Research On The Model Order Reduction Methods For Electromagnetic Transient Simulation Of Active Distribution Networks

Electromagnetic transient simulation of active distribution networks is the foundation to understand the fast dynamic characteristics and to reveal the various related transient problems in the system. However, the simulation efficiency is significantly decreased with the widely integration of various distributed generators, energy storage devices, electric vehicles, and loads. In this dissertation, the model order reduction(MOR) methods for active distribution networks are studied. The research is conducted in several aspects including the MOR framework, the projection algorithms, the realization methods, and extension for nonlinear problems. The work in this dissertation is summarized as follows:1) The MOR framework for large-scale active distribution networks is studied with the employment of the Krylov subspace based MOR theory. The principle of model division is proposed for the application of linear MOR methods. The MOR procedure is explained based on the basic Krylov subspace methods for active distribution networks, including the construction of projection basis, the improvement of numerical characteristics, and the realization of space projection.2) Under the aforementioned MOR framework, a new MOR algorithm, which can preserve the stability and passivity of the original model, is studied for distribution networks. A novel state-space based modeling method for linear distribution network is proposed. Compared with the traditional model, the novel model is smaller in scale and easier for analysis, while it still holds the properties such as the preservation of original passivity and stability. In this way, the applicability of the proposed MOR algorithms is significantly improved.3) The realization of the proposed MOR methods in popular EMTP-type programs is studied. An automated MOR procedure based on error control is proposed for distribution networks. The reduced model is then applied in an EMTP-type program with the employment of the combined state-space nodal analysis framework. This realization method paves the way for practical engineering application of the proposed MOR techniques.4) The extension of the MOR methods based on linear system theory for nonlinear models of active distribution networks is explored. The interface for the integra-tion of linear networks and nonlinear devices is designed. On this basis, different MOR algorithms and simplified simulation methods are studied for the models of active distribution networks with different proportions of nonlinear elements.