Distribution Network Topology Identification Using Asynchronous Transformer Monitoring Data

The topology of a distribution network with a high penetration of fluctuating renewable energy is usually adjusted frequently for the reliable and economical operation. This paper proposes an automatic topology identification method based on measurements from widely distributed transformer terminal units (TTUs) to align the topology model in the management system with the actual topology of the distribution network. The proposed method formulates the topology identification problem as an $L2$ -norm minimization problem. The challenge lies in the absence of phase angle measurements and the sampling errors caused by the non-synchronous sampling time of different TTUs. Adding variables into the minimization problem for the unavailable phase angles introduces nonlinearity and non-convexity, and the non-synchronous sampling time can cause significant sampling errors when measuring the fluctuating power injections. The non-convexity is addressed by proposing a novel linearization and convex relaxation approach to convert the formulated minimization problem into a convex mixed integer quadratically constrained program (MIQCP) problem. The significant sampling error is addressed by generating pseudo measurements from Lagrange interpolation to be used as alternatives to the actual power measurements having significant errors. Case studies conducted in test systems demonstrate the effectiveness and efficiency of the proposed method.