Research Of Transient Stability Constrained Emergency Control Based On Dynamic Optimization

With the development of modern power system, transient stability has increasingly been important. During the normal operation of power system, emergency fault may cause serious damage, so it is necessary to take some strategies to stabilize power system. This paper researches the transient stability constrained emergency control (TSCEC) from the viewpoint of dynamic optimization. The theory of optimal control is applied to the models of dynamic power system. Two methods of dynamic optimization are studied: pseudo-spectral and multiple-shooting. They are all used to implement the emergency control strategy and yield effective solution.The main content of this paper are as follows:(1) TSCEC is modeled as an optimal control problem with differential-algebraic equation (DAE) constraints. Pseudo-spectral method is introduced to this model and dynamic equations are discretized globally to simultaneously solve the transcribed non-linear optimization problem (NLP). This algorithm is also compared with traditional discretization schemes like implicit trapezoids and orthogonal collocation on finite elements to analyze its robustness and accuracy. Fine grained parallelism is utilized to speed up the solving process.(2) Partially discretization method is considered, which is different from the global scheme mentioned above. The basic idea of multiple-shooting is discussed. Based on forward and adjoint sensitivity analysis, the forward and adjoint model of TSCEC is induced to obtain gradient information. A composite sensitivity analysis method is proposed to update inexact Jacobian and Hessian and computation complexity is reduced. The block structure of matrices in the model is used in the parallelism of dynamic simulation and reduced space interior point method.