How supercomputing helps protect the electricity grid

Since 2022, Terna has been actively involved in the National Research Centre for High Performance Computing, Big Data and Quantum Computing (ICSC), one of the five National Research Centres established as part of Italy’s National Recovery and Resilience Plan (PNRR) to strengthen scientific research and technological innovation in the country. The ICSC’s operational hub is the Bologna Technopole, one of Europe’s leading centres for research and innovation, which hosts Leonardo, a supercomputer considered among the most powerful in the world. Thanks to its enormous data-processing capacity, Leonardo supports scientific research and the development of advanced technologies across various fields, including climatology and energy, making it possible to carry out simulations and calculations that would require prohibitive amounts of time and resources using traditional computers.

In recent years, Terna’s contribution to the ICSC has taken shape through several projects submitted to, approved and funded by the centre, focusing on areas of strategic interest for both the company and the national electricity system. One of these is certainly the issue of resilience, and in particular the RETE project (Resilience of the Electric Transmission grid to Extreme events), launched to develop and test advanced technological solutions aimed at increasing the resilience and protecting the national electricity grid from extreme weather events and the impacts of climate change. In particular, researchers have focused on so-called “rapid landslides”, ground-instability phenomena that occur over very short periods of time, often following intense rainfall capable of saturating the soil and compromising its stability. These events can pose a significant threat to electrical infrastructure, putting the safety of power pylons and substations located in the most exposed areas at risk.

The researchers analysed the territory using a multiscale approach, moving from comprehensive overview of the landslide hazard of the overall view of the entire national territory to a detailed analysis of individual slopes considered at risk. Specifically, through the use of artificial intelligence and graph theory — a mathematical method used to represent and study the relationships between different elements of a network — they were able to identify areas where electrical infrastructure is most vulnerable. Among other findings, The approach was tested in Campania and Sicily and demonstrated the potential of the solution, providing a foundation for the development of an effective tool to support the planning of interventions aimed at enhancing the resilience of Italy's National Transmission Grid.

Through the use of supercomputing, the RETE project generated detailed forecasts of how extreme precipitation events are expected to evolve over the next thirty years. These data were then cross-referenced with a national landslide susceptibility map, a tool that identifies areas most prone to ground instability based on several factors, including geological characteristics, slope gradients and soil type. More broadly, the integration of climate scenarios and territorial data made it possible to develop a landslide risk index, enabling the identification of National Transmission Grid (NTG) assets located in areas with a higher susceptibility to landslide initiation and the prioritization of the most exposed network sections, while assessing their strategic role within the grid.

A detailed analysis of individual slopes in the most exposed areas, through the application of a site-specific numerical model, made it possible to accurately quantify the probability of landslide occurrence based on both historical and projected climate inputs, thereby complementing and strengthening the large-scale statistical approach. This made it possible to estimate how a potential disruption to a given piece of infrastructure could affect the continuity and security of the energy supply.

Building on the results of the RETE project, a new initiative called LARA (Landslide rAinfall-induced Resilience Assessment on Power Systems) is now being launched, with Terna as the lead partner alongside leading institutions such as ENEA, the Italian National Agency for New Technologies, Energy and Sustainable Economic Development, CMCC, a research centre specialising in climate and its impacts, and Politecnico di Milano. The project aims to turn the results achieved through RETE into an operational tool capable of providing concrete and continuous support for managing, monitoring and strengthening the resilience of the electricity grid.

One of the main innovations introduced by LARA will be greater accuracy in weather forecasting. Compared with approaches based on daily averages, which may fail to capture sudden and intense events, the project will use convection-permitting models capable of analysing precipitation at hourly or sub-daily resolution, improving the ability to detect extreme and rapidly developing weather events.

The RAPID project will also integrate advanced machine learning techniques to produce more accurate hazard maps, combining data on soil moisture and saturation with records of past landslide events. Another innovative element will be the simulation of landslide propagation, making it possible to estimate the volume of material involved and the trajectory of debris, and thus assess more precisely whether, and which, electrical infrastructure could actually be affected.

To manage the complexity of simulations at the national scale, RAPID will leverage supercomputing and Monte Carlo simulations, an advanced statistical technique used to analyse phenomena characterised by high levels of uncertainty, such as natural hazards, which cannot be described through a single deterministic forecast. Within the RAPID project, this approach will make it possible to model thousands of possible developments of hydrogeological risk, assessing different combinations of weather events and ground conditions. Through these simulations, Terna will be able to identify vulnerabilities in the national electricity grid in advance and shift from a damage-response approach to a preventive strategy, directing investments towards strengthening infrastructure and increasing resilience to the effects of climate change.