University of Technology of Belfort-Montbéliard

Extracting maximum power from photovoltaic systems (PV) is challenging due to the nonlinear relationship between voltage and current. This paper proposes a novel maximum power point tracking (MPPT) controller that leverages the combined strengths of artificial neural networks (ANNs) and robust terminal sliding mode control (RTSMC). An ANN trained using an adaptive particle swarm optimization (APSO) algorithm predicts the maximum power point (MPP) of the PV system. PSO optimizes the ANN’s structure and initial weights, leading to faster training times and improved prediction accuracy under varying environmental conditions. Subsequently, a TRSMC method ensures the system precisely and rapidly tracks the predicted MPP. TRSMC’s inherent robustness guarantees effective operation even under uncertain conditions like sudden changes in sunlight or temperature. This combined approach offers a high-accuracy, fast-response, and robust MPPT controller, maximizing power extraction from PV systems. The effectiveness of the proposed method is validated through simulations, demonstrating superior performance compared to other algorithms such as perturb and observe (P&O) and a P&O-based RTSMC controller (P&O-RTSMC).

Mechanical structures often operate under dynamic loads and complex kinematic conditions, but current design methods primarily adopt static or simplified boundary conditions, leading to inaccuracy in defining the design space and the resulting solutions. To address these limitations, this paper presents a novel generative design framework that integrates kinematic and dynamic factors to establish more precise design inputs. To accurately capture the real-world mechanical states of structures, extensive kinematic and dynamic data are extracted through virtual prototyping technology. A novel design domain determination method is proposed based on kinematic relationships, while a load condition extraction approach is formulated based on material deformation theory, collectively ensuring a more realistic representation of design inputs in generative design. Furthermore, a design for additive manufacturing strategy is incorporated to refine and optimise the generative design outcomes, enhancing the structural manufacturability. The proposed method is applied to a challenging case study: designing a hydraulically actuated quadruped robot leg with complex kinematic relationships and varying dynamic load conditions. Experiment results indicate that the proposed method greatly improves structural efficiency and performance, demonstrating its potential in designing complex mechanical structures.

Profile stretch-bending forming is a complex process influenced by multiple factors, including processing technology, geometric parameters, and material properties. These factors make it challenging to predict and optimize the forming process, especially controlling springback, which significantly affects the final quality. This study investigates the impact of the stretch-bending process on the quality of rocket frame rings by combining both simulation and production data. A Co-Kriging model is developed to approximate the frame ring forming process, and a multi-objective optimization model is applied to minimize springback, deformation, and residual stress in the rings. The optimization is performed using the non-dominated sorting genetic algorithm with a projected target area (NSGA-II/A) to identify the optimal process parameters. The results show that the optimized parameters significantly improve the forming quality and help achieve higher precision in the production of rocket frame rings. This work provides valuable insights for enhancing the stretch-bending forming process and guiding the production of high-quality components.

Electric Vehicles (EV) significantly contribute to reducing carbon emissions and promoting sustainable transportation. Among EV technologies, hybrid energy storage systems (HESS), which combine fuel cells, power batteries, and supercapacitors, have been widely adopted to enhance energy density, power density, and system efficiency. Bidirectional DC-DC converters are pivotal in HESS, enabling efficient energy management, voltage matching, and bidirectional energy flow between storage devices and vehicle systems. This paper provides a comprehensive review of bidirectional DC-DC converter topologies for EV applications, which focuses on both non-isolated and isolated designs. Non-isolated topologies, such as Buck-Boost, Ćuk, and interleaved converters, are featured for their simplicity, efficiency, and compactness. Isolated topologies, such as dual active bridge (DAB) and push-pull converters, are featured for their high voltage gain and electrical isolation. An evaluation framework is proposed, incorporating key performance metrics such as voltage stress, current stress, power density, and switching frequency. The results highlight the strengths and limitations of various converter topologies, offering insights into their optimization for EV applications. Future research directions include integrating wide-bandgap devices, advanced control strategies, and novel topologies to address challenges such as wide voltage gain, high efficiency, and compact design. This work underscores the critical role of bidirectional DC-DC converters in advancing energy-efficient and sustainable EV technologies.

