Digitalization and decarbonization of energy systems: a geospatial and edge computing approach

Abstract

The modernization of power systems into smart grids (SGs) is fundamental to achieving a de carbonized economy through the integration of distributed energy resources (DERs), such as solar PV, electric vehicles and emerging energy carriers such as green hydrogen. This transition requires a paradigm shift from traditional reactive planning toward a proactive framework based on the principles of predict, prevent, and optimize. Advanced data-driven techniques and de centralized decision-making provide a promising framework for the optimization of decentral ized resources, provided that the operational constraints of the physical grid are respected. Ac curate grid state estimation requires a reliable digital representation of the physical network. To support this requirement, synthetic distribution networks (SDNs) are constructed using street map layers and geospatial load data through a GIS-based graph methodology. By employing two-dimensional KD-trees along with a graph reduction process, this approach achieves a 2.6 fold reduction in node count compared to traditional geometric approximations. The effective management of distributed resources further depends on a robust communication and efficient data-processing infrastructure. The increasing volume of internet of things (IoT) data, combined with the strict real-time requirements of grid automation, motivates the developing of a multi tier edge computing architecture. This decentralized infrastructure distributes computational tasks across edge, fog, and cloud layers using the lightweight MQTT communication protocol. By strategically placing fog brokers through spatial-electrical clustering (K-means and DBSCAN), transmission latency is reduced by 81% relative to centralized cloud models. Each of these fronts contribute to moving into the scalable infrastructure necessary for sustainable and intelligent grid operations.