Computing the closest small-signal security boundary in the control parameter space for large scale power systems

Abstract A constrained optimization algorithm is proposed for the determination of the minimum change in the parameter values of a set of controllers that locates a complex pair of eigenvalues of a linearized power system model at a user-defined small-signal security boundary. The intended practical use is to assess the combined robustness of the system controllers in maintaining adequately damped power system oscillations. The oscillation damping security margins, given an operating point and a set of fixed damping controller parameters, is the Euclidean norm of the relative parameter variation vector, also referred in the literature as the minimum distance in the control parameter space. The computational algorithm to find the Closest Security Boundary in the Control Parameter Space, CSBCPS, is based on the rigorous mathematical implementation of the non-linear programming method to the problem, including constraints on the parameter ranges. The resulting equations are solved by the Newton method. Numerical results for a large practical power system dynamic model are detailed described to illustrate the use of the proposed CSBCPS algorithm in small-signal applications involving multiple thyristor controlled series compensators and power system stabilizers.