SureROB at ECSC 2026

Optimal design and operation of mechanical systems need for fast and accurate predictions, while control needs moreover for efficient data assimilation techniques. Recent technologies on model order reduction and data-driven modelling enable the construction of efficient (fast and accurate) digital twins of structural systems, conciliating fast responses with accurate predictions. This presentation will describe the main components of the structure digital twin, with efficient solvers of the physics-based models, efficient data-assimilation techniques, and data-driven modelling, enabling efficient design, operation and structural control.

Industrial robots will play a crucial role in the transition toward sustainable manufacturing – both by becoming more energy- and resource-efficient themselves and by enabling the production of green products through cleaner, smarter processes. Yet achieving this dual role poses significant challenges across the entire robotic system lifecycle. How can robots be designed to reduce material use and energy consumption without compromising performance, robustness, or cost? How can control, planning, and monitoring strategies contribute to greener operation under real industrial constraints?

This keynote frames these questions at the outset of a new research effort, focusing on fundamental trade-offs in mechanical design, actuation, control, and system integration. Rather than presenting final solutions, early ideas and promising directions are discussed. Including lightweight and modular structures, energy-aware control concepts, lifecycle-oriented optimization, and application-driven benchmarking, the perspective covers the full spectrum of industrial robots — from high-performance automation to collaborative systems.

Procedures of control or monitoring structural behavior rely, on the one hand, on measurements and, on the other, on mechanical models. In fact, analyzing data within an appropriate physical framework enhances the effectiveness of the identification methods used for control, prediction, or diagnosis. However, complex mechanical systems may exhibit behaviors that are not captured by standard mechanical models.

This presentation focuses on the identification, via homogenization, of non-standard continuous models involving higher-order gradients.

In the first part, we will present examples of composite reinforced with rigid fibers and laminated glasses, the analysis of which leads, respectively, to second-order gradient media and tri-Laplacian plates. The relevance of these models is confirmed by experimental results.

In the second part, we will focus on media exhibiting an internal mechanism. Consequently, in addition to rigid-body modes, there exists a zero-energy deformation mode (or floppy mode) whose amplitude appears as an additional kinematic descriptor. In the case of pantographic plates, the floppy mode is a deviatoric mode, and the corresponding model is a second-gradient model in shear. In the case of periodic bi-parallelograms articulate system, the floppy mode is of constant curvature, and the 1D continuous model is then a beam whose the elastic energy depends on the curvature gradient.

In practice, these enriched media allow for a description consistent with the physics of the medium. It is also important to note that special attention must then be paid to the boundary conditions, which differ from those of standard media.

It is well known that the prediction of the medium to high frequency vibroacoustic response of an engineering system can pose severe computational difficulties, due to both the large number of degrees of freedom required in a model and the sensitivity of the response to small random manufacturing variabilities. These difficulties have led to the development of analysis techniques which are inherently statistical, such as Statistical Energy Analysis (SEA), which seek to predict the mean and variance of the energy of the response, averaged over regions, or subsystems.Su ch methods are clearly useful in the design of passive vibration control: the response can simply be recalculated for various design changes (for example, the addition of soundproofing) and an optimal combination of performance and cost can be sought. However, these methods are not normally associated with the design of active control systems, since the phase of the response is not predicted by the approach. It is shown here that progress can be made towards employing high frequency methods in active control design by considering hybrid deterministic-statistical models and/or a property known as the analyticity-ergodicity condition.

The acoustic black hole (ABH) effect utilizes variations in structural parameters or material properties to reduce wave velocity within a structure. The most common approach to creating an ABH involves tailoring the thickness profile to concentrate energy in specific regions. Owing to their high efficiency, broadband performance, and design flexibility, ABH structures show great promise for vibration and noise reduction in thin-walled systems. This talk reviews recent progress in the modeling, analysis, implementation, and experimental characterization of ABH structures, together with their applications in vibration damping, noise control, and energy harvesting. Modeling approaches include semi-analytical wavelet methods for one-dimensional ABHs, finite element methods for two-dimensional ABHs, and wave field visualization via laser ultrasonics. Implementation strategies such as embedded ABHs and add-on ABH-based dynamic vibration absorbers (ABH-DVAs) are presented. The mechanisms of panel vibration damping through embedded ABHs and ABH-DVAs, cavity noise reduction using ABH panels, and enhanced energy harvesting performance are discussed. Finally, applications of ABH technology to noise and vibration mitigation in high-speed trains and helicopters are highlighted.