Due to the variation of system parameters in airborne inerting systems, pressure fluctuations can lead to spectral line broadening and distortion of harmonic responses, thereby degrading the inversion accuracy and robustness of conventional wavelength modulation spectroscopy (WMS) methods. To address this issue, this study proposes an oxygen concentration prediction method that integrates a convolutional autoencoder with transfer learning (CAE-TL). Building upon this framework, a dynamic bottleneck mechanism is introduced to enhance multi-scale harmonic feature representation and improve prediction accuracy under pressure-varying operating conditions. By constructing a cross-domain learning framework between simulated and experimental datasets, effective transfer of feature priors is achieved, thereby improving model generalization under limited experimental samples. Oxygen concentration measurements were conducted under variable pressure conditions at 20 °C and pressures ranging from 0.1 to 1.0 bar, and the results demonstrate that the proposed sensing system achieves a detection accuracy of approximately 2%. Regression analysis shows that the coefficients of determination (R2) for CAE-TL, CAE, and WMS are 0.9805, 0.9752, and 0.8662, respectively, indicating that the proposed method outperforms conventional approaches in both accuracy and robustness. These results suggest that the proposed method provides an effective solution for stable oxygen concentration detection under pressure-varying conditions relevant to airborne inerting systems.

Accurate hydrogen leakage detection is critical for the safety of high-pressure systems. Background-Oriented Schlieren (BOS) is a powerful visualisation tool, but its quantitative accuracy, particularly the precision of displacement field extraction, is severely limited by the viewing distance, compromising reliability in far-field measurements. To overcome this fundamental limitation, this study proposes a novel dual-channel MGR-Net framework that synergistically integrates near-field and far-field data. By incorporating transfer learning to accelerate convergence and enhance feature extraction, and employing D-S evidence theory for decision-level fusion of the dual-channel outputs, our method achieves a state-of-the-art diagnostic accuracy of 98.94 %. This represents a significant improvement over traditional BOS techniques, offering a robust and intelligent solution for precise hydrogen leakage identification.

Reusable launch vehicles (RLVs) subject avionics electronics to repeated vibration, shock, and thermal cycling, yet no standardized method exists to evaluate reusability at the device level. We integrate physics-of-failure models—Steinberg vibration fatigue, Engelmaier/Coffin–Manson solder thermal fatigue, and Miner cumulative damage—with AHP-entropy multi-criteria decision making (MCDM). A flight-heritage 6-layer FR4 printed circuit board (PCB) from a rocket data-acquisition unit (108.25 × 108.25 × 1.62 mm, 109 components) is analyzed under Falcon 9 vibration and a DLR re-entry thermal profile (f1 = 310.2 Hz), with 11 core devices extracted from ODB++. Thermal fatigue dominates vibration damage by seven orders of magnitude. The board-level average of 62.9 flights is misleading: the ceramic PGA device D10 limits the unit to 12.3 flights (a preliminary model-based estimate, pending ALT calibration), a factor-of-five discrepancy, with three ceramic families (D10 PGA, D9 LCC, D6–D8 FIFO) forming the bottleneck. The combined weighting assigns 79.5% of the decision to the damage-derived criterion; Comprehensive Reusability Index (CRI) thresholds map flights 0–3 to direct reuse, 4–7 to refurbishment, and 8+ to retirement. The results distill into a weakest-link reuse principle, a damage-threshold service model, and an inverse-square thermal-fatigue relation.

In aerospace missions, the proper functioning of the environmental control and life support system (ECLSS) is critical to ensuring mission success. As a key gas detection device in the ECLSS, the performance of the rhenium–tungsten (Re–W) filament in the ion source of quadrupole mass spectrometers directly affects the stability and data accuracy of the instrument. However, tungsten alloy filaments may still suffer damage due to thermal stress, electromigration, mild oxidation, and other factors during operation. Therefore, investigating the damage mechanisms of Re–W filaments is of substantial practical importance. This study aims to explore the failure mechanisms of Re–W filaments and provide a basis for performance improvement in practical applications. The paper systematically introduces the working principle of quadrupole mass spectrometers, details the construction of the experimental platform, and conducts filament failure experiments under high-temperature electrified conditions in a vacuum environment. Subsequently, scanning electron microscopy and x-ray diffraction were employed to analyze both intact and damaged Re–W filaments, elucidating their failure mechanisms. Experimental results reveal the potential causes of filament damage during failure processes, including electromigration, thermal-stress-induced damage, and oxidation. This study represents the first systematic investigation into the failure mechanisms of Re–W filaments in quadrupole mass spectrometers, offering experimental evidence and technical support for future performance enhancement strategies. These findings contribute to improving the reliability of ECLSSs and the success rate of aerospace missions.

