Publications

The paper investigates the influence of adaptive sampling strategies on the generation of a metamodel based on Non-Uniform Rational Basis Spline (NURBS) entities, obtained from unstructured data, with the purpose of improving accuracy while minimising computational resources. The metamodel is defined as solution of a constrained non-linear programming problem and it is solved through a three-step optimisation process based on a gradient-based algorithm. Moreover, this paper introduces a generalised formulation of the NURBS-based metamodel capable of handling unstructured sampling data, enabling simultaneous optimisation of control points and weights. Sensitivity analyses are performed to evaluate the influence of various adaptive sampling techniques, including cross-validation-based and geometry-based strategies, on the resulting metamodel, in terms of accuracy and computational costs. Analytical benchmarks functions and a complex real-world engineering problem (dealing with the non-linear thermomechanical analysis of a part produced with the fused deposition modelling technology) are used to prove the effectiveness of the NURBS-based metamodel coupled with adaptive sampling strategies in achieving high accuracy and efficiency. © 2025 Elsevier B.V.

When performing upscaling of transport phenomena in multiscale systems it is not uncommon that terms of different physical nature than those present at the underlying scale arise in the resulting averaged differential equations. For diffusive species mass transfer with heterogeneous reaction and conductive heat transfer, additional terms result from upscaling using the volume averaging method, which are classically discarded by means of orders of magnitude estimates. In this work, these two cases are revisited and it is shown that, for single and two-species diffusive mass transfer with heterogeneous nonlinear reaction, the additional term is exactly zero using Green's formula. This conclusion is shown to also be applicable when using the periodic homogenization method. Nevertheless, for heat conduction, with and without considering interfacial resistance, only the dominant conduction-corrective terms are shown to be zero also using Green's formula. In contrast, the contribution of the co-conduction-corrective terms may be relevant depending on the systems characteristics, the properties of the phases and the macroscopic boundary conditions. This is exemplified by performing numerical simulations in a non-symmetric unit cell. © 2025 Elsevier Ltd

In some laser assemblies, optical components are glued on their metal mount using a photocrosslinkable adhesive. All these components are submitted to medium to high temperatures for several hours upon laser's operation time. The subsequent thermal stresses endured by the adhesive could lead to its degradation, alter thus the functioning of the assembly and impact the alignment of the laser, key issues in laser applications. This work focuses on the investigation and the modeling of the lifespan, the degradation in the face of thermal stresses, particularly those generated by a laser, and of the ageing of a photocrosslinkable adhesive, PC373HA. The thermal characterization is performed using ThermoGravimetric Analysis and allows for the development, the validation and the comparison of three models based on Arrhenius' law to estimate the lifetime of adhesives as a function of temperature. A temperature of 150 °C is identified as a threshold to ensure limited degradation over the 15 h requested for laser operation endurances. The Ozawa-Flynn-Wall model and the Kissinger-Akahira-Sunose model correctly reproduce PC373HA's thermal behavior submitted to different thermal scenarii (temperatures, heating rates …). Both show that the higher the temperature, the faster the degradation process, approximately 10 min at 220 °C and 100 min at 180 °C, much less than the required 15 h. Besides, in the event of an incident with the laser, if the temperature exceeds 200 °C for more than 10 min or 180 °C for more than 100 min, corrective action should be taken, and the adhesive should be replaced. These models therefore provide fundamental information for laser applications and will allow the implementation of preventive solutions during use but also in case of incident. © 2025 Elsevier Ltd