Despite the growing importance of human-centered design and ergonomics in various fields, a significant gap exists in applying these principles to robotic systems in airport environments. This paper focuses on a real-time baggage handling monitoring system by proposing a computational ergo-design approach. It presents the optimal system architecture for real-time baggage handling. The proposed architecture, called ARTEMIS (ARchitecture for real-TimE baggage handling and MonitorIng System), is designed for real-time baggage handling and monitoring. The circuit modeling is carried out using a directed graph. Five strategies are simulated to test their effectiveness and evaluate their performance within the system. A simulation that generates key indicators enables preliminary visualization and analysis of AGV behavior through predefined scenarios. These results are presented through an intuitive and ergonomic user interface, designed with a focus on user–computer interaction as a problem-solving process centered on the user’s experience. The results show that, if the goal is to balance energy efficiency with effective baggage handling, the Mixed Advance/Delay Strategy appears to be the best overall choice, as it optimizes both energy consumption and baggage handling while maintaining relatively low waiting times. However, if minimizing queue time and maximizing baggage collection are the highest priorities (with less emphasis on energy efficiency), the Turnstile Strategy remains a solid option. In addition, the simulations show that the operator plays a central role in minimizing delays and ensuring the smooth operation of the system. Both local and global system failures depend heavily on the operator’s response time, decision-making, and overall efficiency. Therefore, operator efficiency and a well-designed support system are critical to maintaining a smooth and effective baggage handling process.

The ferromagnetic material FeSi6.5 exhibits lower magnetic losses compared to the more widely used FeSi3, commonly employed in the production of laminated sheets for stators and rotors. This makes FeSi6.5 a promising candidate for electric motor manufacturing. However, its high silicon content complicates significantly production through conventional lamination processes. Additive manufacturing emerges as a viable alternative for this application. In electric motors, laminated sheets are separated by insulating layers to reduce eddy current losses. While additive manufacturing allows the fabrication of complex shapes, it is currently limited to single-material structures and cannot replicate the laminated architecture of electric motors. Research on multi-material additive manufacturing remains in its early stages of development. The challenge is to manufacture laminated electric motors using additive manufacturing, with thin ceramic insulating layers and ferromagnetic FeSi6.5 material layers. To achieve this, we have developed an innovative system to modify our LPBF manufacturing machine to work with two materials. This paper presents our work on the modification of our LPBF machine for the bi-material process, as well the ceramic/metal interface optimization.

The Gurney flap (GF) is a simple flat plate frequently mounted at the airfoil rear. Several investigations have been devoted to studying the effect of a rigid or even movable GF on the aerodynamic behavior of several devices such as flapping airfoils and vertical or horizontal axis turbines. The present paper proposes a new concept of a deformable Gurney flap (DGF) to improve the output power of a flapping airfoil in vertical mode. The advantage of this model is the full control of the effect on the GF during the flapping movement. The DGF is expandable and contractible which allows for monitoring and adjusting the pressure distribution at the appropriate time and position. By using a 2D transient simulation with a specific dynamic mesh design, an extended numerical analysis has been provided. It was found that this model is able to increase the output power by 19.5%. Furthermore, the concept of the DGF is applied on flapping turbines in hybrid modes such as swing arm mode and D-shaped mode. These modes are investigated to clarify the studied model’s advantage and to demonstrate the possibility of applying this strategy to control the different flapping movements.

Gadolinium-doped ceria (GDC) is a promising electrolyte for metal-supported solid oxide fuel cells (MS-SOFCs) due to its high ionic conductivity at intermediate temperatures (500-700°C). However, the extremely high sintering and densification temperature required for GDC electrolytes limits their practical application. Low-pressure plasma spraying (LPPS) offers a potential solution by enabling the preparation of dense GDC coatings without the need for sintering. This study explores the relationship between spraying distance, coating microstructure, and cell performance and elucidates the deposition mechanism and performance degradation of GDC-based MS-SOFCs. Results show that spraying distance significantly affects coating microstructure and cell performance. At shorter distances, perpendicular cracks dominate, while parallel cracks increase with distance. Increasing the spraying distance from 250 to 300 mm reduces the cell open circuit voltage (OCV) from 0.896 to 0.876 V and the maximum power density (MPD) from 176 to 135 mW/cm 2 at 600°C. After 100 h of degradation testing for 275 mm-cell, the OCV decreases from 0.885 to 0.86 V and MPD drops from 151 to 82 mW/cm 2. However, impedance spectroscopy analysis reveals that only 5% of the reduction in power density is attributed to the GDC electrolyte. These findings highlight the potential of LPPS as a viable technique for preparing dense GDC electrolyte for MS-SOFC.