Reliable monitoring of oxygen concentration in aircraft fuel tanks is essential for explosion prevention, yet conventional TDLAS-WMS sensors require calibration with standard gases, a procedure that is impractical for dynamic airborne environments. To overcome this limitation, we present a robust, calibration-free method that integrates a physics-based WMS-2f/1f model with a parameter inversion algorithm. This approach treats the transition line center frequency v0, integrated absorbance A, and lineshape width dvC as free variables, eliminating the need for standard gas calibration. To ensure robust algorithmic convergence, an orthogonal test design strategy was implemented to systematically optimize the initialization, thereby resolving the instabilities inherent in multi-parameter non-linear fitting. To further ensure robustness under extreme conditions, we developed a pressure-dependent systematic error compensation algorithm and a theoretical temperature correction method. Experimental validation demonstrates that the system achieves a measurement precision of 0.0056% and a maximum relative error of 2.70% for O2 concentrations between 1% and 6% under atmospheric pressure. Furthermore, the compensation strategies maintain relative measurement error within +/- 2% across a wide pressure range (0.1-1.0 bar) and temperature span (266-326 K). A quantitative uncertainty analysis based on the GUM framework confirms an expanded uncertainty of 3.16% (k = 2), validating the system's suitability for aerospace safety monitoring.

Accurate gas analysis plays a critical role in aerospace missions, including spacecraft safety assurance, crew health monitoring, and deep-space scientific exploration. Although conventional gas chromatography (GC) techniques are well established, their large size, high power consumption, and long analysis time limit their applicability in modern aerospace missions that require miniaturized, low-power, and highly integrated analytical systems. The development of microelectromechanical systems (MEMS) technology provides an effective pathway for the miniaturization of gas chromatography. MEMS-based micro gas chromatography columns enable the integration of meter-scale separation channels onto centimeter-scale chips through micro- and nanofabrication techniques, significantly reducing system volume and power consumption while improving analysis speed and integration capability. Compared with conventional GC systems, MEMS µGC exhibits clear advantages in size, weight, energy efficiency, and response time. This review systematically summarizes the fundamentals, structural designs, fabrication processes, and stationary phase preparation of MEMS micro gas chromatography columns. Representative aerospace application cases along with related experimental and engineering validation studies are highlighted; we re-evaluate these systems using Technology Readiness Levels (TRL) to distinguish flight heritage from concept demonstrations and propose a standardized validation roadmap for environmental reliability. In addition, key technical challenges for aerospace deployment are discussed. This work aims to provide a useful reference for the development of aerospace gas analysis systems and the engineering application of MEMS-based technologies.

Hypopharyngeal carcinoma is a malignant tumour with a concealed location and difficult to detect in the early stages. Accurate prognosis assessment of hypopharyngeal cancer can help doctors develop reasonable treatment plans. Given that MRI is a primary modality for early-stage hypopharyngeal carcinoma detection, developing an AI-powered prognostic assessment system for hypopharyngeal cancer MRI examinations demonstrates substantial clinical value. However, hypopharyngeal cancer MRI adopts vertical-axis acquisition with larger slice gaps, resulting in lower pixel density along the vertical axis compared to coronal and sagittal axes. This poses a challenge for traditional 2D or 3D intelligent neural networks to adequately extract features from hypopharyngeal cancer MRI images. To this end, we propose Twist3DNet, a multiscale bidirectional 2D-3D information fusion prognostic classification network for hypopharyngeal cancer. We design a bidirectional fusion module, BT Block, to bidirectionally fuse the context from 2D Branch and 3D Branch. We propose a module, 3D M Module, with multiple receptive fields to capture global and local information in MRI images. We validate our method on the hypopharyngeal cancer dataset, BraTS2018 dataset and 3DLSC-COVID dataset. The experimental results demonstrate that our method achieves exceptional performance on these datasets, achieving mF1 scores of 75.78%, 89.09%, and 89.87%, respectively. Furthermore, a comparative analysis reveals that our method can effectively transform 2D networks into 2D-3D hybrid networks, surpassing the performance of corresponding 3D networks. We have made our code available. The code is available at: https://github.com/yk1842/Twist3DNet.