The upscaling process of coupled (single- and two-species) diffusion with heterogeneous chemical reaction in homogeneous porous media is revisited in this work with several important clarifications following the article from Bourbatache et al. (Acta Mech 234: 2293-2314, 2023. https://doi.org/10.1007/s00707-023-03501-w). It is shown that the upscaled model obtained from the volume averaging method (VAM) or, equivalently, following an adjoint and Green’s formulation technique provides a closed model without any a priori assumption on the form of the solution for the pore-scale concentration involved in the spectral approach used in the periodic homogenization method (PHM) reported in the above reference. Through comparison with direct pore-scale simulations, the VAM model is shown to outperform the predictions of the average concentration and average flux profiles for the simple two-dimensional configuration considered in Bourbatache et al. (Acta Mech 234: 2293-2314, 2023. https://doi.org/10.1007/s00707-023-03501-w) in comparison with the model obtained from PHM in this reference. Finally, identification of the apparent effective diffusion coefficient from these pore-scale simulations, which serve as in silico experiments, proves that the correct dependence upon the Damkhöler number is the one predicted by the model obtained with VAM, in contradiction with the conclusion put forth in Bourbatache et al. (Acta Mech 234: 2293-2314, 2023. https://doi.org/10.1007/s00707-023-03501-w). The physical explanation lies in the corrective contribution of the reactive part to the apparent effective diffusion coefficient, which is positive and adds up to the pure intrinsic diffusive part. The discrepancy between PHM and VAM approaches is proved to originate from the choice of changes of variables in the pore-scale concentration used in the spectral approach while employing PHM. © The Author(s), under exclusive licence to Springer-Verlag GmbH Austria, part of Springer Nature 2025.

This paper presents an experimental and numerical analysis of heat transfer and thermal deformation in small-dimension (1.4 mm) liquid mechanical seals operating in an unstable dynamic tracking mode. The studied non-contacting mechanical seal is used in a liquid pump for turbojets. The study aims to estimate the values of pressure, temperature, and thermal deformations that can prevent excessive wear of the sealing rings and control the increase in leakage rate or power loss during operation. Experimental investigations were conducted under a nominal inner pressure of 0.7 MPa, across a wide range of rotational speeds (from 1000 to 6000 rpm), and at low Reynolds numbers (Re < 70). Two high-viscosity fluids, glycerol and engine oil, were used as sealing fluids. Rotational speeds and inner pressure were set as boundary conditions in the simulations. Temperatures measured by thermocouples during the experiments were used to compare with the simulation results. Simulations were performed using the computational fluid dynamics (CFD) software COMSOL. The two-dimensional numerical models accounted for thermal transfers and face seal deformations, coupled with the pressure field in the lubrication fluid. The effects of various sealing fluids and rotational speeds on the time-dependent behavior of temperature, displacement, and pressure within the thin liquid lubricant film were investigated. Subsequent comparisons between experimental and numerical results, particularly for temperature data measured by thermocouples under various operating conditions, demonstrated strong consistency. The greatest discrepancy observed was less than 1.2 °C. © 2025 The Authors

Over the past few years, femtosecond (fs) laser processing has drawn a growing interest in a wide range of applications as it offers the possibility to process the surface morphologies of metals and semiconductors. In contrast to other polishing techniques, laser polishing offers a flexible and non-contact solution, thereby avoiding potential external contamination, while enabling a precise selection of processing areas. We investigated the influence of fs laser parameters on surface roughness of pure copper and ablation thickness, focusing on highlighting the importance of fluence and scanning overlap. With a two-step processing strategy, composed of coarse and fine polishing steps, surfaces with Sa < 400 nm were achieved, representing a 98% reduction from the high roughness of 15 μm on initial surfaces. This research demonstrated the possibility of directly polishing rough parts using a fs laser with a perpendicular incidence. © 2025

Virtual Reality (VR) sketching is a valuable tool for conceptual understanding, creativity, and design, but quality issues can hinder its adoption. To address this, we conducted a study involving 15 novices and 15 experts who sketched three chair models in static, mobile, and control conditions. The results showed that mental rotation skills, training, model type, and movement impact sketch quality. The static condition negatively affected performance, particularly volume and proportion. Conversely, the mobile condition didn't improve sketch quality compared to the control group. 3D perception seems tied to movement, highlighting the need to adapt VR sketching software for these challenges. Enhancing the user experience and addressing these quality concerns will be pivotal in the widespread acceptance of VR sketching tools.

Functional assessment is a key element in evaluating adult spinal deformity (ASD) patients. The multitude of 3D kinematic parameters provided by movement analysis can be confusing for spine surgeons. The aim was to investigate movement patterns of ASD based on key kinematic parameters. 115 primary ASD and 36 controls underwent biplanar radiographs and 3D movement analysis during walking, sit-to-stand and stair ascent to calculate joint and segment kinematics. Principal component analysis was applied to identify the most relevant kinematic parameters that define movement strategies adopted by ASD. Pelvis and head relative to pelvis kinematics were the most relevant parameters. ASD patients adopted four different movement strategies. Class 1: normative head and pelvis kinematics. Class 2: persistent pelvic backward tilt. Class 3: persistent forward shift of the head. Class 4: both pelvic backward tilt and forward shift of the head. Patients in class 3 and 4 presented sagittal malalignment on static radiographs with increased pelvic tilt, pelvic incidence-lumbar lordosis mismatch and sagittal vertical axis. Surprisingly, patients in class 3 had normal pelvic kinematics during movement, showing the importance of functional evaluation. In addition to being key segments in maintaining static global posture, head and pelvis were found to define movement patterns.

