Dynamic power system equivalence considering distributed energy resources using Prony analysis

Abstract

SUMMARYPower system dynamic equivalence, which has been developed to overcome the computational issues, is faced with new challenges by emerging distributed energy resources (DERs). The switching characteristic between DERs operational modes beside non-dispatchable inherent, that is, generation uncertainties, should be take into account in power system dynamic equivalencing. This paper proposes a switching-based dynamic equivalence modeling for power system considering DERs. Prony analysis is employed to fit a reduced-order model to a high-order system in frequency domain. The developed modelling strategy consists of high and low frequency equivalences. When the external system reconnects to the main network through the breaker, high-frequency equivalence is replaced with external system. This period contains 4 cycles for communication delay beside 5 cycles governor time constant, which give rises to a six-order equivalence model. Afterwards, a four-order low-frequency equivalence model is fitted to the external system in the connected grid mode. Accuracy of the fitted dynamic equivalences is investigated by calculating the mean square error between the measured signal and the Prony approximated one. A synchronous-based DER, a doubly-fed induction generators and a fuel cell DER are added to the CIGRE benchmark in order to analyse the proposed dynamic equivalence modelling. Simulation results demonstrate that the proposed equivalence modelling could exhibit the same results as the real external system in the frequency domain. Simulation results also reveal that increasing penetration of DERs affect the dynamic equivalence in gridconnected mode by degrading the higher frequencies while it has no impact on transient period equivalence.

Dynamic equivalencing of an active distribution network for large‐scale power system frequency stability studies

“…Therefore, the modes are initially reduced based on the non‐coherent generators aggregation and after that, are taken into account. Recently, the present authors have numerically dealt with such representation in [17].…”

“…Eight DG units, including three synchronous generators with rated power between 2–5MVA and five inverter interfaced units with rated power between 1–3MVA, are considered in each MG topology. The synchronous generators are equipped with excitation and governor control systems as represented in [16, 17]. The inverter based sources utilise frequency and voltage droop characteristics, as represented in [18], to provide control on the output active and reactive powers, that is, PQ control.…”

Comprehensive Small-Signal Modeling and Prony Analysis-Based Validation of Synchronous Interconnected Microgrids

“…the dynamic modes of the power system frequency. Fourier transform (Bevrani et al, 2014), digital Taylor-Fourier transform (de la O Serna et al, 2016), eigensystem realization algorithm (TP462, 2012) and Prony analysis (TP462, 2012; Sauer et al, 2017;Zhao and Loparo, 2017;Golpîra et al, 2015;Føyen et al, 2018), are the major signal estimation techniques, which can be used to compare eigenvalues or state variables of the model with the corresponding measured waveforms.…”

“…Fourier analysis, the Prony method has the benefit of being able to estimate free from a fundamental frequency and their harmonics. The Prony analysis is used as a system identification method in many areas such as power systems (de la O Serna et al, 2016;Føyen et al, 2018;Golpîra et al, 2015). In Føyen et al (2018), the Prony method is applied to identify the power system oscillations in smart grids.…”

“…In Bracale et al (2007), an adaptive Prony method is presented to estimate the harmonics and inter-harmonics of measured current and voltage. In another application, authors employ Prony analysis to fit a reduced-order dynamic equivalence model to a high-order power system in frequency domain (Golpîra et al, 2015). In the literature, a Prony analysis-based validation method is reported for a small-signal FACTS model (Jiang et al, 2006).…”