Search Results (523)

Credit card fraud detection is a typical rare-event classification problem because fraudulent transactions usually account for only a very small proportion of all transactions. Conventional evaluation on balanced or resampled test data may lead to overly optimistic performance estimates. To address this issue, this study proposes a leakage-safe credit card fraud detection framework integrating Random Forest Gini impurity-based feature selection, resampling strategy evaluation, and Genetic Algorithm (GA)-based hyperparameter optimization. The framework was evaluated on the public European credit card fraud dataset containing 284,807 transactions, of which only 492 were fraudulent. The original dataset was first divided into a stratified training set and an untouched original-distribution test set. Feature selection, standardization, resampling, GA optimization, and threshold tuning were performed only on the training data or training folds. The final test set contained 85,443 transactions, including 148 fraudulent transactions, and was used only once for final evaluation. Experimental results show that GA-XGBoost achieved the best overall balance among the optimized models, with a PR-AUC of 0.798, ROC-AUC of 0.967, MCC of 0.814, balanced accuracy of 0.865, fraud-class precision of 0.908, fraud-class recall of 0.730, and fraud-class F1-score of 0.809. Compared with baseline XGBoost, GA-XGBoost improved PR-AUC from 0.741 to 0.798, MCC from 0.766 to 0.814, and fraud-class F1-score from 0.764 to 0.809, while reducing false positives from 22 to 11 and false negatives from 43 to 40. The ablation results further indicate that resampling strategies are not universally beneficial and should be evaluated under the original test distribution. These findings suggest that leakage-safe evaluation and fraud-class-oriented metrics provide a more reliable basis for practical credit card fraud detection. Full article

Optimizing the carrier concentration across the entire operating temperature range is crucial for maximizing the power factor in n-type PbTe. However, conventional shallow donors produce a nearly temperature-invariant electron concentration, leading to an increasingly large deviation from the optimal carrier concentration at elevated temperatures. Herein, we implement a dynamic deep-shallow co-doping strategy by combining iodine (a shallow donor) with gallium (a deep-level donor) in PbTe. The Ga-related deep impurity states thermally ionize at elevated temperatures, providing additional electrons and driving the Hall carrier concentration above ~563 K toward its temperature-dependent optimum. Concurrently, our optimized synthesis preserves a high carrier mobility, which synergistically sustains a remarkable peak power factor of 30 μW·cm⁻¹·K⁻² for the optimal composition, Ga_0.02Pb_0.98Te_0.996I_0.004. Combined with a strongly suppressed lattice thermal conductivity, this results in a maximum figure of merit (ZT) of 1.41 at 803 K and an average ZT of 1.00 within 400–773 K for Ga_0.02Pb_0.97Te_0.996I_0.004—a 25% improvement over the I-only doped baseline. These findings establish deep-shallow co-doping as a robust and broadly applicable carrier-engineering paradigm for thermoelectric optimization. Full article

Hot cracking in welding is a very complex phenomenon. It can happen in the weld metal zone during solidification but also in the heat-affected zone (HAZ). Hot cracking defects are material decohesion that occur at high temperatures along grain boundaries when the strain and strain rate exceed a certain level. The cracks can be internal or open to the surface in the weld bead. During a welding operation, different types of hot cracks can appear, such as hot cracking due to solidification, hot cracking due to liquation, hot cracking due to loss of ductility. The main factors favoring hot solidification cracking include the presence of residual elements and impurities, leading to the formation of a low-melting eutectic; the solidification mode; and mechanical restraints. This review paper gives an introduction to solidification cracking in stainless-steel welds, the weldability of the austenite grades, and the causes of solidification cracking occurrence. The main methods with which to detect and inspect cracks are investigated. Particular focus is placed on TIG (tungsten inert gas), also known as Gas Tungsten Arc Welding (GTAW). A review of the literature reveals that considerable progress has been made in terms of the improvement in the properties of the weld joint through the application of mitigation means and strategies. The effort made by researchers in understanding solidification cracking phenomena has been key to enhancing cracking resistance and ensuring the integrity of structures. Full article

