Non-Injected broadband impedance estimation based on injected impedance measurement

• A two-stage operational states classification method based on multi-domain features is proposed. • An operational states matching method based on Pearson correlation is proposed. • Non-injected broadband impedance estimation is proposed based on injected impedance measurement. • The accuracy of the harmonic impedance fitting are more than 90% for simulations. As power electronic devices become increasingly connected to power grids, power systems face issues, such as stability and harmonic resonance; broadband harmonic impedance is a key piece of information that helps analyze these issues. This study proposes a non-injected broadband impedance estimation method based on injected method impedance measurement. First, a two-stage operational states classification method was conceived based on multi-domain features. Second, an operational states matching method was proposed to associate the concerned-time power quality monitoring data with the pre-classified states, based on Pearson correlation. Third, a regularized optimization model was developed to estimate the broadband impedance curve. Finally, a grid-connected inverter system under different control strategies was constructed as a case study. The simulations demonstrated that the accuracy of the proposed methodology was more than 90% for all the operational states defined in the case study, thus it obviates the need for perturbation injection as in traditional methods to enable real-time broadband impedance measurement.