Data science to support the grid

The goal: to find innovative solutions able to enhance the company's data assets and support Terna’s predictive maintenance and monitoring of its electricity infrastructure to deliver a grid that is increasingly secure, efficient, sustainable, and above all resilience to extreme weather events. Terna calls, and the start-ups respond in droves: so say the figures of the Data Science for Resilience call, the first to be launched by Terna on the new Terna Ideas Open Innovation platform.

Launched by Terna on 15 December 2022, the call received 71 applications - a resounding success - each proposing solutions focused on data analytics and artificial intelligence for image recognition. Following the conclusion of the initial screening process on 6 April, Terna has selected the 10 finalists who will have the opportunity to present their projects during demo day, held on 17 May at IDEASpark in Turin, one of the company's dedicated innovation venues.

The projects were assessed against seven evaluation criteria: business opportunity, proposed innovation level, compatibility with Terna processes, consistency and robustness of the solution, effectiveness of presentation, team quality, and project scalability.

Terna will now assess the most suitable start-ups that align with its development strategy, with a view to launching new pilot projects to test the most promising solutions. The next step will involve a testing phase to produce the Proof of Concept, which will be funded by Terna. The innovative companies that earned a place on the podium deployed all their skills and the most innovative tools to promote their solutions: drones, artificial intelligence, sensors, video cameras, telematics platforms, automatic calculation, and analysis systems to monitor the grid and increase its resilience to increasingly extreme and frequent weather events.

Terna Ideas: corporate innovation opens its doors to the outside world. The Data Science for Resilience challenge is one of the initiatives promoted through Terna Ideas, Terna’s new Open Innovation platform open to start-ups, innovative SMEs, researchers, professionals, and even members of Terna’s own workforce. The Terna Ideas platform represents an open innovation tool open to anyone wishing to play a key role in the energy transition and contribute actively to the process by proposing ideas and initiatives.

Why is data science important in the electricity sector? The amount of data produced and used has grown enormously in recent years, and effectively analysing it represents a huge challenge due to both its volume and diverse nature. System operators are improving grid observability by installing electrical meters that provide data at frequent time intervals, as well as smart meters in customers’ homes.

Meanwhile, sensors are increasingly used on renewable energy plants to monitor various components (turbines, panels, etc.) which, again, generate a large volume of data. Finally there is the environmental aspect, and all of the data from ground-based weather stations and satellites that this involves. Data analytics and artificial intelligence technologies are already widely used in the energy sector and are expected to be deployed increasingly on an industrial scale to improve energy efficiency, reduce costs and facilitate the transition to renewable energy sources.

The challenge. Terna has organised several initiatives focused on the innovation ecosystem, but the Data Science for Resilience call is the first to be launched on the Terna Ideas open innovation platform. The aim of the initiative is to identify innovative solutions based on the analysis of data and/or images to improve predictive maintenance along the national electricity grid operated by Terna, and to optimise investments to transfer the company's data assets into usable information for the business.

The initiative also calls for the use of weather forecasting data and artificial intelligence to recognise and classify images. The ultimate aim is to make the high-voltage transmission infrastructure safer, more efficient, resilient and sustainable. The challenge focuses on identifying hi-tech solutions able to process data analysis models and algorithms through a “what if” approach.

Terna selected ten finalist start-ups. Their ideas and proposed projects are described below:

• BlueTensor specialises in the development of artificial intelligence for image reconstruction and recognition. The proposed idea, entitled Eyerus, is a product developed to recognise the quality levels of Terna's assets, using images acquired from video surveillance cameras, drones, etc. The company also provides customised artificial intelligence solutions.
• Elif Lab develops innovative solutions using data science and artificial intelligence to solve complex problems. The VisioDian project is able to recognise and assess asset anomalies thanks to monitoring using fixed cameras and sensors. The solution runs on low-power edge platforms and can generate real-time alerts to report anomalies.
• Latitudo 40 combines the skills and experience of a large company with the innovation ability of a lean organisation. Earth Analytics for Energy Infrastructures concentrates on monitoring Terna’s power lines. Using satellite imagery, it provides services such as building detection, landslide risk analysis and green monitoring along the grid.
• Studiomapp specialises in geo-intelligence and top-down image analysis. With Strix by Studiomapp it is able to process remote-sensor data from different platforms, such as satellites, aircraft and drones, to extrapolate information to manage power lines. The solution can quickly analyse large amounts of data and provide relevant information through dashboards, reports or APIs.
• Deix uses mathematical tools, optimisation algorithm and artificial intelligence to find solutions to complex industrial problems. Through NAMO it provides support for the monitoring, maintenance and optimisation of complex plants and systems. The DDO platform enables the integration of data from different sources for asset monitoring and analysis, enabling the construction of complex correlations and estimates of future trends.
• Clearbox AI offers statistical models based on the “what if” approach to data analysis and AI. It uses generative artificial intelligence to create hypothetical scenarios and synthetic data.
• Fast Computing specialises in statistical models based on the “what if” data analysis approach, focusing on diagnostics and predictive maintenance of complex systems using digital twins and automatic analysis and calculation methods, including Principal Component Analysis (PCA), Mahalanobis distance and autoencoder neural networks.
• Eoliann uses satellite data and machine learning algorithms to predict the probability and impact of extreme weather events. The Eoliann Flood API proposal provides information on natural disaster forecasts worldwide, also considering climate change.
• Amigo provides climate services based on geo-spatial data. It conducts studies and provides consultancy services on climate risk analysis for the energy, agriculture, insurance and transport sectors. The NUA project uses statistical methods to identify the areas most susceptible to the effects of increasingly frequent extreme weather events.
• AlphaEdge is a start-up incubated by the European Space Agency (ESA) and focuses on data analytics for weather and climate forecasts. It offers grid risk management solutions for electricity transmission operators, using sensor data, weather models and satellite data to monitor network infrastructure, predict high-risk weather events and assess potential damage.

Innovation: the cornerstone of Terna's business. In the latest update of the Driving Energy 2021-2025 Business Plan, Terna has allocated €1.2 billion in investments (out of a total €10 billion) for innovation and digitalisation. In particular, the company is focusing on new technological solutions for the remote management of substations and transmission infrastructure, thanks to the use of sensors, monitoring and diagnostic systems, predictive management tools, virtual reality and artificial intelligence, aimed at improving the security, reliability and resilience of the electricity grids and providing benefits for local communities.