Large-signal black-box behavioral modeling of grid-supporting power converters in AC microgrids

Abstract

Control of power converters in AC microgrids is well understood and established in the literature and in the industry, allowing the high penetration of distributed generation, electrical energy storage systems and controllable loads. These building blocks connected to the grid through power electronic converters open new opportunities to the expansion of microgrids in electrical power systems. A behavioral large-signal nonlinear polytopic model for grid-supporting converters operating as a current source is proposed in this paper. This model is able to represent the nonlinear behavior of the power converter and it is also very well suited for system-level modeling and simulation. The proposed modeling strategy can be easily extended to power converters in any other of the possible operating modes.

Blackbox Large-Signal Modeling of Grid-Connected DC-AC Electronic Power Converters

“…Furthermore, a detailed description of the tests needed to perform the identification of three-phase Voltage Source Inverters (VSI) is presented in [34]. Finally, a preliminary work to consider the large-signal behavior of grid-connected DC-AC converters with droop control is presented in [35] at simulation level.…”

“…As mentioned in Section 2, this methodology is based on a weighted combination of small-signal models identified around different operating points. This approach has been successfully applied in DC systems [17,19,20,31], and extended to AC EPCs in [35] at simulation level. Therefore, the present work is the first evidence of a polytopic model applied to represent an actual three-phase DC-AC EPC.…”

DQ-Frame Admittance Estimation of Three-Phase Converters with a Wide Range of Operating Points

“…However, the admittance estimation at different operating points is challenging due to the nonlinear relationship between the converter admittance model and operating points. To address this challenge, a multi-operating-point small-signal modelling method for three-phase converters is reported in [7]- [10]. In the method, a set of dq-frame admittance models at a series of operating points are identified based on the timedomain measurement data.…”

DQ-Frame Admittance Estimation of Three-Phase Converters with a Wide Range of Operating Points

“…However, there is a lack of clear guidance for the ANN training process, i.e., the initialization of ANN structure, the selection for the number of hidden layers, and the number of neurons in each layer, which significantly influences the estimation accuracy and is challenging for the practical application. Further, the multioperating-point admittance model identified from either the time-domain data [7]- [10] or the frequency-domain measurement results [13] requires the utilization of nonlinear weight functions to estimate the unmeasured admittance models through the different weighted combinations of measured admittance models, which makes the approach tedious and complicated.…”