It was the most awaited topic of this 7th edition of Viva Technology, a prominent event on innovation and technology held in Paris in June 2023. Artificial intelligence has been on everyone's mind since the recent achievements of generative AI, popularized by ChatGPT and MidJourney. So, what are the implications for ENGIE?
To optimize energy infrastructure
AI represents a powerful driver for optimizing energy infrastructure. It was the focus of a panel discussion at VivaTech, moderated by Mihir Sarkar, Head of AI at ENGIE Research & Innovation. The Group is building solar and wind parks in many countries. Traditionally, to ensure sufficient productivity, operators measure sun or wind for a year or two before giving the green light. This is where AI comes into play. "By utilizing data on climate, temperature, precipitation, irradiance, etc., artificial intelligence can significantly accelerate the pre-evaluation phase of sites," emphasizes Ludovic Quesnelle, Chief Digital IT Officer Renewables Trading & Nuclear at GEMS (Global Energy Management & Supply). This also ensures determining, for example, the optimal location and orientation of a wind turbine.
To document industrial infrastructure
Moreover, the Group is harnessing AI to generate a reliable technical inventory of industrial infrastructure firmly grounded in real-world conditions. Storengy is collaborating with the startup SAMP to update its knowledge of storage sites, some of which date back several decades. This patented technology, based on 3D artificial intelligence, reconciles all technical documentation of a facility with 3D terrain scans. “We can also exhaustively catalog all 12,000 site components, such as pumps, valves, pipes, etc.,” says Antoine Boudehent, Industrial Digital Project Manager at Storengy. "Using traditional methods, we would still be working on it, taking several years," he adds. The result is valuable: enabling all involved teams (operators, maintenance, engineers, subcontractors), and especially newcomers, to better grasp the complexity of a storage site and share a common reality, thus facilitating and accelerating all interventions. "In terms of safety, we avoid the risks of unpleasant surprises", Antoine Boudehent further highlights.
To predict our clients'consumption
The promises of AI extend far beyond infrastructure and encompass all entities within the Group. Another panel discussion at VivaTech, titled "Data-driven success at ENGIE" and moderated by Jean-Pierre Pélicier, Group Chief Data Officer, precisely addressed the value created by AI through various examples, including that of GEMS. As the core of the Group's integrated model, the Trading Global Business Unit (GBU) has the mission of ensuring supply security and a constant balance between what the Group produces and what customers consume. It deals with a petabyte of data. Estimating the impact of even half a degree temperature variation is crucial. The growing influence of renewable energies and the development of electric vehicles have made forecasting much more complex than in the past. "Network operators now ask us to predict our clients' consumption every fifteen minutes, and in some countries, soon down to five minutes", emphasizes Ludovic Quesnelle. Machine learning algorithms will prove essential in meeting such a challenge, alongside an underlying task: recruiting and retaining the talent required for this revolution.