Femoro-acetabular impingement (FAI) may present as alterations in the skeletal morphology of the hip. Repercussions of FAI can be witnessed in self-selected speed walking as well as physical exercise such as running or fast speed walking. The aim of this study was to investigate changes in kinematics at different gait speeds in subjects presenting with radiological findings invoking FAI. One hundred thirty asymptomatic adults underwent biplanar X-rays with a calculation of 3D hip parameters: acetabular anteversion, abduction and tilt, vertical center edge angle (VCE), femoral anteversion, neck-shaft angle, acetabular coverage of the femoral head, femoral head diameter and neck length. Parameters were classified according to FAI clinical thresholds. Two groups were created: Control group (63 subjects having up to one subnormal hip parameter in favour of FAI) and Radiographic FAI group (67 subjects having ≥2 subnormal hip parameters that might cause FAI). All subjects underwent 3D gait analysis at self-selected and fast speed, from which kinematic parameters were generated. Arithmetic differences between fast and self-selected speed gait were considered as gait changes. Subjects in the Radiographic FAI group had decreased acetabular tilt (24 vs. 19˚), anteversion (19 vs. 16˚), abduction (55 vs. 53˚), femoral anteversion (18 vs. 14˚) and increased VCE (29 vs. 33˚, all p<0.05), compared to controls. Changes from self-selected to fast speed showed that subjects in the Radiographic FAI group had lower range of motion (ROM) pelvic rotation (7 vs. 4˚) and ROM hip flexion/extension (10 vs. 7˚), reduced hip extension (-4 vs. -2˚) and step length (16 vs. 13 cm; all p<0.05). The Radiographic FAI group had decreased acetabular abduction, anteversion and femoral anteversion in favour of FAI. When adapting from self-selected to fast speed gait, the Radiographic FAI group seemed to limit pelvic rotation and hip flexion/extension resulting in a decrease in step length. These kinematic limitations were previously reported in subjects with symptomatic FAI. Gait analysis could be considered as a functional diagnostic tool to assess FAI along with radiological assessment.

Purpose To investigate kinematic adaptations from self-selected to fast speed walking in ASD patients. Methods 115 primary ASD and 66 controls underwent biplanar radiographic X-rays and 3D gait analysis to calculate trunk, segmental spine and lower limb kinematics during self-selected and fast speed walking. Kinematic adaptation was calculated as the difference (Δ) between fast and self-selected speed walking. ASD with 7 or more limited kinematic adaptation parameters were classified as ASD-limited-KA, while those with less than 7 limited kinematic adaptation parameters were classified as ASD-mild-KA. Results 25 patients were classified as ASD-limited-KA and 90 as ASD-mild-KA. ASD-limited-KA patients walked with a lesser increase of pelvic rotation (Δ = 1.7 vs 5.5°), sagittal hip movement (Δ = 3.1 vs 7.4°) and shoulder–pelvis axial rotation (Δ = 3.4 vs 6.4°) compared to controls (all p < 0.05). ASD-limited-KA had an increased SVA (60.6 vs − 5.7 mm), PT (23.7 vs 11.9°), PI–LL (9.7 vs − 11.7°), knee flexion (9.2 vs − 0.4°) and a decreased LL (44.0 vs 61.4°) compared to controls (all p < 0.05). Kinematic and radiographic alterations were less pronounced in ASD-mild-KA. The limited increase of walking speed was correlated to the deteriorated physical component summary score of SF-36 (r = 0.37). Discussion Kinematic limitations during adaptation from self-selected to fast speed walking highlight an alteration of a daily life activity in ASD patients. ASD with limited kinematic adaptations showed more severe sagittal malalignment with an increased SVA, PT, PI–LL, and knee flexion, a decreased LL and the most deteriorated quality of life. This highlights the importance of 3D movement analysis in the evaluation of ASD.