Multivariable Regression Equivalent Model of Interconnected Active Distribution Networks Based on Boundary Measurement

Abstract

With the increasing complexity of the active distribution network (ADN) due to distributed generation (DG) integration, together with the electricity market evolution, the traditional ADN is divided into multiple areas to operate independently. Due to technical problems or business privacy, the internal network regional control center cannot grasp the changes of the external regional network in time. In order to accurately reflect the distribution network operation state, a multivariable regression equivalent model is proposed in this paper. Firstly, the external network is made equivalent to a multi-port Norton model. The multivariable linear regression model is then derived based on the equivalent distribution network, and the regression model variables are constructed using boundary node information collected by the measurement equipment. Finally, the maximum likelihood estimation (MLE) is used to estimate the parameters of the multivariable linear regression model. Furthermore, case studies demonstrate the effectiveness and robustness of the proposed method, and detailed information of external ADN is unnecessary, except for the boundary node information. The proposed method can also be applied for three-phase unbalanced ADN efficiently.