Page 36 - Sirius_Annual_Report_2021

Analysis of Digital Twins
A digital twin is typically a system which collects data about a physical asset such as a plant or a reservoir.
The digital twin is a vision for a technology, originally conceived for NASA’s space program, enabling industry to significantly improve the life-cycle management of physical assets. A digital twin is typically a system which collects data about a physical asset (such as a plant or a reservoir), continuously revises this data set through, e.g., updates reflecting changes to the asset’s structure and sensor data reflecting the physical asset’s state and uses this data to monitor and make predictions about the physical asset. The digital twin can be thought of as a three-layered structure: the data sources, an information layer, and an insight layer. Industrial focus is today mainly on collecting data into shared, and increasingly structured, data sets which we think of as the information layer, and on providing dashboard- like insights into the system.
This project will focus on analysis support for digital twins, by building or combining tools which can leverage the information layer into insights. The purpose of these tools can be to reproduce and explain past events, to explore alternatives for decision making, to prepare for incidents or to optimize production. A central goal for this project is to combine semantics, behavioral and conceptual modelling techniques, and analysis methods in the context of digital twins.
Methodological background for the work is an integration of ontology-based conceptual modelling techniques, formal methods, and data-driven techniques for system analysis.
• Understand how conceptual modelling can be used to integrate analyses results.
• Develop experience with semantics foundations for co-simulation.
• Develop methods for decision making with digital twins. Objectives Activities
The work has synergies with, and feeds technology to the PeTWIN and other Digital Twins projects.
• •
Understand the design space for coupling behavioral and conceptual models.
Develop methods that combine structured information with behavioral analyses.
• Develop a formal theory of coupled behavioral and conceptual models.
• Develop prototype tool for programming with semantics.
36
| SIRIUS ANNUAL REPORT 2021