Abstract:Thin overlays have become a primary pavement maintenance measure in recent years. Research on thin overlays primarily focuses on optimizing material composition and applying cost-effective overlay types. The influence of commonly used raw materials on the service performance of thin overlays remains insufficiently understood. Therefore, this study compared the service performance of thin overlays prepared with coarse aggregates (basalt, limestone, steel slag, and finely separated reclaimed asphalt pavement (RAP)). Specifically, two typical gradations, stone matrix asphalt (SMA-10) and asphalt concrete (AC-10C), were first selected. Then, the 2.36–4.75 mm, 4.75–9.5 mm, and 9.5–13.2 mm fractions were completely replaced with the four target coarse aggregates on an equal-volume basis. Finally, the service performance of the resulting specimens was tested and evaluated. The results showed that steel slag exhibited significant advantages in high-temperature stability, shear resistance, and skid resistance. It was identified as the optimal coarse aggregate. The SMA-10 overlay demonstrated superior overall performance compared to the AC-10C, and should be preferred when the budget allows. Finely separated RAP can be used in thin overlays, but it exhibited the poorest overall performance, with its freeze–thaw resistance approaching the lower specification limit. This indicates a need for further field validation.

Abstract:To investigate the characteristics of air-extraction-induced negative-pressure penetration of bucket foundations and the effects of negative pressure level and bucket diameter on penetration behavior and pore-water pressure response, a series of 1g model tests were conducted using bucket diameters of 108, 127, and 141 mm under negative pressures of ?10, ?20, and ?30 kPa. A penetration resistance model was also established by considering seepage along the inner and outer bucket walls and the seepage-induced reduction in resistance at the bucket tip. The results show that, compared with the ?10 kPa condition, the penetration times under ?20 and ?30 kPa were reduced by 81.1%–87.4% and 87.7%–91.6%, respectively. However, when the negative pressure was further increased from ?20 to ?30 kPa, the penetration time decreased by only an additional 2.5–4.0 s, indicating a diminishing gain in penetration efficiency with further increases in negative pressure. The effect of bucket diameter weakened as the negative pressure increased, with the maximum difference in penetration time among the three bucket diameters decreasing from 36.5 s to 3.5 s and 2.0 s. The peak magnitude of negative excess pore-water pressure increased with increasing negative pressure, whereas the normalized pore-pressure response gradually decreased, and the peak values mainly occurred during the later stage of negative-pressure penetration. The theoretical model showed good agreement with the experimental results under relatively low negative pressure and during the near-quasi-static penetration stage, with R² values ranging from 0.875 to 0.954.

Abstract:To address the issues of DC bus voltage drop, source-load power imbalance, and battery overcharge/overdischarge in off-grid PV-storage-charging systems under photovoltaic power fluctuations and load transients, a control strategy combining energy storage-side control and system energy management is proposed. First, linear active disturbance rejection control (LADRC) is introduced into the DC bus voltage regulation of the Buck-Boost converter. A linear extended state observer is employed to estimate and compensate for the total system disturbance, thereby improving the dynamic response delay and poor parameter adaptability of traditional PI control. Second, a virtual capacitor model is introduced, and a three-loop competitive control strategy consisting of bus voltage regulation, battery constant-voltage charging, and battery constant-voltage discharging is designed to achieve DC bus power balance and battery overcharge/overdischarge protection. Furthermore, by integrating the MPPT/derating switching control of the PV Boost converter and the EV-side load derating control, a load-priority energy management strategy is established. Experimental results demonstrate that the proposed method can reduce the DC bus voltage drop during load switching, achieve coordinated operation among PV generation, energy storage, and loads, and improve the stability and reliability of the off-grid PV-storage-charging system.

