Identifying the Relationship Between a Transformer and Its Downstream Premises

Academic Program: Electrical Engineering

Purpose/Essence of the Project

The project focuses on solving the problem of identifying customer-to-transformer assignments in a power grid in the absence of direct topological information (ground truth), using consumption data alone. To this end, we developed and compared static and dynamic approaches while examining their robustness under conditions of uncertainty, such as missing data, noise, and anomalies.

The algorithm serves as a key component in detecting bypasses at the distribution transformer level, even when not all downstream premises connected to the transformer are covered by smart meters.

Describe the Product/Project

Static and dynamic algorithms (NNLS and UKF) that estimate and classify the network topology using information collected from smart meters distributed throughout the electricity network polygon and from the transformer.

How Was the Idea Born?

The research proposal builds on previous studies in the field of smart meters. Additionally, there is a need to identify the relationship between a transformer and its downstream premises without relying on an event log, and no solution to this problem currently exists worldwide.

Why Did You Choose This Project?

We chose the project out of a desire for a challenge, curiosity, and the wish to address a need that has not yet been solved and to be pioneers in the field, with the possibility of developing the work into an extended article for a master’s thesis.

Is the Project Connected with a Personal Story of Yours or of Someone You Know?

No.

What Need Does It Address and for What Population?

The project addresses the problem of identifying customer-to-transformer assignments in the absence of topological information.

The project is intended for entities operating in the power grid sector, particularly distribution and infrastructure management companies that need to identify customer assignments and analyze electricity consumption anomalies in the absence of complete topological information. The system can serve as a decision-support tool for detecting faults, electricity theft, and inconsistencies in the grid.