Archives ouvertes de Paris-Saclay
Not a member yet
272730 research outputs found
Sort by
Mid- and Far-Infrared Spectral Signatures of Mineral Dust from Low- to High-Latitude Regions: significance and implications
No full textInternational audienceMineral dust absorbs and scatters solar and infrared radiation, thereby affecting the radiance spectrum at the surface and top-of-atmosphere and the atmospheric heating rate. While half of the outgoing thermal radiation is emitted in the far infrared (FIR, 15–100 μm), knowledge of the optical properties and thermal radiative effects of dust is currently limited to the mid-infrared region (MIR, 3–15 μm). In this study we performed pellet spectroscopy measurements to evaluate the MIR and FIR contribution to dust absorbance and explore the variability and spectral diversity of the dust signature within the 2.5–25 μm range. Thirteen dust samples re-suspended from parent soils with contrasting mineralogy were investigated, including low and mid latitude dust (LMLD) sources in Africa, America, Asia, and Middle East, and high latitude dust (HLD) from Iceland. Results show that the absorbance of dust in the FIR up to 25 μm is comparable in intensity to that in the MIR. Also, spectrally different absorption (position and shape of the peaks) is observed for HLD compared to LMLD, due to differences in mineralogical composition. Corroborated with the few available literature data on absorption properties of natural dust and single minerals up to 100 μm wavelength, these data suggest the relevance of MIR and FIR interactions to the dust radiative effect for low to high latitude sources. Furthermore, the dust spectral signatures in the MIR and FIR could potentially be used to characterise the mineralogy and differentiate the origin of airborne particles based on infrared remote sensing observations
Remaining Useful Life Prediction for Aircraft Maintenance Using Machine Learning
No full textInternational audienceABSTRACT Ensuring regular equipment maintenance is critical for any business that relies on machinery. Predictive maintenance (PdM) is a strategy for scheduling maintenance tasks, with a primary focus on predicting the remaining useful life (RUL) of equipment in advance. This approach helps optimize maintenance schedules, reduce downtime, and detect unexpected faults. Predictions are based on analyzing data collected from the equipment, with machine learning (ML) facilitating these forecasts by training models on historical input data and corresponding outputs. The trained model can then estimate the RUL of the equipment before it reaches the end of its operational capacity. Various ML techniques have been employed for the accurate estimation of the RUL. In this paper, we aim to identify the most effective ML regression methods for PdM and RUL prediction for an auxiliary power unit (APU), focusing on performance indicators such as the root mean squared error (RMSE), the mean absolute error (MAE), and the correlation coefficient ( R ). The process begins with a dataset, followed by feature selection methods such as random forests and normalization during the preprocessing stage. Then, the ML models are trained and evaluated. To assess the effectiveness of the proposed approach, data from the NASA Ames Research Center, along with on‐wing sensor data from the Shenyang Maintenance Base of China Southern Airlines (SYMOB), are used. Six ML algorithms and a hybrid model are employed: Support Vector Machines (SVM), long short‐term memory (LSTM), gated recurrent unit (GRU), decision tree (DT), K‐nearest neighbors (KNNs), gradient boosting trees (GBTs), and a hybrid model (GBT + LSTM). The results for the regression techniques, based on the RMSE and R, are as follows: SVM (37.62, 0.84), LSTM (20.22, 0.91), GRU (31.29, 0.87), DT (17.89, 0.94), KNN (10.98, 0.98), GBT (22.62, 0.97), and (GBT + LSTM) (24.32, 0.96). The KNN method is the most effective approach for this study, as it demonstrates the lowest RMSE and the highest correlation coefficient ( R ) compared to other methods. Therefore, we highly recommend utilizing the KNN technique for PdM analysis of APUs
Understanding Charge Radii with Machine Learning: Discovering Physics Expressions
No full textInternational audienceWe introduce a robust, interpretable machine learning (ML) framework that combines numerical regression for high-accuracy predictions with symbolic regression to uncover the underlying physics. This hybrid approach effciently derives analytical expressions by leveraging the smoothed predictions of optimized ML models, a significant acceleration over direct symbolic regression on raw experimental data. We apply this framework, as an example, to nuclear charge radii across the nuclear chart, notably including light nuclei that are often excluded from such studies. We employ Light Gradient Boosting Machine (LGBM) and Gaussian Process Regression (GPR) models to map correlations between charge radii and key physical features: mass and proton number dependencies, total binding energy, and for the first time, the pairing gap. Our models are rigorously trained using four-fold cross-validation with automated hyperparameter optimization, ensuring robustness and generalizability, which is critical for the typically small and skewed datasets in nuclear physics. Finally, we distill the knowledge from the initial LGBM and GPR models into simplified, interpretable mathematical expressions via symbolic regression, white-boxing these ML models. The derived formulas provide physical insights comparable to traditional many-body models and demonstrate a powerful pathway for physics expression discovery guided by ML
