École nationale supérieure d'arts et métiers
Recent publications
Automated detection of lathe checks in wood veneers presents significant challenges due to their variability and the natural properties of wood. This study explores the use of two convolutional neural networks (U-Net architecture) to enhance the precision and efficiency of lathe checks detection in poplar veneers. The approach involves sequential application of two U-Nets: the first for detecting lathe checks through semantic segmentation, and the second for refining these predictions by connecting fragmented lathe checks. Post-processing techniques are applied to denoise the mappings and extract precise lathe check characteristics. The first U-Net demonstrated strong performance in predicting lathe check presence, with precision and recall scores of 0.822 and 0.835, respectively. The second U-Net refined predictions by linking disjointed segments, improving the overall lathe checks mapping process. Comparative analysis with manual methods revealed comparable or superior performance of the automated approach, especially for shallow lathe checks. The results highlight the potential of the proposed method for efficient and reliable lathe check detection in wood veneers.
The finite‐element method (FEM) is a well‐established procedure for computing approximate solutions to deterministic engineering problems described by partial differential equations. FEM produces discrete approximations of the solution with a discretisation error that can be quantified with a posteriori error estimates. The practical relevance of error estimates for biomechanics problems, especially for soft tissue where the response is governed by large strains, is rarely addressed. In this contribution, we propose an implementation of a posteriori error estimates targeting a user‐defined quantity of interest, using the dual‐weighted residual (DWR) technique tailored to biomechanics. The proposed method considers a general setting that encompasses three‐dimensional geometries and model nonlinearities, which appear in hyperelastic soft tissues. We take advantage of the automatic differentiation capabilities embedded in modern finite‐element software, which allows the error estimates to be computed generically for a large class of models and constitutive laws. First, we validate our methodology using experimental measurements from silicone samples and then illustrate its applicability for patient‐specific computations of pressure ulcers on a human heel.
The demand for efficient Industry 4.0 systems has driven the need to optimize production systems, where effective scheduling is crucial. In smart manufacturing, robots handle material transfers, making precise scheduling essential for seamless operations. However, research often oversimplifies the Robotic Flexible Job Shop problem by focusing only on transportation time, ignoring resource allocation and robot diversity. This study addresses these gaps, tackling a Multi-Robot Flexible Job Shop (MRFJS) scheduling problem with limited buffers. It involves non-identical parallel machines and robots with varying capabilities overseeing material handling under blocking conditions. The case study is based on a real Industry 4.0 scenario, where the layout restricts each robotic arm’s access, requiring strategic buffer placement for part transfers. A Mixed-Integer Programming (MILP) model aims to minimize makespan, followed by a new Genetic Algorithm (GA) using Roy and Sussman’s Alternative Graph. Computational tests on various scales and real data from a manufacturing plant demonstrate the GA’s efficacy in solving complex scheduling problems in real-world production settings. Based on the data, the Proposed Genetic Algorithm (PGA), with an average Relative Deviation (ARD) of 0.25%, performed approximately 34% better compared to the Basic Genetic Algorithm (BGA), with an average ARD of 0.38%. This percentage indicates that the PGA significantly outperforms in solving complex scheduling problems.
Objective To assess the accuracy of three commercially available and one open‐source deep learning (DL) solutions for automatic tooth segmentation in cone beam computed tomography (CBCT) images of patients with multiple dental impactions. Materials and Methods Twenty patients (20 CBCT scans) were selected from a retrospective cohort of individuals with multiple dental impactions. For each CBCT scan, one reference segmentation and four DL segmentations of the maxillary and mandibular teeth were obtained. Reference segmentations were generated by experts using a semi‐automatic process. DL segmentations were automatically generated according to the manufacturer's instructions. Quantitative and qualitative evaluations of each DL segmentation were performed by comparing it with expert‐generated segmentation. The quantitative metrics used were Dice similarity coefficient (DSC) and the normalized surface distance (NSD). Results The patients had an average of 12 retained teeth, with 12 of them diagnosed with a rare disease. DSC values ranged from 88.5% ± 3.2% to 95.6% ± 1.2%, and NSD values ranged from 95.3% ± 2.7% to 97.4% ± 6.5%. The number of completely unsegmented teeth ranged from 1 (0.1%) to 41 (6.0%). Two solutions (Diagnocat and DentalSegmentator) outperformed the others across all tested parameters. Conclusion All the tested methods showed a mean NSD of approximately 95%, proving their overall efficiency for tooth segmentation. The accuracy of the methods varied among the four tested solutions owing to the presence of impacted teeth in our CBCT scans. DL solutions are evolving rapidly, and their future performance cannot be predicted based on our results.
