• Liudmila V. Massel, Melentiev Energy Systems Institute SB RAS (Irkutsk, Russia)
• Aleksey R. Tsybikov, Melentiev Energy Systems Institute SB RAS (Irkutsk, Russia)
• Nikita I. Shchukin, Melentiev Energy Systems Institute SB RAS (Irkutsk, Russia)

Abstract. In the context of the global transition to a low-carbon economy, the design of renewable energy facilities such as solar and wind power plants require consideration of numerous hard-to-predict factors: variability of natural resources, terrain topography, environmental constraints, and economic parameters. Digital twins (DTs) represent a powerful tool for addressing these challenges by providing a virtual representation of a physical asset throughout its entire lifecycle. This article examines proposed methods for constructing Smart Digital Twins, (SDTs) for use in renewable energy facility design. The foundation of the proposed approach is a modified digital twin model based on ontological models. These models formalize key concepts, entities, their attributes within the renewable energy domain, as well as the semantic relationships among them. The ontology provides a unified glossary and data structure, which is critical for integrating heterogeneous information sources and ensuring mutual understanding among system components and specialists. The next step involves transforming this ontology into a smart digital twin model by incorporating intelligent components such as knowledge bases, a virtual environment, artificial intelligence models, schemas, and diagrams. To describe the relationships between models and components, a fractal stratified model is proposed. This model formalizes the knowledge structure and interconnections among ontological, informational, and mathematical models. The article details an adapted ontological engineering methodology tailored for digital twin design tasks, along with a method for constructing a virtual environment that enables debugging of both digital twins and smart digital twins in conditions of intermittent or entirely absent connectivity to the physical asset. To emulate external parameters such as weather conditions, a modified process based on CRISP-DM (Cross-Industry Standard Process for Data Mining) is proposed, facilitating the integration of machine learning models. The practical relevance of the approach is demonstrated through the development of a visualization component for identifying optimal power plant locations and supporting their design. This tool leverages an interactive 3D model of the Earth, satellite data, and meteorological APIs. The implemented solution confirms the feasibility of applying smart digital twins to renewable energy facility design.

digital twin, smart digital twin, renewable energy, fractal-stratified model, ontology, machine learning, visualization

2026-03-05