Formal series for linear representations of timed colored Petri nets
No full textInternational audienc
Long-duration in situ monitoring of H2O and CH4 in the equatorial tropopause with the Pico-STRAT Bi Gaz balloon borne laser diode spectrometer during the Strateole 2 campaign
No full textInternational audienceThe Pico-STRAT Bi Gaz spectrometer provides in situ mixing ratio measurements of H2O and CH4 (or CO2 ) under balloon. The instrument was flown in the tropical upper troposphere and lower stratosphere in 2019/2020 and 2021/2022 during the Strateole 2 campaigns for a total of five flights of 20 to 80 days between 18 and 20 km altitude. In this frame, in situ measurements of water vapor and methane were performed every 4 to 12 minutes in the equatorial tropopause layer. On several occasions, water vapor measurements of Pico-STRAT Bi Gaz have been compared with localized measurements from the FLASH-B Lyman-α hygrometer and vertical profiles of the NOAA Global Monitoring Laboratory (GML) frost point hygrometer over Hilo, Hawaii. Pico-STRAT Bi Gaz measurements agreed with the FLASH-B hygrometer to within 2.2 ± 5.3 % between 18.2 and 18.7 km in 2021 and to within 1.3 ± 5.3 % near 19 km in December 2019. Pico-STRAT Bi Gaz agreed with NOAA’s FPH hygrometer to within 1.2 ± 4.1% between 18 and 19 km on four occasions during the two campaigns. These are within both instruments’ uncertainties. Methane measurements from Pico-STRAT Bi Gaz have been compared with in situ measurements from the Whole Air Sampler instrument (WAS), flown aboard the NASA WB-57 aircraft during the ACCLIP 2022 campaign over South Korea, eight months after the Pico-STRAT Bi Gaz overpass. The relative difference between both instruments is found to be of (−0.1 ± 0.9) % within the altitude range from 17 to 19 km and within the Pico-STRAT measurement uncertainty
Reduced order modelling for shell finite element structures using the direct parametrisation of invariant manifolds: Hardening/softening transition, resonant dynamics and mode selection
No full textInternational audienceThe accurate simulation of the nonlinear dynamics of thin-walled structures is a critical but computationally demanding task. In this contribution, a 7-parameter solid-shell finite element formulation is combined with the direct parametrisation method for invariant manifolds (DPIM), in order to derive accurate and efficient reduced-order models (ROM) accounting for geometric nonlinearity. The method is illustrated in its ability to be used with different yet complementary purposes. On the one hand, low-order tractable models can be obtained, providing simple ROMs that are amenable to giving physical insights and understanding. On the other hand, higher-order solutions are available within the same framework, hence providing accurate and converged solutions. This scheme is carried out on examples with increasing complexity. First, the transition from hardening to softening behaviour for thin shells with shape imperfections is investigated. The 1:2 resonance as a driver of the change of type of nonlinearity is analysed, and a full understanding of the smooth transition is illustrated. Then, shells with varying thicknesses are investigated, and the case of 1:2 internal resonance is further investigated, showing the emergence of isolated solution branches (isola). In the course of the numerical simulations, it is shown how the reduced basis needs to be enlarged to take into account more and more complex resonance scenarios, and some guidelines are provided in order to help the analyst in selecting the master modes. The numerical results highlight the ability of the reduced-order models to provide a fully comprehensive and integrated framework for the understanding and accurate prediction of thin shells' nonlinear dynamics
Using Ghost Ownership to Verify Union-Find and Persistent Arrays in Rust
No full textInternational audienceThe type system of Rust enforces the "shared xor mutable" principle, which forbids mutation of shared memory. This principle eases verification in Rust, but certain programs require circumventing it with the mechanism of interior mutability. Thus, supporting interior mutability in a deductive verification tool is difficult. The Verus tool demonstrated the use of ghost resources to that end. So far, this mechanism has only been applied to Verus in order to verify primarily system code.We extend the deductive verification tool Creusot with support for linear ghost resources. We show how Creusot's full support for mutable borrows enables better specifications for primitives of linear ghost code. We apply this methodology to the verification of two data structures using sharing and mutation: union-find and persistent arrays
Levosimendan to Facilitate Weaning From ECMO in Patients With Severe Cardiogenic Shock