This work introduces a novel methodology for identifying critical sensor locations and detecting defects in structural components. Initially, a hybrid method is proposed to determine optimal sensor placements by integrating results from both the discrete empirical interpolation method (DEIM) and the random permutation features importance technique (PI). Subsequently, the identified sensors are utilized in a novel defect detection approach, leveraging a semi-intrusive reduced order modeling and genetic search algorithm for fast and reliable defect detection. The proposed algorithm has successfully located defects with low error, especially when using hybrid sensors, which combine the most critical sensors identified through both PI and DEIM. This hybrid method identifies defects with the lowest errors compared to using either the PI or DEIM methods alone.
A new methodology for determining the representative strain in spherical indentation is presented in this paper. This methodology is based on the one defined by Hernot et al. (Mech. Mater. 68:1-14, 2014) for the case of Vickers indentation. It consists in calculating the gradients of a measured quantity according to the mechanical parameters of the behavior law of the tested material. The application of the proposed method is performed based on a new numerical approach to spherical indentation. From the representative strain and stress values obtained, it can be concluded that there is no constant universal constraint factor independent of the dimensionless ratio (a/R) and the properties of the indented material. In the case of an experimental application, these representative strain and stress values allow obtaining a strain hardening curve very close to the one obtained classically by a tensile test.
Background Few studies have assessed the participation of the spine in arm elevation. The primary aim of this exploratory study was to specify spinal movements during unilateral arm elevation. Methods We used an EOS imaging system to assess 2D global posture (Sagittal Vertical Axis [SVA], T1 and T9 tilt and Central Sacral Line [CSL]) and segmental spine curves (C3-C7 in the sagittal plane only, and T1-T6, T7-T12 and L1-L5 in the sagittal and frontal planes) for four different left arm elevation levels: in the sagittal (Sa) plane (30°Sa: reference position, 140°Sa and 180°Sa), and in the scapular (Sc) plane (180°Sc), in ten right-handed asymptomatic participants (5 women; mean age 24.6 SD 3.0 years]. In addition, we estimated C1, head and pelvic orientation and head and pelvic linear displacement. We used Bayesian statistics (BF10 > 3 indicates a significant variation: moderate, strong, very strong or extreme evidence). Results From 140°Sa to 180°Sa or Sc, the significant decrease in SVA and the T1-T9 tilt angles indicated a global backward spine bending (moderate to very strong evidence). The significant reversal of the C3-C7 lordosis at 30°Sa (-1.34 [2.53]°) to kyphosis at 180°Sa (13.88 [3.53]°, strong evidence) and 180°Sc (11.85 [2.75]°, extreme evidence) and the significant decrease in the T7-T12 kyphosis (26.58 [2.84]°at 30°Sa to 16.40 [2.65]° at 180°Sa and 17.60 [2.78]° at 180°Sc [all extreme evidence]) showed a global spine straightening. We found significant pelvic anteversion between 30°Sa and 140°Sa (moderate evidence) and persistent right spine bending and leftward head displacement (extreme evidence). The change in C1 orientation (extreme evidence) showed an atlanto-occipital extension. Conclusion Simple unconstrained movements of unilateral arm elevation involve the whole spine, pelvis and head, including significant backward spinal bending, a reduction in the low cervical spine lordosis and the thoracic kyphosis, and atlanto-occipital extension.
