This paper is devoted to the development of a neural network structure which implements the line,power flow state estimator algorithm for solving the set of nonlinear equations of power system automatic generation control analysis. The principal context is that of on-line network analysis in energy management systems with particular reference to the automatic generation control function. The author shows that the complete line power flow state estimator formulation maps into an array of two-layer neural networks. The development starts from a formulation for solving as a minimization the equation system to which the state estimator sequence leads at the each iteration. A neural network structure is given which implements the steepest descent method for minimizing the objective function. It is shown that some of the input values of neural networks are formed from power flow measurements as on-line. A principal feature of the extensive parallel processing capability of the architecture developed is that the computing time of state estimator analysis is independent of the number of nodes in a power network for which analysis is carried out. (C) 2003 Elsevier B.V. All rights reserved.