In this study, we introduce an innovative approach for addressing fractional partial differential equations (fPDEs) by combining Monte Carlo-based physics-informed neural networks (PINNs) with the cuckoo search (CS) optimization algorithm, termed PINN-CS. There is a further enhancement in the application of quasi-Monte Carlo assessment that comes with high efficiency and computational solutions to estimates of fractional derivatives. By employing structured sampling nodes comparable to techniques used in finite difference approaches on staggered or irregular grids, the proposed PINN-CS minimizes storage and computation costs while maintaining high precision in estimating solutions. This is supported by numerous numerical simulations to analyze various high-dimensional phenomena in various environments, comprising two-dimensional space-fractional Poisson equations, two-dimensional time-space fractional diffusion equations, and three-dimensional fractional Bloch–Torrey equations. The results demonstrate that PINN-CS achieves superior numerical accuracy and computational efficiency compared to traditional fPINN and Monte Carlo fPINN methods. Furthermore, the extended use to problem areas with irregular geometries and difficult-to-define boundary conditions makes the method immensely practical. This research thus lays a foundation for more adaptive and accurate use of hybrid techniques in the development of the fractional differential equations and in computing science and engineering.

Diagnosis of overcharging in lithium-ion batteries (LIBs) is crucial to guaranteeing the long-term thermal stability and operational lifespan of a battery system. Compared with conventional diagnosis methods that rely on cell temperature and voltage measurements, the dynamic impedance spectrum (DIS) provides novel insights into assessing battery charging and overcharging processes. In this work, the perturbation signals are superimposed onto the charging current for the real-time monitoring of the dynamic impedance variation. A quantitative analysis is conducted to examine the applicability of the characteristic parameters, which are extracted by fitting the DIS with an equivalent circuit model (ECM), in assessing the overcharging in LIBs. Experimental investigations confirm the validity of the DIS measurements by performing the Kramers-Kronig (K-K) tests, where the maximum absolute residuals are lower than 0.5%. The proposed method is capable of reliably warning the battery overcharge when the batteries are charged up to 98% state of charge (SOC).

Bismuth-doped antimony tungstate (Bi-doped Sb2WO6) microspheres were synthesized via a novel hydrothermal synthesis approach. These microspheres were then used as active layers in gas sensors for the detection of carbon dioxide (CO2), a significant greenhouse gas and a critical parameter for evaluating air quality. The incorporation of bismuth significantly enhances the gas-sensing performance of the Sb2WO6 microspheres, with the 4% Bi-doped sensing active layer achieving a remarkable response value of 15 when exposed to 200 ppm of CO2, outperforming the undoped Sb2WO6. Furthermore, the selectivity of the 4%Bi-Sb2WO6 sensor toward CO2 gas was enhanced relative to the Sb2WO6 sensor. The fundamental mechanisms of gas sensing and the factors contributing to the improved CO2 response of 4%Bi-Sb2WO6 microspheres were investigated using density functional theory. Bi-doped Sb2WO6 materials exhibit significant advantages in gas-sensing applications, including improved conductivity, enhanced gas adsorption capacity, increased reaction rates, good chemical stability, excellent selectivity, and the ability to adjust electron density. These characteristics enable Bi-doped Sb2WO6 to demonstrate higher sensitivity and rapid response capabilities in gas sensors, making it suitable for practical applications.

This study examines the effects of internal governance mechanisms on the issuance of green bonds and investigates whether firms issuing green bonds exhibit distinct corporate governance characteristics, especially regarding board gender diversity and corporate social responsibility (CSR) committees. The analysis is based on a sample of 64 green bond announcements between 2013 and 2022. Based on the Generalized Least Squares Regression model, empirical results show that the presence of a CSR committee is positively and significantly associated with the issuance of green bonds. In other words, companies with a CSR committee are more likely to issue green bonds. In addition, companies with a lower debt ratio are more likely to issue green bonds.

Dynamic electrochemical impedance spectroscopy (DEIS) provides critical insights into the kinetic pathways of LIBs under dynamic operating conditions, establishing its significance in advanced onboard diagnostics. However, accurate DEIS implementation faces challenges due to voltage drifts arising from load fluctuations, state of charge (SOC) variations, and temperature changes. To address these issues, this work proposes a novel differencing framework that systematically suppresses drifting components while preserving the integrity of perturbation and response signals, enabling precise in-situ frequency response analysis. Analytical criteria are established to select differencing lags and orders, effectively mitigating disturbances without compromising frequency-domain responses under dynamic driving conditions. Comprehensive experimental investigations across diverse SOC levels, driving profiles, thermal environments, and battery aging states validate the method, achieving consistently accurate and stable DEIS measurements with maximum absolute residuals below 0.6% under Kramers-Kronig (K-K) validation.

