Neural network aided design for metering system of power system state estimation

This paper presents an artificial neural network (ANN) aided design approach for the determination of the measurement scheme, employed for on-line power system state estimation (PSSE). According to predetermined values of both the state estimator performance objectives (accuracy, running time), the proposed technique provides decisions regarding the adequate number of measurements, meter type, and meter locations for the a specific network configuration. An efficient modeling to the output layer of the proposed ANN is developed in this paper. The network is trained using off-line prepared I/O data pairs and the back-propagation learning algorithm. Numerical experimentation is conducted on a simple 6-bus system and simulation results are assessed.