Optimal allocation of smart metering systems for enhanced distribution system state estimation

Smart meters are the backbone of advanced metering systems for accurate billing, load estimation and demand side response in smart distribution grids. By appropriately selecting their installation locations, it is feasible to achieve more precise distribution state estimation. This paper presents a semidefinite programming method to optimally allocate smart metering systems in distribution grids in order to minimize the state estimation error variances. The placement problem is formulated as an M-optimal experimental design model and is transformed into a binary semidefinite programming model, with binary decision variables, minimizing a linear objective function subject to linear matrix inequality constraints. Simulations using a 55-bus distribution grid are carried out in order to verify the optimality of the chosen meter locations in terms of minimizing bus voltage magnitude and angle variances.