Analysis and Comparison of Reduced-order Model and Modeling Method for Grid-connected Inverter

The small-signal model of grid-connected inverter is an important tool for judging system stability and parameter design. The complete state equation of the grid-connected inverter will be accompanied by more redundant information. It is convenient to tune and debug system parameters by adopting appropriate reduced-order model. However, some of the existing typical reduced-order models have narrow applicability, low accuracy under specific conditions or loss of generality with the change of parameters. Firstly, this paper develops a general power-loop model based on the concepts of equivalent inertia constant and equivalent damping coefficient. Through single-side Fourier analysis, it is pointed out that the traditional third-order model is only suitable for the analysis of the systems with large inertia and high damping. Through the analysis of output impedance of grid-connected inverter, it is pointed out that the accuracy of the electromagnetic fifth-order model is sensitive to the control parameters. In order to solve the accuracy problem of reduced-order model, a model order-reduction method based on time constant is proposed by using the characteristics of natural separation of the time scale of inverters. The proposed method is simple and effective. And a simple reduced-order model of the inverter is derived. The obtained model has high accuracy and wide application. Finally, the correctness of the theoretical analysis and the effectiveness of the order-reduction method and the reduced-order model are verified by comparative experiments. © 2020 Automation of Electric Power Systems Press.