To address the escalating water resources consumption, the need for intelligent automation in urban sewage treatment has become imperative. However, existing ultrasonic sludge interface instruments have been found to exhibit poor environmental noise suppression and low identification accuracy of mud-water interface, thereby necessitating a new approach. This paper analyzes the problems and proposes a mud-water interface monitoring method based on extraction of ultrasonic energy peak point. The method involves an array ultrasonic system structure comprising multiple ultrasonic transmitting and receiving elements, median value average filtering to suppress peak noise interference, and interval maximum method to separate the target peak point from residual noise. The proposed method was tested using a high-precision ultrasonic probe in Beijing Ma-Fang Sewage Treatment Plant and corrected using a fault diagnosis method with 3σ threshold. Results demonstrated a high measurement accuracy with a mean deviation of less than 2% and a precision variance below ±0.02 m. Compared to conventional ultrasonic and optical methods, this approach exhibited superior robustness in turbid environments, fully meeting the technical requirements of sludge thickness monitoring.

The variation of the open-cell polyurethane foam thickness within the seating system can affect the dynamic response of the occupant-seat system and the riding discomfort. This study was aimed to develop and optimize a finite element model of the occupant-seat system incorporating the skeleton and muscle tissue for predicting the seat transmissibility and assessing the riding discomfort with different foam thicknesses. The key model parameters influencing the prediction of the seat transmissibility were investigated with the sensitivity analysis, and the correlation between the parameters and the seat transmissibilities was quantified by fitting the polynomial functions. The genetic algorithm was also utilized to iterate the squared error between the predicted and the measured seat transmissibilities, thereby obtaining the optimal parameter values. The best fit of the vertical inline and fore-and-aft cross-axis seat transmissibility predicted by the optimized model increased by 13.34 % and 14.41 %, respectively. When the dynamic stiffness of the foam decreased with the increase of the thickness, the peak frequency of seat transmissibilities, the weighted root-mean-square acceleration and the SEAT value exhibited the same downward trend.

Addressing the challenge of sharply increasing heat flux density due to spatial constraints in missile launch platforms and high-power integrated packaging modules, this paper proposes a novel split-type cold-plate (ST-CP) to achieve efficient thermal management via liquid cooling technology. Utilizing the mathematical expression defining the diverter hole size, diverter hole structure parameters are optimized using COMSOL Multiphysics 3.5a software. Compared to a traditional rectangular channel cold-plate, the ST-CP demonstrates significant advantages in temperature uniformity and heat dissipation efficiency, achieving a maximum temperature reduction of 15% and a 30% improvement in temperature uniformity. Furthermore, a mathematical model correlating the volume flow rate and flow resistance of the ST-CP is fitted, providing a design basis for engineering applications. The results solve the thermal dissipation problem for high-power density equipment and offer a new solution for enhancing heat dissipation uniformity.

Accurate detection of hydrogen leakage is critical for safety in high-pressure hydrogen systems used in fuel cell vehicles, where even minor leaks can lead to severe deflagration risks. Conventional detection methods, relying on fixed-location sensors and manual inspection, are labor-intensive, slow, and unsuitable for large-scale, realtime monitoring. Background-Oriented Schlieren (BOS) imaging offers a non-invasive visualization approach, but its diagnostic accuracy is limited by conventional displacement extraction algorithms such as crosscorrelation and optical flow. This study introduces a novel dual-channel Multi-Granularity Residual Network (MGR-Net) with transfer learning to overcome these limitations. The architecture employs a dual-channel input structure and specialized MGR Block and MGR Module for enhanced feature fusion, significantly improving diagnostic precision. Experimental results show a Dice coefficient of 81.88 %, exceeding ResUNet++ (77.34 %) by 4.54 percentage points, while reducing model size by 78 % and maintaining real-time processing at 42 frames per second, demonstrating both superior accuracy and efficiency for BOS-based hydrogen leak detection.

