Avaliação da proteção de perda de excitação convencional e baseada em RNA de geradores síncronos na presença de SVC

Abstract

As the electrical power system (SEP) expands, it becomes necessary to take precautions regarding stability and power transmission issues, as they become increasingly complex. In this scenario, the use of Flexible Alternating Current Transmission System (FACTS) devices improves the stability and power flow optimization of the network. However, the integration of FACTS into transmission lines (TLs) can modify the voltage and current amplitude and angle values of the system where they are inserted, thus altering the network has characteristics and consequently, the trajectory of the measured impedance. Such effects can then be reflected in the operation times and coverage ranges of relay protection zones that relays on impedance measurement, such as the loss of excitation protection for synchronous machines connected to these TLs. In this context, this study presents computational simulations and laboratory tests to analyze the response of protection systems for synchronous machines connected to TLs with Static Var Compensators (SVC) FACTS devices. Various events with and without the presence of SVCs are simulated and stored in a database for offline closed loop testing with a generator protection relay, evaluating the response time of traditional loss of excitation protection characteristics. These characteristics are compared with a method based on a Multilayer Perceptron Artificial Neural Network (ANN) implemented to enhance the protection response in these scenarios.