Background/Objectives: The handstand is an exercise performed in many sports, either for its own sake or as part of physical training. Unlike the upright bipedal standing posture, little is known about the sagittal alignment and balance of the spine during a handstand, which may hinder coaching and reduce the benefits of this exercise if not performed correctly. The purpose of this study was to quantify the sagittal alignment and balance of the spine during a handstand using radiographic images to characterize the strategies employed by the spino-pelvic complex during this posture. Methods: Nineteen national-level artistic gymnasts participated in this study and underwent a low-dose biplanar (frontal and lateral) radiograph in both upright bipedal standing posture and during a handstand. Then, 3D reconstruction of the spine, based on biplanar radiographic images, enabled the determination of key pelvic (pelvic incidence, sacral slope, pelvic tilt) and spinal (lumbar lordosis, thoracic kyphosis, T9 sagittal offset) parameters in both postures. Results: The results showed that most gymnasts performed pelvic retroversion during the handstand, which was accompanied by an average decrease in lumbar lordosis, thoracic kyphosis, and T9 sagittal offset. Additionally, lumbar curvature was found to depend on pelvic orientation in upright bipedal standing posture, whereas it was associated with the thoracic spine during the handstand. Conclusions: This study provides new insights into how the spine kinematically adapts to an inverted body load. The results may help coaches and physiotherapists in teaching the handstand or using it to rehabilitate and strengthen the spine through the handstand posture.
Background/objectives: The aim of this study was to evaluate changes in trunk height and variations in spino-pelvic parameters during trunk self-elongation. Two populations were studied: non-athletes and gymnasts, who differ in their engagement with core-strengthening exercises. Methods: EOS biplanar radiographs were taken on 14 non-athletes and 24 gymnasts in both neutral and trunk self-elongation positions. Three-dimensional reconstructions of the pelvis and spine were used to calculate effective trunk height, thoracic and lumbar contributions, and spino-pelvic parameters. Results: Trunk self-elongation resulted in a significant increase in trunk height for both groups (7 mm on average, range: -1 to 14 mm), accompanied by a reduction in thoracic kyphosis for all participants (-10° for non-athletes and -17° for gymnasts, on average) and a reduction in lumbar lordosis in most participants (-5° for non-athletes and -7° for gymnasts, on average). However, some individuals in both groups exhibited an increase in lumbar lordosis, which reduced the contribution of the lumbar region to overall trunk height. Conclusions: Trunk self-elongation instruction effectively increases trunk height, but additional instructions, such as pelvic retroversion, may enhance its effectiveness.
In the evolving landscape of engineering collaboration, Extended Reality (XR) demonstrates transformative potential for transdisciplinary work. XR approaches are promising in enhancing flexibility and efficiency in diverse, global work settings by facilitating remote and asynchronous collaboration. This paper introduces XR methods designed to address challenges such as communication gaps and project misalignment in asynchronous collaboration. Building upon our previous work, which introduced Avatar Replay for non-present users, this paper presents an extension that enables interaction with these users through their avatars. This development promotes a more dynamic and responsive approach to asynchronous collaboration. Additionally, the paper shares insights from a test scenario using a Head-Mounted Display (HMD) with an improved prototype in a transdisciplinary guidance scenario. The findings of this research highlight the significant potential of XR in enhancing collaboration and understanding across disciplines, particularly in asynchronous collaboration between experts and non- experts in the engineering domain. This enhanced interaction across realities not only promises increased societal acceptance of immersive technologies but also signifies a transformative shift in the way we work together, seamlessly blending virtual and real-life interactions.
We are considering the problem of observer design for heat diffusion-reaction partial differential equations (PDEs) that are monitored through delayed boundary sensor measurements. The novelty with respect to (w.r.t) existing observers lies in the fact the PDE is subject to parameter uncertainties coming, not only in the domain, but also in the sensor. We develop an adaptive observer that provides online accurate estimates of both the state and parameter estimates. The observer design makes use of modal decomposition theory and decoupling transformation method. The former is resorted to reformulate the observer design and analysis in a finite-dimensional space. The decoupling transformation makes the state estimation problem decoupled from the parameter estimation problem. Compared with existing adaptive observers, the new observer features: (i) the injection of extra actions in the domain allowing a better compensation for the effects of parameter uncertainties; (ii) a better consideration of all available information (at a given time) in the parameter estimator. The observer analysis is established using a Lyapunov- Krasovskii functional, under a persistent excitation (PE) assumption. The analysis highlights sufficient conditions, involving the maximal time delay, for the state and parameter estimation errors to be exponentially convergent to zero.