Hydrogen, a clean energy source, is pivotal in achieving carbon neutrality but presents safety challenges due to its high diffusivity and flammability. This study explores hydrogen leakage acoustics and spatial characteristics under high-pressure, unconfined conditions. Helium, validated as a surrogate gas with a sound pressure difference of <5 dB, was used for safe experimentation. Using Background Oriented Schlieren (BOS) imaging, significant directional and spatial dependencies were identified, with optimal detection at 1.5 m height and highest sound levels between 300 degrees and 360 degrees. The influence of nozzle geometry was also highlighted, with flat nozzles producing progressively higher sound pressure levels compared to circular ones. Threshold flow rates for audible detection were determined, approaching the maximum permissible leakage limit of 118 NL/min in noisy environments. These findings provide a foundation for safer and more effective hydrogen detection systems.

The O-ring made of nitrile rubber (NBR) is widely used in high-pressure hydrogen storage systems due to its excellent sealing characteristics. The sealing performance of the O-ring is directly related to the safety of the high-pressure hydrogen storage system. The stress and strain at the sealing interface are key factors in evaluating the sealing performance. Under extreme conditions, excessive stress and strain is also the reason for the failure of rubber materials. Prolonged exposure to high-pressure hydrogen gas can cause hydrogen permeation in O-rings, altering the stress distribution at the sealing interface. The paper establishes an analytical function model based on the modified hyperelastic effective modulus for the distribution of hydrogen pressure and stress on the sealing surface, revealing the stress extremum and stress distribution at the sealing interface under hydrogen permeation of the sealing ring. The paper proposes the influence of hydrogen concentration on hydrogen-induced strain in Orings, refining the composition of strain in O-rings under hydrogen permeation. By comparing the analytical model with the simulation model, the correctness and effectiveness of the analytical model were validated, while the accuracy of the simulation model was verified through experimental data from Kyushu University in Japan. The paper proposes two indices, S-hpsd (Skewness Coefficient of Hydrogen Permeation Stress Distribution) and K-hpsd (Kurtosis Coefficient of Hydrogen Permeation Stress Distribution), to describe the stress distribution at the sealing interface in Hydrogen Permeation environments. The results indicate that under the same compression ratio, the higher the hydrogen pressure within 0 MPa-70 MPa, the greater the normal stress at the sealing interface. The stress distribution at the sealing interface changes from U-shaped to V-shaped, with both Shpsd and K-hpsd gradually increasing. The kurtosis coefficient of hydrogen permeation stress distribution increases significantly. Under the same hydrogen pressure, within the range of 8%-20%, the higher the compression ratio, the higher the normal stress at the sealing interface. As the shape changes from U-type to Vtype, both S-hpsd and K-hpsd gradually decrease. The research findings of this study can provide references and theoretical support for maintenance strategies and failure prevention of O-ring seals in hydrogen fuel cell vehicles.

The modeling of the seat transmissibility is necessary to advance the understanding of the dynamic interactions between compliant seats and occupants. Within this investigation, an optimized artificial neural network (ANN) model was employed to clarify contributions associated with anthropometric parameters to the seat transmissibility. Anthropometric parameters underwent dimensionality reduction through the principal component analysis, and resultant principal components served as input features for the ANN model. Additionally, the ANN structure’s weights and biases values were adjusted using the genetic algorithm (GA), resulting in a PCA-GA-ANN model for the prediction of seat transmissibilities. The results indicated root mean square error (RMSE) values for predicting vertical in-line and horizontal cross-axis transmissibilities from the developed model were 0.061 and 0.055, respectively, demonstrating superior effectiveness in the prediction error and trends when compared with both the ANN and GA-ANN models. The seat transmissibility predicted from the PCA-GA-ANN model exhibited resonance behaviors similar to that observed in the whole-body vibration test. The sensitivity analysis showed that the subject’s age was the most predominant anthropometric parameter for the prediction, followed by gender and body mass index. The ANN model optimized with principal component analysis (PCA) and GA effectively eliminates the redundant information of anthropometric parameters, enhancing the generalization of the seat transmissibility prediction.

This paper focus on the advanced multi-energy storage systems interconnection by DC smart grids with high efficiency and high compactness. A non-isolated modular high conversion ratio bidirectional soft switching DC-DC converter and its extended multi-ports structure are proposed in this paper. In each power module of the converter, an auxiliary zero-current transition (ZCT) cell is integrated to achieve the zero-current turn-on and turn-off for all semiconductor devices. Therefore, the switching losses are reduced greatly and the efficiency is improved considerably. The presented converter is suitable for the small and medium power scale interconnection in smart DC grids, and can realize the flexible multi-energy storage system interaction owing to its extended multi-ports structure. Besides, various control schemes are adopted to optimize the control system of this converter, which can improve the performance on energy interaction. The operation principles and design considerations are discussed in details and a 5 kW experimental prototype is constructed to justify the validity of the theoretical analysis.