Low-melting-point impurities in powder feedstock can trigger local melting phenomena in laser powder bed fusion (LPBF) parts and may initiate defects in printed components. Here, we combine bulk chemistry with automated, high-throughput particle-by-particle SEM/EDS to identify and quantify Sn-containing impurity particles in two gas-atomized Al–Mg–Zr powder batches with different bulk Sn levels. The aim was not to establish a direct batch-to-batch performance comparison, but to clarify whether Sn was uniformly distributed among the powder particles or concentrated in rare impurity particles. Although ICP analysis indicated only 0.07 ± 0.02 wt.% Sn in the Sn-higher batch and <0.01 wt.% Sn in the Sn-lower batch, automated SEM/EDS screening of 20,001 particles per batch revealed that Sn was present as a very small number of highly enriched particles with Sn > 45 wt.% (eight particles in the Sn-higher batch and three particles in the Sn-lower batch). In the Sn-higher batch, Sn-rich particles were predominantly spherical and fell within the LPBF feedstock size window (Dmax ≈ 25–40 μm), implying that standard sieving would not remove them. BSE imaging and EDS mapping of polished sections and fracture surfaces of LPBF specimens built from the Sn-higher batch revealed spatially localized Sn-rich features associated with pores and Sn-rich phases on the fracture surface, supporting a direct powder-to-part transfer. These results demonstrate that low bulk impurity levels can mask highly localized, particle-scale contamination and highlight the need for particle-level compositional screening to support robust powder qualification and reuse decisions in LPBF. Full article

To address prevalent industrial challenges, including the high cost of fabricating microstructures via photolithography and 3D printing, impurity residues easily generated by conventional physical/chemical pore-forming techniques, and the limited sensitivity of regular capacitive sensors, this paper innovatively proposes an integrated low-temperature in situ gas foaming strategy using ammonium bicarbonate for the fabrication of porous TPU-based ionic gels. Relying on the complete gaseous decomposition property of ammonium bicarbonate upon heating, a three-dimensionally interconnected continuous porous network is spontaneously constructed inside the polymer matrix. Thermoplastic polyurethane (TPU) is selected as the continuous polymer phase, and [EMIM][TFSI] imidazolium ionic liquid is blended as the ion source to synthesize composite ionic gel substrates. A PDMS composite slurry filled with graphene is employed to prepare flexible substrates, followed by low-temperature oxygen plasma surface modification to introduce polar functional groups such as hydroxyl and carboxyl onto electrode surfaces. A standard sandwich-structured ionic pressure sensor with the configuration of “top modified electrode—porous ionic gel dielectric layer—bottom modified electrode” is finally assembled. The porous framework and modified electrodes constitute a dual synergistic enhancement system: the porous structure markedly reduces the equivalent elastic modulus of the gel and improves its compressive deformation capacity; polar-modified electrodes optimize the interfacial compatibility between electrodes and gels, shorten ion migration paths and lower interfacial contact resistance. Systematic calibration of multiple batches of parallel samples reveals that the as-fabricated sensor achieves a high sensitivity of 25.3 kPa⁻¹ across the full measuring range from 0 to 1000 kPa with a linear fitting coefficient R² = 0.992. The loading response time and unloading recovery time of the device are 60 ms and 80 ms respectively, with a performance degradation of less than 3% after 1000 consecutive loading–unloading cycles, featuring low hysteresis error and excellent signal repeatability. Multi-scenario in vivo wearable tests on human subjects verify that the device can precisely capture subtle fluctuations of radial artery pulse and periodic laryngeal deformation during swallowing, distinguish characteristic waveform patterns of various English words according to differences in vocal cord vibration, and accurately detect bending motions when attached to finger joints. The entire fabrication process adopts common chemical raw materials and standard laboratory equipment without expensive micro-nano processing facilities, featuring convenient raw material procurement and high process fault tolerance, which enables large-area coating-based mass production. This work delivers a novel technical route for the low-cost large-scale production of high-performance ionic flexible sensors and bears significant industrialization reference value for applications in wearable medical monitoring, bionic robotic electronic skin, flexible human–machine interactive touch panels and other related fields. Full article