To enhance the water pre-cooling rate, postharvest quality, and prolong the shelf life of fruit and vegetables, static magnetic field-assisted water pre-cooling technology (SM-WPT) was proposed. The study investigated the effects of static magnetic field (0–100 Gs) on the cherry cooling rate during water pre-cooling process and evaluated the physicochemical properties during its 21-day storage (T = 4 ℃). Moreover, the analytical hierarchy process combined with entropy weight method (AHP-EWM) was utilized to analyze the optimal parameters of SM-WPT for preserving cherries. Results demonstrated that SM-WPT significantly prevented the deterioration of postharvest cherry quality during storage. The cherry for the SM40 group (40 Gs) reduced the cooling time by up to 40% compared to non-magnetic field treatment, alleviated changes in color difference and weight loss with 22.70% and 12.90%, respectively, and an obvious decrease in respiration rate and decay rate. Meanwhile, the SM80 group (80 Gs) was the best treatment for improving antioxidant enzyme activity and inhibiting malondialdehyde accumulation. The AHP-EWM comprehensive evaluation indicated that the SM40 group had the best overall preservation effect on cherries.
Products made with veneers such as Laminated Veneer Lumber can reach higher properties than solid wood because the defects such as knots (but not only) are distributed among the various layers. Visual sorting, even using automatic grading, is only partially efficient for evaluating mechanical properties, which mainly depends on the fiber orientation and the density both at local scale. An experimental protocol has been established to correlate the nondestructive (fiber orientation and density) estimation and the destructive (tensile test) measurement of beech veneer wood. The aim is to understand the impact of the small imperfections on clear wood in fiber orientation on the mechanical properties. In the present study, wood veneer is assumed to be a transverse isotropic material due to the predominant wood fiber direction in the growth direction of the tree. Experimental measurements of modulus of elasticity are based on Digital Image Correlation (DIC) and virtual extensometer. The Young modulus model is based on a composite material model that considers fiber orientation and density. The Young modulus model is used to determine longitudinal, transverse, and shear moduli, for specimens with angles ranging from 0 to 45°. Mechanical properties are obtained by mathematical minimization between experimental and model data. The coefficient of determination obtained was 0.97. The measure of the fiber angle with a resolution of 1 × 1 mm² and the tensile test with the DIC, both local, significantly improve Young modulus measurement compared to previous studies’ assessment accuracy and allow for a better understanding of the wood behavior.
In the wake of the prominence of language models such as ChatGPT/GPT4 and the emergence of various Natural Language Processing (NLP) approaches, there has been growing interest in their applications. However, a gap exists in scientific documentation regarding Small and Medium Enterprises (SMEs) within the industrial sector. This paper addresses this gap. This is the first systematic review of the literature associated with the context of NLP in industry. Through five research questions, it provides an overview of NLP applications, goals, technical solutions, obstacles, and applicability to SMEs, which is useful for both researchers and manufacturers. Following the PRISMA 2020 methodology, this study reveals a lack of literature addressing the use of NLP in industrial SMEs. The findings suggest that NLP is predominantly applied in specific industrial domains, including design, process monitoring, and maintenance. NLP applications mainly aim to enhance operational performance, notably in support functions like maintenance, safety, and continuous improvement. Practical implementations include automatic data analysis, similarity searches, information retrieval, and the conversion of raw text into standardized data. When looking at the technical solutions implemented, the paper demonstrates a strong diversity in the encountered algorithmic approaches. Challenges include remaining up-to-date, scaling, addressing low-quality or insufficient data issues, and navigating domain- or operator-specific vocabulary. In particular, maintaining up-to-date data presents a critical challenge for NLP applications but with limited identified solutions. Finally, the study indicates that only a fraction of the proposed NLP algorithmic solutions may apply to SMEs because of a lack of resources, expertise, and standardized procedures.
