Decoupling the influence of operating factors such as temperature, humidity, and back pressure is essential for improving fuel cell efficiency. This study employs a regression model combining Locally Weighted Scatterplot Smoothing (LOESS) with the Sobol index method to analyze the effects of two major parameter categories: temperature and humidity, and stoichiometry and back pressure, on fuel cell output voltage. Using limited experimental data, it uncovers complex interactions and sensitivities among these factors. The results indicate that the stack temperature has the greatest impact, accounting for 60%, with minimal influence from the changes in the current density. Cathode humidity impacts output voltage by about 10%, while anode humidity accounts for approximately 5%. Significant interactions between temperature and both anode and cathode humidity contribute around 9% each. Among stoichiometries, cathode stoichiometry has the largest impact, exceeding 50% at low current density and growing to over 70% as current increases. Back pressure and anode stoichiometry each have an impact of around 10%, with minimal mutual influence. This study highlights the extent to how different operating parameters influence fuel cell performance, offering valuable insights for optimizing fuel cell operating conditions.

This study investigates the static stability and flow field of a cruising Electric Vertical Take-off and Landing (eVTOL) aircraft with a compound-wing configuration using Lattice Boltzmann Method (LBM). Numerical accuracy of Finite Volume Method (FVM) and LBM is validated and comparatively analyzed based on isolated-rotor simulations. The LBM-based simulations are 8–9 times faster than those based on FVM for weakly compressible flows, which exhibits a significant advantage in computational efficiency and cost. The maximum error of thrust is ∼ 11% for LBM-based simulations and that of torque is ∼ 7%, which is similar to that of FVM. Unsteady simulations based on LBM are conducted for the whole eVTOL aircraft, and flow field structures and static stability are further analyzed. The results reveal that four varieties of vortical structures dominate the flow field, including wingtip and root vortices, vortices generated by static lift rotors, and vortices originating from the counter-rotating tail-propeller pair, varying with angle of attack and angle of sideslip. Their developing mechanisms and mutual interactions are examined in detail. The aircraft is unstable in terms of longitudinal and directional static stabilities, while it is stable in lateral, enhanced by wingtip vortices. Moreover, tail-mounted propulsion places higher demands on thrust propellers. In addition, vortices generating at the leading edges of the upper lift rotors delay separation on the wing upper surface but imposes negative effects on both longitudinal and lateral stability. The high computational efficiency, accurate flow prediction, and realizable aerodynamic forces demonstrated by LBM provide engineering insight for future aerodynamic performance analysis and development of eVTOL aircraft.

Low-altitude unmanned aerial vehicle (UAV)-based traffic object detection is challenged by substantial scale variations from aerial perspectives, the extremely small pixel proportions of distant traffic participants, complex road background interference, unstable illumination, and severe occlusion in dense traffic scenes. To address these problems, this paper proposes ACP2-YOLO, an improved YOLO11-based detection framework for low-altitude UAV traffic scenarios, with the goal of enhancing the detection of vehicles, pedestrians, and non-motorized traffic participants. The proposed framework introduces two key improvements. First, a lightweight hybrid ACmix module that integrates convolution and self-attention is embedded into the network, enabling the model to jointly capture local detailed features and global contextual dependencies and thereby strengthen feature representation under complex backgrounds. Second, a P2 small-object detection layer is added to the original three-scale detection structure of YOLO11 to construct a four-scale P2–P5 feature pyramid. By allowing shallow high-resolution features to directly participate in object prediction, this design effectively reduces spatial information loss caused by deep downsampling and improves small-object perception. Experiments on the VisDrone2019 dataset show that the improved model achieves 53.1% Precision, 41.1% Recall, 42.9% mAP@50, and 26.3% mAP@50–95, outperforming the baseline YOLO11 by 4.2, 4.2, 5.0, and 3.6 percentage points, respectively. Comparisons with mainstream YOLO-series detectors further demonstrate its superior overall accuracy, small-object detection capability, and adaptability to complex scenes, indicating its potential for UAV-based traffic monitoring, road safety inspection, and intelligent transportation perception.

1. Rapid and reliable unmanned aerial vehicle (UAV)-based person detection is a key capability for post-disaster search-and-rescue and emergency surveillance. In real deployments, UAV imagery is of...