Optimal Meter Placement in Distribution Feeders Using Branch-Current based Three-Phase State Estimation: A Quest for Observability Enhancement

The electricity distribution networks are undergoing rapid changes. Integration of renewable energy sources and new loads demands to have a higher level of observability require more advanced protection and control schemes. Hence, the system operator should have online measurements of nodal voltages and branch currents to deal with these issues. Measuring all variables of real systems with a large number of nodes and lines as well as high cost of measurement devices are two major reasons for solving a three-phase state estimation with only a few meters. Therefore, optimal allocation of measurement devices is an effective solution to minimize the total investment cost while improving the quality of voltage and current estimates. This paper presents a novel approach for finding the best locations of the voltage meters and reducing the estimation error. A branch-current-based three-phase state estimator is coded to calculate the voltage profile according to the obtained branch currents and limit the relative errors between the calculated and actual nodal voltages. Moreover, a mesh adaptive direct search-based mixed-integer non-linear optimization algorithm is proposed to find the optimal locations of the voltage measurements and achieve the voltage magnitude and phase estimation errors of less than 1% and 5%, respectively. The proposed approach is suitably validated with simulations on a 25-node unbalanced distribution feeder. It is able to find the best locations of the voltage meters in calculation time less than 16 min. Moreover, the state estimation algorithm converges to more accurate nodal voltages and branch currents in less than 1 sec., especially in the case of optimally allocated meters.