Objective This study investigates the relationship between motion sickness and body movements experienced by car passengers during non-driving related activities. Background The theory linking motion sickness to postural instability is well-documented in static environments. However, evidence supporting this theory in dynamic environments, such as moving vehicles, is still lacking. Method Using an experimental approach replicating a naturalistic 15-min car ride, 56 participants were equipped with an in-ear sensor to measure the linear accelerations of the head. Participants reported their motion sickness severity at 3-min intervals during the experiment and once more post experimentation. Additionally, the UniPG numerical model was used to estimate motion sickness severity. Results The study identified significant relationships between specific head movement patterns and motion sickness severity, even though the overall symptoms reported were mild. Nonlinear interactions were identified between the standard deviation ( p = .032) and the skewness ( p = .028) of longitudinal head acceleration, as well as for the skewness ( p = .004) and kurtosis ( p = .008) of lateral head acceleration. Predictions from the UniPG model correlated with subjective ratings for 67% of participants with some motion sickness symptoms. Conclusion Highly variable longitudinal movements appear more tolerable when lateral movements remain symmetric; however, when both variability and asymmetry in head movements are present together in a specific pattern, they may exacerbate motion sickness symptoms. Application Incorporating motion sickness prediction models in vehicles, based on the measurement of head movements, might improve detection of the escalation of symptoms in car passengers.

Self-reconfiguration of modular robots is one of the most challenging problems in the robotics field. The objective is to determine how a set of identical modular robots, with local knowledge of the system and limited capacities, can reorganize themselves into a target topology or shape. The problem has received great interest from the research community giving birth to many centralized and distributed algorithms. However, the lack of comparative study of these algorithms makes it difficult to choose one when faced with a given configuration. In this paper, we present a kind of high-level hybridization approach of these algorithms using a neural network technique. The objective is to propose a centralized pre-processing procedure that allows, according to the self-reconfiguration problem, to determine which algorithm is most suitable. We applied the Neural Network technique to two self-reconfiguration algorithms: C2SR and TBSR. The obtained results show that the machine learning tool succeeds 96.67% of the time to determine the suitable algorithm based on the initial and the final shape. Consequently, using machine learning directly leads to the reduction of the required number of moves for the reconfiguration.

In recent years, the fusion of deep learning and computer vision technologies has significantly advanced the development of autonomous vehicles that are present more and more in road traffic. In this context, this paper proposes a vehicle vision system that combines two techniques, the first uses artificial intelligence algorithm to accurately identify vehicles in the path of vehicle’s trajectory, the second uses stereovision algorithm to precisely estimate inter-vehicle distances. This solution effectively reduces overall processing time by exploiting the advantages of the You Only Look Once real-time vehicle detection and limiting the region of interest in image to the computation area of the disparity map for the stereovision. Detection and distance estimation of numerous vehicles consumes an important computation time; therefore a parallel data processing based on the Open Multi-Processing library is used to optimize data processing performance. The proposed solution is implemented on an embedded platform, the experiment results show that the system successfully detects vehicle and estimate distance with an error rate of less than 10%, achieving a real-time processing of 30 frames per second.

Electrochemical impedance spectroscopy (EIS) offers valuable insights into battery state monitoring and failure diagnosis; however, the impedance measurements are constrained by the high implementation costs. Within this context, the fast measurements of the battery impedance spectrum during a brief relaxation process are investigated in this work. The sources of impedance measurements' inaccuracies are first analyzed, revealing that the undependable impedance responses are mostly attributed to voltage hysteresis effects. A zero-lag voltage hysteresis cancellation (ZVHC), integrated with a gradient descent algorithm, is then proposed for the extraction of voltage hysteresis components intertwined with the perturbation responses while eliminating the data latency. Experimental studies indicate that the proposed method is capable of accurately capturing the battery impedance responses ranging from 1.5 kHz to 0.1 Hz within 20 seconds. The requested resting period is less than 0.3% that of the conventional method. The impedance measurement results consistently match the reference trajectory under diverse operating conditions, confirming the high efficiency and validity of the proposed method.

As global energy transition and electrified transportation continue to advance, Traction Power Supply Systems (TPSS) are becoming increasingly vital in rail transit. To solve problems that traditional planning for TPSS often overlooks the integrated energy access and the collaborative optimization of economic-security throughout the life cycle, a Life-cycle Economic-Security Collaborative Planning (LE-SCP) model for TPSS with Integrated Energy Access (IEA-TPSS) is proposed. This model integrates an Equipment Performance Correction Matrix (EPCM) to address the performance degradation over the lifecycle of equipment. Additionally, it considers N -1 security to ensure the resilience of IEA-TPSS under contingent conditions. Further, a Scenario-based Hierarchical Decomposition (SHD) method is introduced to solve this model. A case study analysis using actual data was conducted to validate the LE-SCP model and SHD method. Simulation results show that the LE-SCP model and SHD method effectively address the comprehensive lifecycle planning issues of IEA-TPSS, considering both equipment degradation and N -1 security. Moreover, compared to traditional solution algorithms, the proposed SHD method can effectively reduce the model solving time by 8.61%.