Research On Some Issues Of Power System State Estimation

Power system state estimation is the fundation and core part of Energy Management System. The accurate and efficient performing of state estimation software guarantees the safety, stability and economy of power system. Power system state estimation has become a research focus since 1970s. For now, great progress has been made in this area, both in theory and engineering. However, there still exist some problems to be solved. The traditional state estimation methods are limited by solving the nonlinear and non-convex programming problems. This paper solve the prominent problems from the prospect of exact linearization of measurement equations, which optimizes the accuracy, efficiency and global optimum performance. The work of this thesis is listed as follow.1) To ensure the estimation accuracy, power system state estimation has to identify gross measurements while conducting the estimation of state variables. The existing robust state estimation approaches usually adopt gradient based methods to solve the nonlinear and non-convex programming problems, therefore cannot guarantee to obtain the global optimum mathematically, and may encounter convergence problems. Based on the exact linear measurement equations, this thesis proposes a mixed integer linear programming formulation for robust state estimation with the following three main advantages. First, it is with strong anti-gross measurements ability. Second, the global optimum can be guaranteed theoretically. Third, no nonlinear iterations are needed, thus there is no convergence issue. Simulation results demonstrate the effectiveness of the proposed approach.2) When the state of power system change adruptly, the traditional method becomes difficult to track the state accurately. What’s more, its tracking efficiency is also high enough when applied to large scale power system. After thoroughly analyzing reasons of the drawback, a novel linear state tracking model is derived by applying the exact linear measurement equations to tracking model. Tests on IEEE 14-node benchmark systems illustrate the effectiveness and efficiency of the proposed approach.3) Only if the system parameters are correct, can state tracking software perform perfectly. Based on exsiting motheds, a novel joint tracking of state and parameter method is derived. Tests carried out on IEEE system illustrate that the method can track the state variables accurately, free from the affect of parameter error.