Background An innovative solution has been recently proposed for the treatment of heart failure with preserved ejection fraction (HFpEF), using a centrifugal mechanical circulatory support (MCS) device. We sought firstly to assess the hemocompatibility of the proposed device. HFpEF treatment requires the blood pump to operate at low blood flow rate (0.05–0.5 L/min). Given high blood trauma expected in these conditions, we sought secondly to investigate design improvements in order to reduce this risk. Method Computational fluid dynamics was used to analyze the blood fluid filled within the centrifugal pump and to estimate the blood trauma due to its operation. This assessment is based on the computation of integrated quantities such as the hemolysis index and the turbulent dissipation energy. Results The hemolysis index associated with the present device is comparable to the figures from HVAD. With the proposed pump, for instance, an index of 0.0036 is obtained at 1 L/min and 3000 rpm. Using thinner blades for the impeller allows a 12% reduction of the hemolysis index in average, while reducing its diameter leads to an index 2.8 folds lower at 0.5 L/min. Conclusion The present investigation shows the promising hemocompatibility of our centrifugal pump. Concerns about very high hemolysis generated at low flow rates could be overcome by reducing the impeller diameter. Experimental validations are planned to support our findings.
Background A hip flexion contracture (HFC) results in an inability to extend the hip by reducing the ROM of the affected hip. The condition affects one in four patients with above-knee amputations on the amputation side. While HFC in other disorders is known to decrease hip ROM and increase pelvic tilt during gait, its impact on the gait of patients with above-knee amputations remains unexplored. Typically, prosthetists design the socket with a flexion angle matching the HFC, potentially leading to compensations during the posterior stance phase of the gait cycle. To our knowledge, little is known about how or whether these compensations relate to the socket’s flexion alignment. Questions/purposes (1) Is the presence of HFC associated with modifications of spatiotemporal and kinematic parameters during gait in patients with an above-knee amputation? (2) Is there a correlation between the socket flexion angle and the spatiotemporal and kinematic parameters during gait in patients with an above-knee amputation with and without HFC? Methods A comparative observational study was conducted between February 2022 and June 2023. Thirty-two participants with unilateral above-knee amputations who had undergone amputation at least 1 year prior and had a minimum of 1 month of experience with their current prostheses were eligible for consideration and included in the study. After the trial, 1 of 32 participants was excluded due to other impairments affecting gait, and 9% (3 of 32) were excluded because of pain or discomfort during data acquisition on their gait, leaving 88% (28 of 32) of participants included in the analysis. The median (IQR) age of participants in the HFC group (n = 13) was 50 years (26 to 56); 85% (11) were male and 15% (2) were female. The median (IQR) age of participants in the noHFC group (n = 15) was 41 years (32 to 56), and 100% were male. Time since amputation was similar between groups (HFC median 8 years [IQR 3 to 21], noHFC median 6 years [IQR 1 to 9], difference of medians 2; p = 0.31). Thirty-two percent (9 of 28) of patients were classified according to the Medicare Functional Classification Level system as K4 (exceeding basic ambulation skills) and 68% (19 of 28) were classified as K3 (ability to walk with variable cadence and traverse most environmental barriers). Clinical and prosthetic measurements were made, which comprised measurement of the HFC using a hand-held goniometer with the patient in the modified Thomas test position, the socket flexion alignment, and the difference (δ) between the HFC and socket flexion alignment. A gait analysis was performed with an optoelectronic system equipped with six infrared cameras and two force plates to analyze the time-distance and kinematic parameters of gait. To answer our first question, we quantitively compared the gait spatiotemporal and kinematic parameters between groups, and for the second question, we evaluated the correlations between the same parameters and prosthesis alignment for both groups. Results During gait, the HFC group