This study experimentally evaluates hydrogen recovery from synthetic syngas and water–gas shift (WGS) syngas using a laboratory-scale pressure swing adsorption (PSA) unit equipped with layered activated carbon/zeolite 5A beds. Breakthrough tests were first performed to determine adsorption-time limits and identify the critical impurity controlling product quality. Continuous PSA experiments were then carried out using two cycle configurations: a two-bed Berlin-type cycle and a four-bed Linde-type cycle. CO was the first impurity breakthrough experimentally detected and it therefore defined the practical adsorption-time cut-off, whereas CO₂ exhibited the strongest retention, especially in beds with an increased activated-carbon fraction. The results showed a clear trade-off between purity and recovery. The four-bed Linde-type cycle provided a wider operating window than the two-bed Berlin-type cycle, owing to pressure equalization and product-purge steps. The best overall performance was obtained for WGS syngas with the 1.6:1 AC:zeolite bed, reaching 99.5 vol.% H₂ at 84% recovery and maintaining 99.2 vol.% H₂ at 86% recovery. The tail gas was enriched in CO₂ up to approximately 72 vol.%, indicating potential for integration with downstream CO₂ management. Full article

In order to improve the accuracy of the lifting state detection of an oil–gas jack-up platform, a lifting state detection method based on improved random forest is proposed to solve the problems of low detection efficiency caused by the interference of lifting data and redundant features. To detect outliers in the lifting data of an oil–gas jack-up platform by the K-means clustering method and clean abnormal data, the principal component analysis method is introduced into the random forest algorithm to reduce the dimension of lifting data, and an improved random forest is constructed with the Gini impurity criterion to preliminarily classify the lifting state. Then, fuzzy comprehensive evaluation is used to refine the state of the classification result and realize lifting state detection. The test results show that the proposed method has good stability in lifting state timing detection, a high-inter class-to-intra-class distance ratio, and accurate platform displacement detection under different incident angles/motion responses. Full article

The oscillator strengths of absorptive transitions from the ground to the resonant excited impurity states for the impurity positioned in and near the GaAs/AlGaAs rectangular quantum well are studied. Due to the resonant nature of the final states, the absorption peaks are broadened. The shape of the peaks is reproduced numerically as a function of impurity position with respect to the well and the well width. Peak parameters, such as maximum, broadening, and integral absorption, are analyzed numerically; the Fano parameter is considered qualitatively. Full article

Catalytic oxidation has become a crucial technology for removing CO from industrial flue gas. However, the complex composition of flue gas (including NH₃, NO, SO₂, H₂O, etc.) poses significant challenges to the catalytic activity and stability of catalysts. In this work, we propose a new strategy for constructing highly efficient catalysts by loading a Pd component onto TiO₂ nanosheets (NSs) with predominantly exposed {001} facets. It has been revealed that the well-connected channels, abundant oxygen vacancies and Ti³⁺ species on the TiO₂(NS) support facilitate the formation of highly dispersed and electron-rich Pd nanoparticles. The weak adsorption of impurities such as NH₃, SO₂, NO and H₂O on these active sites promotes the adsorption and activation of the target reactants (CO and O₂), thereby enhancing catalytic activity. Furthermore, such reduced adsorption inhibits the aggregation of Pd nanoparticles and synergizes with the intrinsically weak NH₃ adsorption of TiO₂(NS) to suppress ammonium sulfate species deposition, thereby enhancing long-term catalytic stability. This work advances TiO₂ facet engineering in catalysis and offers new design concepts for efficient CO oxidation catalysts in complex atmospheres. Full article