No full textInternational audienceImportance: Levosimendan may facilitate weaning from venoarterial extracorporeal membrane oxygenation (VA-ECMO) and improve survival, but supporting evidence remains limited.Objective: To assess whether early administration of levosimendan reduces the time to successful VA-ECMO weaning in patients with severe but potentially reversible cardiogenic shock.Design, setting, and participants: Randomized, double-blind, placebo-controlled trial conducted across 11 intensive care units (ICUs) in France. Between August 27, 2021, and September 10, 2024, 205 adult patients with acute cardiogenic shock who had started VA-ECMO in the preceding 48 hours were enrolled. Final follow-up was completed on November 10, 2024.Interventions: Patients were randomized in a 1:1 ratio to receive levosimendan, 0.15 μg/kg per minute, to be increased to 0.20 μg/kg per minute after 2 hours (n = 101), or placebo (n = 104).Main outcomes and measures: The primary outcome was time to successful ECMO weaning within 30 days following randomization. Secondary outcomes included ECMO-, mechanical ventilation-, and organ failure-free days, ICU and hospital lengths of stay, serious adverse events, and all-cause 30- and 60-day mortality.Results: Among the 205 randomized patients (median age, 58 [IQR, 50-67] years; 149 [72.7%] male), main cardiogenic shock etiologies were postcardiotomy (79 [38.5%]), acute myocardial infarction (56 [27.3%]), and myocarditis (28 [13.7%]). Treatment dose was increased to 0.20 ± 0.01 μg/kg per minute in 93% of patients receiving levosimendan and in 96% of those receiving placebo. Within 30 days, 69 of 101 patients (68.3%) had a successful ECMO weaning in the levosimendan group compared with 71 of 104 (68.3%) in the placebo group (risk difference, 0.0% [95% CI, -12.8% to 12.7%]; subdistribution hazard ratio, 1.02 [95% CI, 0.74-1.39]; P = .92). In the levosimendan and placebo groups, respectively, median ECMO duration (5 [IQR, 4-7] days vs 6 [IQR, 4-11] days; P = .53), mean ICU length of stay (18 [SD, 15] days vs 19 [SD, 15] days; P = .42), and 60-day mortality (27.7% vs 25.0%; risk difference, 2.7% [95% CI, -9.0% to 15.3%]; P = .78) did not differ significantly. Ventricular arrhythmias occurred more frequently with levosimendan (18 [17.8%] vs 9 [8.7%]; absolute risk difference, 9.2% [95% CI, 0.4%-18.1%]).Conclusions and relevance: Among patients with severe but potentially reversible cardiogenic shock supported by VA-ECMO, early levosimendan administration did not significantly reduce the time to successful weaning of ECMO compared with placebo.Trial registration: ClinicalTrials.gov Identifier: NCT04728932
Recommendations for Human Sperm Morphology Assessment in 2025: An Expert Review From the French BLEFCO Group
No full textInternational audienceNumerous publications have questioned the lack of analytical reliability and clinical relevance of sperm morphology assessment for infertility workup and before use of assisted reproductive techniques (ART). There is a huge variability in the performance and interpretation of this test. It has become necessary to evaluate its true medical service rendered to the patient. Objectives To develop clinical guidelines for use of spermatozoa morphology assessment during male fertility check‐up and before ART. Materials and Methods These guidelines were produced following a pre‐defined standard methodology for narrative and Patient Intervention Comparison Outcomes (PICO) questions. The French Working Group (WG) on Sperm Morphology Assessment consisted of 15 members including an expert in statistics. Results R1: WG does not recommend systematic detailed analysis of abnormalities (or groups of abnormalities) during sperm morphology assessment. R2: WG recommends that the laboratory should use a qualitative or quantitative method for detection of a monomorphic abnormality (globozoospermia, macrocephalic spermatozoa syndrome, pinhead spermatozoa syndrome, multiple flagellar abnormalities). The result may be given as an interpretative commentary or as a numerical report of the percentage of detailed abnormalities. R3: There is insufficient evidence to demonstrate the clinical value of indexes of multiple sperm defects (TZI, SDI, MAI) in investigation of infertility and before ART. Accordingly, the working group does not recommend the use of sperm abnormality indexes (TZI, SDI, MAI) in sperm morphology assessment. R4: WG gives a positive opinion on the use of automated systems based on cytological analysis after staining after qualification of the operators, and validation of the analytical performance within their own laboratory. R5: WG does not recommend using the percentage of spermatozoa with normal morphology as a prognostic criterion before IUI, IVF, or ICSI, or as a tool for selecting the ART procedure. Discussion This article examines the clinical interest of sperm morphology assessment during fertility check‐up and before ART. The overall level of evidence from studies is low, challenging current practices regarding sperm morphology assessment. Conclusion These guidelines suggest a significant simplification of sperm morphology assessment in the light of the examined publications while maintaining the detection of monomorphic sperm abnormalities
An extensible Julia toolkit for symmetry-aware dual-space phasing in arbitrary dimensions
No full textInternational audienceWe present an open-source Julia-based software toolkit for solving the phase problem using dual-space iterative algorithms. The toolkit is specifically designed for aperiodic crystals and quasicrystals, supporting general space-group symmetries in arbitrary dimensions. A key feature is the symmetry-breaking anti-aliasing sampling scheme, optimized for computational efficiency when working with strongly anisotropic diffraction data, common for quasicrystals. This scheme avoids sampling redundancy caused by symmetry constraints, imposed during phasing iterations. The toolkit includes a reference implementation of the charge-flipping algorithm and also allows users to implement custom phasing algorithms with fine-grained control over the iterative process