exhibited reduced mean ± SD residual hip ROM in comparison with the noHFC group (35° ± 6° versus 44° ± 6°, mean difference -9° [95% CI -13° to -6°]; p < 0.001), increased pelvic tilt (11° ± 6° versus 7° ± 3°, mean difference 4° [95% CI 1° to 8°]; p = 0.02), increased pelvic rotation (12° ± 3° versus 9° ± 2°, mean difference 3° [95% CI 2° to 6°]; p < 0.001), and increased trunk rotation (15° ± 5° and 12° ± 2°, mean difference 3° [95% CI 0° to 6°]; p = 0.04). Greater δ correlated with decreased ROM in the contralateral hip (r = -0.71; p = 0.006), pelvis (r = -0.77; p = 0.002), and trunk (r = -0.58; p = 0.04) in the sagittal plane and with increased residual hip ROM (r = 0.62; p = 0.02). In terms of spatiotemporal gait parameters, in the HFC group, the δ correlated with an increase in contralateral step width (r = 0.58; p = 0.04) and a decrease in prosthetic step length (r = -0.65; p = 0.02). Conclusion Our findings further suggest that physiotherapists should consider the pelvic and trunk compensations associated with HFC in their rehabilitation because of potential long-term effects, such as low back pain or osteoarthritis. In addition, the correlation between the socket flexion angle and the parameters involved may support prosthetists in their choices of prosthetic settings. For now, we cannot consider these compensations as an impaired gait syndrome, and future studies are needed to evaluate their impact on patients’ quality of life. Level of Evidence Level III, therapeutic study.
This research delves into the fracture analysis of cellular thermosetting polymers using Mode I fracture tests with a compact tension geometry subjected to both monotonic and cyclic loadings. The equivalent linear elastic fracture mechanics concept effectively described crack initiation and propagation. Numerical simulations estimated the equivalent linear elastic crack length aligning with experimental measurements. Resistance curves revealed a transient regime followed by a self‐similar regime with relatively constant fracture energy, typical of quasi‐brittle materials. Finally, the fracture energy evolution correlated with macroscopic density evolution exhibiting a linear relationship relaying the influence of microstructure to a second order.
Live-cell imaging generally requires pretreatment with fluorophores to either monitor cellular functions or the dynamics of intracellular processes and structures. We have recently introduced full-field optical coherence tomography for the label-free live-cell imaging of fungi with potential clinical applications for the diagnosis of invasive fungal mold infections. While both the spatial resolution and technical set up of this technology are more likely designed for the histopathological analysis of tissue biopsies, there is to our knowledge no previous work reporting the use of a light interference-based optical technique for direct mycological examination and monitoring of intracellular processes. We describe the first application of dynamic full-field optical transmission tomography (D-FF-OTT) to achieve both high-resolution and live-cell imaging of fungi. First, D-FF-OTT allowed for the precise examination and identification of several elementary structures within a selection of fungal species commonly known to be responsible for invasive fungal infections such as Candida albicans, Aspergillus fumigatus, or Rhizopus arrhizus. Furthermore, D-FF-OTT revealed the intracellular trafficking of organelles and vesicles related to metabolic processes of living fungi, thus opening new perspectives in fast fungal infection diagnostics.
Institution pages aggregate content on ResearchGate related to an institution. The members listed on this page have self-identified as being affiliated with this institution. Publications listed on this page were identified by our algorithms as relating to this institution. This page was not created or approved by the institution. If you represent an institution and have questions about these pages or wish to report inaccurate content, you can contact us here.
2,843 members
Pierre-Yves Rohan
  • Institut de Biomecanique Humaine Georges Charpak (IBHGC)
Christophe giraud-audine
  • Département de Fluides et Systèmes Energétiques ( FISE)
Carlos Junqueira Junior
  • Laboratoire de Dynamique des Fluides (Dynfluid)
Lahcen Bih
  • Laboratoire des Sciences et Métiers de l'Ingénieur
Sofiane KHELLADI
  • Centre de Paris
Information
Address
Paris, France
Head of institution
Prof. Laurent CHAMAGNEY