Uncertainty quantification for branch-current state estimation in power distribution systems

The growing number of prosumers in the distribution grid pushes the system to the edge of its operational capacities. This makes it essential to assess the current state of the system. However, the availability of grid measurements is often times not sufficient. To achieve observability, inaccurate pseudo measurements can be included in the state estimation. The obtained results are subject to high uncertainty. This paper proposes an algorithm for quantifying the uncertainty of the branch current and node voltage estimate in a branch current-based weighted least squares state estimation. The algorithm uses the same inputs as the state estimation, including the measurement vector, its standard deviation, and the grid parameters. The uncertainty quantification is twofold: first, the uncertainty of the currents is estimated using the diagonal of the gain matrix. Second, the voltage uncertainty is obtained using a novel error propagation calculation. The proposed algorithm is based on a modified version of the branch current state estimation algorithm.We evaluate the proposed algorithm in a variety of scenarios with different measurement availabilities and pseudo measurement uncertainty distributions. As a guiding example, we utilize a benchmark distribution system from the Simbench dataset. The algorithm demonstrates high accuracy when dealing with Gaussian distributed input uncertainty

Subjects

current state estimation | uncertainty quantification | low voltage grid | error propagation

DDC Class

600: Technology

Options

Uncertainty quantification for branch-current state estimation in power distribution systems