State estimation (SE) is a crucial tool for power system state monitoring since the control center requires a process to deal with a large number of imprecise measurements. Several SE methods have been applied and developed for the electric power system in the transmission level in the past several decades. Meanwhile, SE for the distribution level remained in the background for a long time since the distribution networks were mainly radial with uni-directional power flows, making classical monitoring and control functions sufficient. Recently, due to the liberalization of the energy market, growing penetration of distributed generation, mainly renewable energy sources, and distributed energy resources such as electric vehicles, the distribution system has been gradually changing from passive into active grids. This requires more sophisticated monitoring and control of the distribution network via a distribution management system (DMS) to ensure optimal integration and maximize the grid hosting capacity. Since one of the key functions of DMS for real-time operation is the SE procedure, this calls for the following: (i) development of SE techniques for the distribution level, so-called distribution system state estimation (DSSE); (ii) more deployment of time-synchronized devices like phasor measurement units (PMU) that can directly measure and acquire accurate and time-aligned phasors with typical refresh rates up to 20-60 times per second.

Two types of DSSE algorithms were developed and implemented on the real-life 50 kV ring distribution grid composed of PMU devices by using the real-time simulation platform. The first is a static approach. The problem is formulated as a WLS problem to be solved based on the iterative Newton method, known as static state estimation (SSE). The second is a dynamic approach, which is more advanced, known as the forecasting-aided state estimation (FASE). It is one of the particular applications of the dynamic state estimation (DSE) concept based on the quasi-steady-state operating conditions. The dynamic formulation is solved using the extended Kalman filter (EKF) technique. This research aims to implement distribution system state estimation (DSSE) algorithms coupled with the auxiliary function, the so-called anomaly detection discrimination and identification (ADDI), into the distribution network. Both normal and abnormal operation scenarios of the power system are simulated to validate the algorithms.

The results reveal that the FASE is superior to the SSE algorithm in terms of estimation accuracy and computational time under normal operating conditions. However, under abnormal conditions, if there is no ADDI module, the performances of both algorithms are degraded significantly due to erroneous measurements. The FASE algorithm loses the system states' trajectory when sudden load change occurs. These issues point out the necessity of using the ADDI module against possible disturbances in real-life networks. In the end, the results show that the proposed FASE algorithm coupled with the ADDI module can accurately estimate the states under both normal and abnormal operations. One significant contribution is that the proposed algorithm can perform adequately fast so that it can process every high-speed measurement from PMU devices.