Distribution Grid Topology Calibration Based on a Data-Driven Approach

Abstract

With the introduction of advanced metering infrastructure and smart meters at the customers' premises, an unprecedented amount of data becomes available to improve and validate distribution grid models. Therefore, assuming there are distribution grid topological errors, data-driven methods can utilize smart meter data to remedy the real-time topology in which the grid is currently operated and correct the topology errors stored in the database of the distribution management system. In this work, a hybrid methodology, encompassing graph theory and data-driven approaches based on statistical inference, is used to identify the errors in the underlying operational grid topology models. The methodology relies on voltage magnitude timeseries data, which are easily obtained from smart meters.