AI and Energy Management

Short-term electricity generation and transmission planning is a fundamental optimization problem for utilities such as Hydro-Québec. In addition to relying on complex modelling that takes operational realities into account, this planning must evolve in a highly stochastic environment. In this presentation, I will demonstrate how recent advances in artificial intelligence can solve this type of problem. I will also discuss the challenges that must be addressed when implementing such solutions in critical IT systems.

This presentation explores the growing role of artificial intelligence (AI) in energy management, following the different stages of an energy project: from planning to consumption, including operation and maintenance. Through concrete and current examples, we will see how AI is already transforming the sector, in particular through:

- Long-term demand forecasting, integrating the impacts of climate change;

- Predicting short-term energy supply, optimizing production and integration of renewable sources;

- Optimizing infrastructure maintenance, reducing costs and interruptions;

- Intelligently managing consumption, promoting energy efficiency and reducing emissions.

In the second part, the conference will focus on potential and emerging applications of AI, such as large language model (LLM) applications in energy and applications in projects of common interest. Projects of Common Interest (PCI) are energy infrastructure projects (gas, electricity, hydrogen) considered crucial by the European Union for achieving the objectives of energy security, affordability, sustainability, and decarbonization. They are evaluated according to criteria such as contribution to EU energy policy, market integration, and climate objectives. Artelys assists developers in evaluating these criteria and will present a use case illustrating the various possible applications of AI to this type of project.