A combined ANN/simulation tool for electric power system dynamic security assessment

Abstract

The increasing complexity of power systems due to their growth, operation close to technical limits, economical constraints and new regulations, calls for the inclusion of dynamic security assessment (DSA) into energy management systems where the overall system control is performed. Recent researches on DSA have shown a significant interest in the accuracy of tools based on hybrid methods, a combination of time domain simulation and transient energy functions. A synthetic review of a hybrid method based procedure, implemented by the authors using a suitably adapted simulation code, is first recalled. Successively the possible application of a tool based on artificial neural networks (ANN) for operating a contingency pre-screening by identifying stable and unstable situations is proposed. The ANN capability of identifying the stable and unstable clusters which may originate after a perturbation is also investigated. The proposed ANN structure has been trained and validated by simulating properly arranged scenarios from a test system constituted by a real size HV network of the Mediterranean area.