Non-Dispersive Infrared (NDIR) is an optical, non-contact, multi-component measurement technology that can be used for online monitoring of gas components and contaminants in liquids. It offers advantages such as good selectivity, strong anti-interference capability, long service life, and high accuracy. NDIR sensors are widely used in industrial, environmental, medical, food, and aviation fields. Based on the principle of gas absorption of specific wavelength infrared light, NDIR sensors provide long-term stability and resistance to environmental changes. This paper reviews the applications of NDIR sensors in monitoring gas concentrations and assessing oil conditions. Additionally, it summarizes the substances that NDIR sensors can monitor according to their absorption wavelengths and outlines the principles for selecting light sources in NDIR sensors.

Hydrogen is regarded as a crucial low-carbon energy carrier for energy storage. Due to their excellent sealing properties, rubber O-rings have been widely used in the field. However, rubbers are prone to hydrogen permeation and swelling, which can compromise their service life and ultimately result in sealing failure, posing significant safety risks. Therefore, predicting the lifespan of sealing components subjected to hydrogen is of critical importance. This study first interprets the macroscopic phenomena of hydrogen permeation and swelling from the perspective of molecular chain statistics. Subsequently, a lifespan prediction model for hydrogen permeation and swelling is proposed, and FEM of the sealing component is developed to calculate strain energy density. This method can predict the service life of sealing components under certain conditions, thereby improving the efficiency of hydrogen utilization safety.

Based on the Goos-Hänchen effect and the intensity attenuation of the evanescent wave penetrating and traveling, a new theoretical model of total internal reflection from the interface of turbid media is proposed and an analytical reflectance expression in a wide incident angle range is developed. The Goos-Hänchen angle displacement between the critical reflectance and the saturated reflectance is discovered. A sensor, for measuring the complex refractive index of turbid media in real-time, with divergent light source is designed. The captured images show that the light distribution reflected from the transparent medium has a sharp boundary, but for turbid media, the reflected light intensity attenuates during the transition from total to non-total internal reflection regions. It is successful and accurate that the new model fits the experimental data of the reflectance and the complex refractive index of turbid media is measured by our sensor. The results show that measuring has advantages in real-time, in situ, and with high accuracy.

This study aims to figure out some safety issues by investigating the roof shape of infrastructures such as zero-carbon huts that people are already using to prevent dangerous hydrogen clouds from building up in places with little airflow when there is a leak. We designed three different obstacles. The horizontal and vertical concentration distributions of the free-jet hydrogen cloud when encountering different obstacles were determined by concentration measurements. In addition, the diffusion characteristics of the hydrogen jet in the presence of obstacles were investigated by using the Background Oriented Schlieren (BOS). The results show that there is a linear relationship between the measured axial concentration of the hydrogen jet and the leakage parameters at lower flow rates. The measured radial concentration of the hydrogen jet shows a "Gaussian-like" distribution when encountering a flat plate obstacle, a "W" distribution when encountering a flat plate obstacle with sidewalls, and a "W-like" distribution when encountering a dome-type obstacle. A method of inverting the radial concentration distribution of the hydrogen jet using the BOS is also proposed, and the error between the theoretical and experimental values is calculated to be 3.81 %. This study stresses the significance of geometric and dimensional parameters on hydrogen leakage characteristics and, as a result, provides insights into developing performance standards for the availability and reliability of safety-critical systems.

This study investigated the relationship between electrocardiography (ECG) and serum potassium levels in 100 children aged 6–18 years with chronic kidney disease admitted to the Department of Nephrology, Beijing Children's Hospital Affiliated, Capital Medical University, China. The research data were obtained from children in their growth and development stages with significant differences in their physical characteristics. We established a promising indicator, Ts/a, calculated Ts/a for V2–V6 in 100 children, and correlated it with their serum potassium levels. Despite significant differences in age and body shape between developing children and adults, the results showed that the Ts/a values of V5 and V6 were more strongly correlated with blood potassium values. T wave at the V5 and V6 leads in children can reflect the direction of the primary amplitude of the T-wave and T-wave downward slope, and the heart space position in the chest characteristics during childhood lead to these results. This study provides an essential foundation for the future mobile monitoring of serum potassium levels.