Data quality challenges in existing distribution network datasets

TL;DR

This paper discusses the challenges of data quality in distribution network datasets, emphasizing the need for improved data-driven methods and aligning academic research with industrial realities to support energy transition efforts.

Contribution

It highlights discrepancies between academic assumptions and industrial data realities, guiding future research towards more practical and applicable solutions for distribution network data quality.

Findings

Existing datasets often contain errors affecting simulations

Smart meter data offers new opportunities for data improvement

Academic research needs to better reflect industrial data conditions

Abstract

Existing digital distribution network models, like those in the databases of network utilities, are known to contain erroneous or untrustworthy information. This can compromise the effectiveness of physics-based engineering simulations and technologies, in particular those that are needed to deliver the energy transition. The large-scale rollout of smart meters presents new opportunities for data-driven system identification in distribution networks, enabling the improvement of existing data sets. Despite the increasing academic attention to system identification for distribution networks, researchers often make troublesome assumptions on what data is available and/or trustworthy. In this paper, we highlight some differences between academic efforts and first-hand industrial experiences, in order to steer the former towards more applicable research solutions.