Abstract:Automobiles are increasingly transforming into user-centered intelligent service platforms, and driver identification models capable of perceiving driver identity in real time are essential for delivering personalized services. The identification performance of these models can be affected by variations in fatigue states. However, most existing models are trained using data collected under alert states, without considering the effects of fatigue. Therefore, this study proposes a driver identification method that adapts to fatigue evolution. First, multimodal data, including driving operation behaviors, facial behaviors, and fatigue ratings, were collected. Then, the Kruskal-Wallis test was employed to select discriminative driver identification features. Finally, by integrating driving data from different driving states, a driver identification model based on principal component analysis and a one-dimensional convolutional neural network was developed. The results show that the proposed model achieved average accuracy, precision, and recall of 88.72%, 89.97%, and 88.80%, respectively, for driver groups of 3, 9, 15, and 21 participants, representing increases of 23.34%, 22.53%, and 23.24%, respectively, over the conventional model. The proposed method effectively improves the driver identification accuracy and the adaptability to fatigue evolution, providing support for accurate driver identification and the development of personalized driving systems.

Abstract:To address the insufficient discriminability of fault-section features caused by disturbances such as train starting, load fluctuations, and rectifier ripple in metro DC traction networks, an active harmonic injection-enhanced CNN-LSTM method is proposed for short-circuit fault location. Harmonic signals are actively injected into the traction network to enhance the fault-discriminative information contained in the input features. The injected harmonic response is then extracted using the empirical wavelet transform and employed as the fault-location feature. A convolutional neural network is used to extract local waveform features, while a long short-term memory network is introduced to learn temporal variation characteristics, thereby enabling the classification of train-starting conditions and different fault sections. The results show that, compared with the original traction current input, using the harmonic response induced by active injection as the input significantly improves the fault-location performance of all evaluated models. Among them, the CNN-LSTM model achieves both an accuracy and an F1-score of 98.03%, effectively distinguishing train-starting conditions from short-circuit faults and accurately identifying different fault sections. The proposed method provides an effective approach for short-circuit fault-section location in metro DC traction networks.

Abstract:To investigate vibration transmission and mitigation in a phononic crystal embedded rail track under transient wheel–rail impact, full-scale drop-wheel tests were conducted in comparison with a conventional fastened track. Three conditions were examined: 15 mm unloaded, 15 mm fully loaded, and 25 mm unloaded. Vertical accelerations were recorded simultaneously at the rail, track slab, and foundation. Six time–frequency analysis methods were compared in terms of energy concentration and consistency with vibration decay. Wavelet synchro squeezing transform (WSST) was selected to examine the time–frequency evolution of the impact response, while one-third-octave vibration levels were used to evaluate the mitigation effect. WSST achieved the highest overall evaluation index among the six methods, at 0.52192. In the phononic crystal embedded rail track, the vibration response decreased progressively from the rail to the track slab and foundation, with the foundation response concentrated mainly within 20–80 Hz. Compared with the conventional fastened track, its overall vibration levels under the three conditions were lower by 29.5, 15.1, and 3.9 dB within 1–80 Hz, and by 28.8, 14.4, and 3.4 dB within 4–200 Hz, respectively. WSST characterizes the dominant frequency ridges and their decay following drop-wheel impact, providing a basis for a more detailed evaluation of track vibration mitigation under transient impact.

Abstract:Discretionary lane-changing (DLC) is a complex and safety-critical maneuver in highway driving that presents substantial challenges to traffic safety and operational stability. Accurately modeling DLC decisions remains difficult. Many existing approaches have limitations. They often fail to capture dynamic risk interactions with surrounding vehicles. They also inadequately model the temporal evolution of driver intent. These limitations often lead to limited prediction accuracy and generalizability. To address these gaps, we propose a new model, the Spatiotemporal Risk Field-integrated DLC (SRF-DLC) model. This model integrates time-sequential vehicle kinematics, driving behavior, and risk representation using an Attention-based LSTM network. The model was trained on trajectories from the highD dataset. We used 5,631 trajectories, each lasting 1 to 4 seconds. The model forecasts three maneuvers: lane-keeping, overtaking, and fold-down. Results show that the model trained on 3-second trajectories performs best. It achieves a peak accuracy of 96.74%. This optimal history length aligns with natural human decision-making timelines. Ablation studies were also conducted. By that we confirm the critical role of the Spatiotemporal Risk Field. Excluding this component caused a significant drop in accuracy. The highest accuracy decrease reached 47%. In comparisons, the SRF-DLC model outperformed some existing methods. This demonstrates that explicit spatiotemporal risk contextualization and appropriate history length selection is crucial in accurate DLC prediction.

