Separating low-frequency electromechanical modes from spurious modes in ambient data analysis by means of modal stabilization diagrams

The separation of low-frequency electromechanical modes from spurious modes is a challenging problem in ambient data analysis performed by identification and signal processing techniques. In order to accurately extract all relevant dynamic properties from power system ambient measurement data, a filtering pre-processing step and high-order models are usually adopted during the identification process for an effective handling of noise, which may lead to the appearance of the so-called spurious modes into the identified system model. These spurious modes have no physical relevance, but they may present similar characteristics to the low-frequency electromechanical modes, affecting the accurate evaluation of the latter ones. Hence, this paper proposes a post-processing technique based on modal stabilization diagrams to distinguish low-frequency electromechanical modes and spurious modes from ambient data analysis. This proposed post-processing technique is applied in simulated and real measured ambient data with the purpose of estimating the dynamic properties of inter-area modes in large interconnected generation and transmission power systems (IPSs), as well as, global and local modes in active distribution networks (ADNs) with distributed generation (DG) based on synchronous generators.