The adsorption of H₂S impurity in the gas flow on carbon adsorbents produced from coconut shells and sugarcane bagasse was studied. The runs were carried out in pulse mode. An original chromatographic method for determining the degree of H₂S absorption on carbon adsorbents has been developed, which makes it possible to determine the amount of absorbed H₂S in air and water environments. The results obtained show that the successive treatment of carbon adsorbents first with a solution of a strong oxidizer (HNO₃, KMnO₄) and then, after washing, with an alkali solution (KOH) leads to a sharp increase in the amount of H₂S adsorbed. The efficiency of H₂S absorption on the obtained adsorbent reaches 85.5%, which corresponds to 27.7 mg/g H₂S and is comparable to the results obtained on commercial coconut carbon (CAU). The data allow one to conclude that the rise in the H₂S adsorption on the carbon sorbents studied can be due to the increase in the micropores’ volume in the activated carbon, as well as the formation of surface functional groups containing an alkali metal (i.e., C-OK, C-COOK) that promotes irreversible chemisorption of H₂S impurity on the carbon adsorbent. The absorption of H₂S occurs through the chemisorption mechanism, which is confirmed by IR spectroscopy data. Full article

The antifoaming performance of natural gas desulfurization solvents is critical for maintaining product gas quality and ensuring the safe operation of processing units. Hydrocarbon impurities can enter amine solutions through feed-gas entrainment, wellhead flowback carryover, and leakage of equipment lubricants. These contaminants may gradually accumulate in the solvent system and become a significant contributor to foaming. To address the industrial demand for rapid quantitative determination of hydrocarbon contaminants in desulfurization solvents, this study investigates in-service UDS-series solvents and representative samples collected from a natural gas purification plant in western Sichuan. NMR spectroscopy and GC-MS analyses reveal that the impurities are predominantly n-alkanes in the C₁₃-C₁₈ range, based on which a corresponding reference standard oil was prepared. COSMO-RS calculations combined with molecular interaction analysis identify n-hexane as the optimal extraction solvent. The ultraviolet spectrophotometric method commonly used to determine hydrocarbons in environmental water samples shows limited sensitivity to long-chain n-alkanes and requires strong acid pretreatment that disrupts the amine solvent matrix, rendering it unsuitable for UDS solvents. In contrast, the n-hexane extraction-GC-FID method showed good linearity, precision, and accuracy, meeting engineering analytical requirements for monitoring hydrocarbon contamination in MDEA-based UDS desulfurization solvents. Full article

To accelerate the transition to sustainable energy, efficient methods for CO₂-free hydrogen production and carbon utilization are needed. This study presents a new, sustainable approach for the simultaneous production of hydrogen, valuable hydrocarbons, and functional carbon materials by converting methane in low-pressure microwave plasma. Compared to traditional methane reforming methods (such as steam reforming), our plasma-based process operates at low temperatures, eliminates direct CO₂ emissions, and enables the conversion of methane into three valuable products: (1) environmentally friendly hydrogen for fuel cells and energy storage systems, (2) a range of valuable organic products (C₂H₂, C₂H₄, C₂H₆), and (3) functional carbon films with self-improving catalytic properties. Optical emission spectroscopy (OES) and the Langmuir double probe method were used for plasma diagnostics, revealing an increase in the concentration of active species (CH, Hα, C₂) and electron temperature upon argon addition. The structure, morphology, and impurity composition of the deposited films were investigated using X-ray diffraction (XRD), scanning electron microscopy (SEM), and inductively coupled plasma mass spectrometry (ICP-MS), respectively. Gas-phase byproducts were analyzed using gas chromatography–mass spectrometry (GC-MS). Argon addition at an Ar/CH₄ ratio of 1 leads to the formation of carbon films with a more ordered structure, as confirmed by XRD data, and improved surface morphology. It was established that argon, by effectively participating in the excitation and dissociation processes of methane molecules through energy transfer from metastable states and increased electron temperature, optimizes plasma–chemical reactions, promoting the deposition of higher-quality carbon coatings. Full article

