of two decoupled numerical algorithm.
    SIAM Journal on Applied Mathematics 10/1998; 58(5). DOI:10.1137/S0036139995289325 · 1.41 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: We consider the problem of optimal design of hybrid car engines which combine thermal and electric power. The optimal configuration of the different motors composing the hybrid system involves the choice of certain design parameters. For a given configuration, the goal is to minimize the fuel consumption along a trajectory. This is an optimal control problem with one state variable.The simultaneous optimization of design parameters and trajectories can be formulated as a bilevel optimization problem. The lower level computes the optimal control for a given architecture. The higher level seeks for the optimal design parameters by solving a nonconvex nonsmooth optimization problem with a bundle method.
    Optimization and Engineering 11/2004; 5(4):395-415. DOI:10.1023/B:OPTE.0000042032.47856.e5 · 0.96 Impact Factor