PC4: SYNCHRONIC

Overview

SYNCRONIC addresses the process that supports bidirectional data flows between physical and digital twins, examining the non-functional performance of DTs. The primary objective is to explore theories, methods, and technologies for the digital coupling process.

While significant attention has been given to constructing models that accurately represent a system’s behavior, there is a need for studies focusing on the non-functional performance attributes of digital twins (DTs), such as reliability, maintainability, security, and connectivity.

Establishing a closed-loop process between a physical system and its digital twin generally assumes continuous observation or measurement of the system’s full state. However, this assumption becomes less reliable when communication networks or buses are involved, whether linking the sensors to the twin or the twin to the actuators. The main challenge is to define and verify the necessary conditions, such as data freshness, availability, reliability, and security, that ensure accurate execution of digital twin models and real-time synchronization with the physical world. Efficient synchronization also brings to the forefront critical questions about emerging technologies like IoT, 6G networks, and the Cloud-Edge continuum—particularly how they can address constraints such as latency, precision, and efficiency (e.g., through data compression).

Designing a data collection strategy for digital twins (Smart data) presents a significant challenge and involves multiple dimensions: balancing data collection with twin requirements, determining the optimal sensor placement, minimizing energy consumption, and ensuring high-quality data collection. While data is important, maintaining the link between the data and its context is crucial. Identifying metadata models required by digital twins and how metadata can be produced and used to help managing models are open questions.

Maintaining synchronization between the digital twin and its physical counterpart, particularly in dynamic environments where conditions change frequently, requires either the consistent and accurate integration of real-world data into the digital model (data assimilation) or the adjustment of the digital twin’s state at specific intervals by delaying, triggering, or rolling back the effects of events within the model (complex event processing). However, implementing this is highly challenging due to the complexity of nonlinear, multiscale, and multiphysics phenomena involved in complex systems, which are often tied to computationally intensive simulations. Achieving effective sequential data assimilation and control through artificial intelligence (AI) methods—such as deep learning—that combine both physical models and data-driven algorithms remains an open challenge.

The project focuses on smart data collection, synchronization, adaptation and control, and network optimization with clear objectives: developing comprehensive frameworks for efficient data utilization, establishing real-time synchronization methods, creating adaptive control mechanisms, and implementing network co-optimization for scalable deployment.

Associated Use Cases

Use cases in PC4 require complex systems or systems operating under fluctuating environmental and operational conditions, necessitating real-time monitoring and dynamic adjustments. Secure and safe integration within a continuum computing environment is also critical.

Investigator & Project Partners

Principal Investigator:

Hind Bril El Haouzi

Full Professor at Université de Lorraine

Hind Bril El Haouzi is a full Professor at the Université de Lorraine where she teaches and conducts research in the fields of Computer Engineering and Production Management Control. Since 2018, she has been co-leading the Sustainable Industrial System Engineering Research group at the (CRAN,CNRS) Laboratory. She obtained her PhD in Computer Science, Automatic Control, and Production Engineering in 2008. She has a wealth of industrial experience, having worked for many years as a digitization project leader before joining Université de Lorraine. She currently coordinates several collaborative projects with the industrial sector that focus on the challenges of digital transformation and the societal transition of industry (ANR PRCE, ADEME, CPER,CIFRE…). Her research interests include modeling and control of cyber physical production systems, with a focus on digital simulation and distributed manufacturing control. She has published over 100 papers in international conferences and journals.

Participating Partners:

Project Implementation

The project is structured into three complementary work packages, covering the essential dimensions of the coupling between the physical system and its digital twin: frugal data collection and optimal network deployment, spatio-temporal synchronization and verification, as well as adaptation and control based on real-world data. This organization is intended to ensure reliable, secure, high-performance, and resource-efficient exchanges between the physical world and its digital counterpart.