Estimation of Modal Parameters for Inter-Area Oscillations Analysis by a Machine Learning Approach with Offline Training

• DOI: 10.3390/en13236410
• Publication Date: 2020-12-04T16:59:00Z
• Authors: Carlo Olivieri; Francesco de Paulis; Antonio Orlandi; Cosimo Pisani; Giorgio Giannuzzi; Roberto Salvati; Roberto Zaottini
• Keywords: Mode (computer interface); Identification

An accurate monitoring of power system behavior is a hot-topic for modern grid operation. Low-frequency oscillations (LFO), such as inter-area electromechanical oscillations, are detrimental phenomena impairing the development itself and also integration renewable sources. interesting countermeasure to prevent occurrence continuously identify their characteristic mode parameters, possibly realizing an online system. In this paper attempt develop modal parameters identification done using machine learning techniques. approach based on proper artificial neural network exploiting frequency measurements coming from actual PMU devices presented. The specifically developed offline training stage fully detailed. output results dynamic decomposition method considered reference in order validate approach. Some presented effectiveness proposed data recordings real events. main key points affecting performance technique discussed by means validation scenarios. This contribution first step more extended project whose final aim networks (ANN) architecture able predict (in given time span) terms LFO classify contingencies/disturbances that has memory passed samples.