PATTERN-ANALYSIS IN POWER-SYSTEM STATE ESTIMATION

In recent years, interest in the application of artificial intelligenc e technologies to power system operation, planning and design has grow n rapidly. The application of non-symbolic techniques, particularly Ar tificial Neural Networks (ANNs), is a new area of research in this fie ld. In this paper, intelligent systems for solving power system state estimation problems are investigated. A new framework for the solution of the topology determination, observability analysis and bad data pr ocessing tasks is proposed. Pattern analysis techniques have been deve loped to deal with noisy environments. An ANN for topology determinati on and a supervised learning algorithm for very large training sets, t he Optimal Estimate Training 2 (OET2), are introduced. OET2 overcomes the major shortcomings of the back-propagation learning rule and can a lso he very useful for other problems. Power system network decomposit ion techniques are used to decrease the computational burden of the to pology classifier training session. Tests using the IEEE 24- and 118-b us systems illustrate situations in which the existent tools for data processing fail.