Abstract:This study proposes a maintenance task decision-making method that integrates FMEA, RCM, and PHM. An FMEA failure risk assessment model is established in accordance with EN 50126 to quantify failure risks and determine risk grades across three dimensions: failure severity, occurrence frequency, and detectability. Reliability-centered maintenance analysis is implemented to screen critical maintenance items, analyze functional failures, classify failure impacts, and make logical decisions about maintenance tasks, thereby forming standardized maintenance tasks. Combined with PHM technology, maintenance strategies are optimized based on a matrix of failure impact and occurrence frequency; candidate items for PHM construction are selected with priority defined; and an integrated maintenance decision-making system covering risk assessment, reliability decision, and intelligent health management is ultimately established. Full-process verification is performed with the IGBT module, the core component of traction systems, as the research object. The results reveal that the proposed method reduces the risk level of IGBT modules from unacceptable to acceptable and extends the tertiary maintenance mileage from 1.2 million km to 1.6 million km, effectively improving the economic efficiency and intelligence of equipment operation and maintenance. The hybrid-driven decision-making methodology establishes a progressive closed-loop framework of FMEA - RCM - PHM - RCM. By leveraging a unified risk quantification value as the linking bridge, it achieves the integrated fusion of the three approaches. This provides both theoretical support and practical engineering reference for optimizing maintenance systems and constructing intelligent operation and maintenance frameworks for railway vehicle traction systems.

Abstract:As environmental concerns caused by waste polyethylene terephthalate (PET) become increasingly severe, utilizing it as a resource for asphalt modification has emerged as an important direction in road engineering. However, existing studies have predominantly focused on PET-modified SBS or rubber asphalt individually, leaving the effects of PET-derived additives on SBS/CR composite modified systems inadequately understood. In this study, a PET chemical depolymerization product (PET EA) was prepared via aminolysis and incorporated into both low content (3% SBS + 5% CR) and high content (7% SBS + 10% CR) composite modified asphalts at various dosages. The influence of PET EA content on asphalt properties was systematically investigated, and fluorescence microscopy was employed to elucidate the modification mechanism. The results show that PET EA increases the softening point and viscosity of the SBS/CR modified asphalt systems, while adversely affecting ductility; these effects are more pronounced in the high content system. Specifically, at 7% PET EA, the softening point increases by 4.8?°C, the viscosity at 135?°C rises by approximately 185%, and the maximum tensile stress during ductility testing increases by 85%, whereas ductility decreases from 19.7?cm to 6.4?cm. Furthermore, PET EA acts as a stabilizer in the composite system, inhibiting sedimentation and phase separation of the modifiers during high temperature storage, thereby improving storage stability. The amide groups present in PET EA promote the migration of light components from the asphalt into the SBS phase, facilitating further swelling of SBS, which in turn affects macroscopic properties such as the three conventional indices and viscosity. Overall, PET EA can serve as an effective additive for SBS/CR composite modified asphalt, offering both performance enhancement and stabilization. The findings provide a theoretical reference for the preparation of PET SBS CR composite modified asphalt.

Abstract:To achieve rapid and accurate prediction of the natural carbonation depth of in-service bridge concrete under ordinary atmospheric environments, a dataset consisting of 496 groups of field-measured concrete carbonation samples from bridges is first established. Artificial Neural Network (ANN), Support Vector Machine (SVM) and Random Forest (RF) algorithms are adopted to screen and optimize feature combinations. Regression trees and four tree-based ensemble learning (EL) algorithms are separately applied to develop prediction models for the natural carbonation depth of bridge concrete, followed by interpretability-based machine learning analysis. The results show that: the semi-empirical semi-theoretical model specified in relevant current Chinese standard produces highly scattered predictions for natural carbonation depth of bridge concrete, while the EL model built with the Gradient Boosting Regression Tree (GBRT) algorithm delivers outstanding performance and enables fast and precise prediction results. Among the factors affecting natural carbonation of bridge concrete, concrete compressive strength and curing age stand out as the two most dominant features, negatively and positively correlated with carbonation depth respectively. The carbon dioxide concentration and annual average ambient temperature at the bridge site also exert considerable positive impacts on carbonation depth, while the remaining factors impose marginal effects.