Rapid Power System Transient Stability Limit Search Using Signal Energy and Neural Networks

In very recent work, researchers showed that the signal energy of the system response is indeed such a useful physical quantity, and that there is a sure relationship between this physical quantity and the system transfer limits. However, this relationship is still unable to be modeled. Here, artificial neural networks are the best to be used in such cases where input-output mappings are unknown or difficult to be modeled. The basic strategy for this paper is to use artificial neural networks for mapping power system signal energies into the power system transfer limits; i.e. to extract the relationship between power system transfer limits and the system signal energy. Primary results show that the neural network, after being trained, is able to accurately predict the power system transfer limit by looking to the signal energy measured at different buses in the system

I. Introduction

One of the most challenging problems in real-time operation of power systems is the assessment of transient stability. Its importance has increased due to the reduction of operational safety margins caused by the increasing load demand in power systems without accompanying investments in generation and transmission. This has affected the dynamic security analysis requiring more reliable and faster tools since analytical techniques and time domain simulations are time consuming and do not allow to take preventive or corrective actions in due time.