Reconciliation of Measured and Forecast Data for Topology Identification in Distribution Systems

Abstract

The status of switches in distribution systems might be changed for protection purposes or to achieve the operation objectives. However, the system topology should be known for the efficacious management and control of these systems. In most practical systems, due to the lack of enough measurements, the topology cannot be identified accurately and speedily. The forecast output can be transformed into the joint probability density function (<small>pdf</small>) of uncertain parameters, e.g., power demands. The measured data and this <small>pdf</small> are first conflated to update the joint <small>pdf</small> to best comply with the measurements. The topology identification problem, i.e., finding the probability of each topology given a set of measured values, is converted to a simpler form, i.e., finding the probability of observing these measured values in this topology. These probabilities are then calculated by tracing how the measurements affect the shape of the new consolidated <small>pdf</small> in each topology compared to the original <small>pdf</small>. The proposed technique employs as much data as available, is able to find the accurate probabilities, and is yet quite fast. The mathematical treatment of extracting such probabilities is presented and the proposed technique is validated in numerical studies in terms of accuracy, speed, and versatility.