Severe foaming and a significant decrease in desulfurization performance were noted in a novel UDS solvent applied in a natural gas field in western Sichuan, China. The effects of hydrocarbon and ionic impurities on foaming behavior and the purification performance of candidate adsorbents were investigated. An extraction-gas chromatography method was established and validated for determining total hydrocarbons in amine solutions, enabling quantitative evaluation of hydrocarbon contamination. Controlled contamination experiments revealed that hydrocarbons had the strongest effect on foaming, while sulfate and chloride strongly promoted foam formation; organic acid anions showed only minor effects. Fixed-bed screening identified A-98FM anion-exchange resin as the most effective for anionic impurity removal and AC-02 activated carbon as the best candidate for hydrocarbon purification, with a cumulative adsorption capacity q_0–12 of 14.86 mg/g over 12 h. Pore-structure and thermal-release analyses suggested that conventional pore descriptors alone could not fully explain the dynamic purification performance, while hydrocarbon-related loadings in spent AC-02 occupied accessible pore space and contributed to performance decay. Treatment of a field-aged UDS lean solvent further showed that reductions in target impurities were accompanied by lower foam height and shorter defoaming time. This work provides experimental support for impurity monitoring, foaming-risk identification, and adsorptive purification of UDS desulfurization solvent under flowback-contamination conditions. Full article

Given the increasing concern around greenhouse gas emissions and the decline in the availability of fossil fuels, there is increasing global demand to develop alternate fuels for maritime transportation that are sustainable and which have lower greenhouse gas emissions. Methanol is one such alternative fuel that has garnered considerable attention given its potential to be produced by more sustainable processes and its more favorable greenhouse gas emission profile in comparison with current fossil fuels. Understanding the physical and chemical properties of methanol under a range of conditions is essential for its development as a marine fuel. In this study, we seek to define physical and chemical properties of different methanol samples to simulate real-world storage conditions as these data are lacking in the literature. Several methanol samples were evaluated: nearly pure methanol; International Organization for Standardization (ISO) marine methanol (MM) grades A, B, and C; and methanol plus higher alcohols. We first evaluated all methanol samples for impurities, acetic acid content, density, and distillation range. We then characterized the effects of water absorption and found that methanol can easily absorb unacceptable water content from humid air within hours, necessitating storage conditions that prevent this process. In eight-week aging experiments at 20 °C and 40 °C in ambient air, we did not observe significant oxidation for any of the methanol samples; however, we did observe increases in acid number. We assessed the impact of contamination of methanol with water, marine gas oil (MGO), and an MGO–biodiesel mixture on density, viscosity, distillation range, and lubricity. Finally, we show that MGO contamination of methanol results in a slight increase in sooting tendency. In aggregate, our results provide an in-depth analysis of physical and chemical properties of methanol as well as the impacts of storage conditions and impurities on the properties of fuel methanol. Full article

Aiming at the problems of low separation efficiency and high impurity content in the cleaning process of moist rice threshed materials, this study systematically explored the influence mechanism of moisture on the segregation behavior of rice threshed products by combining physical experiments and CFD-DEM coupling simulation. Physical test results show that moist conditions significantly change the properties of threshed materials: the impurity mass fraction increases from 3.1% to 6.8%, the straw breakage rate rises from 5.97% to 6.99%, and the working load and unbalanced dynamic load of the threshing unit increase noticeably. On this basis, a gas–solid coupling simulation model of the cleaning device embedded with the Johnson–Kendall–Roberts (JKR) cohesive contact model is established. It is found that moist threshed materials exhibit an obvious stratified movement on the sieve surface, in which the bottom straw layer moves slowly while the upper material layer flows rapidly, resulting in a 50% increase in the sieve surface load. This phenomenon directly accounts for the rising impurity content observed in experiments. Through integrated experimental and simulation analysis, this research clarifies the macroscopic laws of moisture effects and reveals three key pathways for moisture to deteriorate cleaning performance: changing physical characteristics of threshed materials, enhancing inter-particle adhesion, and forming stratified movement. The findings provide an innovative research scheme and reliable theoretical support for the design and optimization of high-efficiency cleaning devices for rice combine harvesters. Full article