Fast and accurate technique for power system state estimation

Abstract

The paper presents a fast and accurate technique for power system state estimation. The proposed technique is based on the exact linearisation of the Cartesian co-ordinate formulation of nodal load flow equations. The complete Taylor series expansion is used to solve these equations while retaining the nonlinearity. In this technique, the Jacobian and Hessian matrices are kept constant, and hence need to be computed once only, at the beginning of the solution, and then stored in the computer memory using the sparsity storage procedure. Because, in this technique the active and reactive power problems are attempted simultaneously, not sequentially, there is no need to use decomposition into two subproblems, namely P–Θ and Q–V. The linear programming (LP) technique is used to solve the state estimation problem without any approximation for the load flow equations. For small changes in active and reactive power, the right-hand side of the LP is compensated only when the state estimation solution of the real component of the bus voltages is obtained with a minimum number of iterations compared with the fast decoupled state estimation (FDSE). For large band-data points, however, the imaginary components of the bus voltages must be computed, in addition to the real components. The proposed technique is more suitable than FDSE for solving ill-conditioned networks which have large bad-data points. An extensive study was carried out on a standard 23-bus test system to demonstrate the capability of the proposed technique for solving the state estimation problems.