Research On Network Analysis Algorithms Of Active Distribution Network

Active Distribution Network (ADN) is the distribution network which can control dispersed energy resources including Dispersed Generation (DG), flexible load and storage. This paper focuses on certain network analysis algorithms of Active Distribution Network. Topics including bad data identification, state estimation and load flow calculation are discussed. First, this paper proposes a new method for bad data identification based on clustering analysis. The proposed method can identify bad data without using the weights of measurements. Tests show that bad data can be identified successfully and quickly when branch measurements (active power, reactive power at either side of the branch and voltage magnitude at either node) have one single bad data. Then, this paper presents the nonlinear-programming-based state estimation model with four detailed models using different state variables. In this approach, constraints can be easily considered in nonlinear-programming-based state estimation method which makes the estimation more reasonable. Tests and comparative studies on both test systems and a real system show that the proposed method is applicable for real engineering. Last, this paper proposes a novel three-phase power flow algorithm based on loop-current-method for meshed active distribution networks. Models of various dispersed generation including induction machines are also proposed. Tests show that the proposed method can deal with meshed network with DGs.