Helical idler gear transmission systems can adapt to high-speed, heavy-load working environments and are thus widely used in aerospace, shipbuilding, and other heavy industry sectors. Root crack is one of the common fault types. Prior studies generally only considered cracks at a single working flank, lacking comparative analysis between the crack at the working flank and the non-working flank. This paper examines the dynamic response of helical idler gears with root cracks at different working flanks, comparing dynamic response differences between working and non-working flank cracks. First, a comprehensive dynamics model of the helical idler gear system is established. Second, the influence of root crack location (the working flank or the non-working flank) on time-varying meshing stiffness is considered based on potential energy method, and a flexible model is established by finite element method for the faulty gear. Finally, solution results of the rigid-flexible coupling dynamics model are analyzed. The dynamic response signal characteristics of root cracks at the working flank and the non-working flank are analyzed in time domain, frequency domain and time frequency domain, respectively. Corresponding experiments are designed based on the FZG experimental platform, and the experimental results are in good agreement with the simulation results, which verified the accuracy of the model.
Accurate fault diagnosis of gear transmission systems is crucial for ensuring mechanical reliability and preventing catastrophic failures. However, existing research predominantly focuses on single-gear crack faults, often overlooking the complex coupling effects when cracks occur simultaneously on meshing gears in practical engineering scenarios. To address this research gap, a multi-degree-of-freedom dynamic model incorporating time-varying mesh stiffness under normal, single-crack, and coupled-crack conditions is established. Experimental validation is conducted based on an FZG closed test rig for power flow. The results indicate that the mesh stiffness under coupled-crack conditions is generally lower than that under single-crack conditions. In the time-domain vibration response, the periodic impact amplitudes induced by coupled cracks are significantly amplified, with the impact period jointly influenced by the rotational speeds of both the driving and driven gears. According to frequency-domain analysis, coupled cracks result in a notable increase in harmonic peaks of the mesh frequency, enhanced sideband amplitudes, and a modulation period that is between the rotational frequencies of the driving and driven gears. The simulation results from the dynamic model show high consistency with the experimental signals in terms of time-frequency characteristic trends and time-domain indicators such as the crest factor, thereby validating the effectiveness of the dynamic model. This study elucidates the unique influence mechanism of coupled cracks on the dynamic behavior of gear systems and can provide theoretical guidance for the accurate diagnosis and condition assessment of multi-tooth faults in subsequent research.
11 December 2024 in IEEE Transactions on Intelligent Transportation Systems
In autonomous driving, efficiently and accurately recognizing taillight states using dashcams is essential for interpreting other vehicles’ intentions. Recent video-based taillight recognition methods outperform earlier image-based approaches. However, they face challenges in efficiently integrating spatiotemporal information and managing high computational costs. In this paper, we introduce ETR, an accurate and efficient Transformer-based video classification model designed to enhance taillight recognition. Specifically, we first design a lightweight backbone to extract temporal and spatial features from videos and generate queries with prior information. Next, we develop a hierarchical progressive Transformer decoder that integrates feature maps from different levels of the backbone to enhance the model’s global information. Finally, we employ a classification head to predict the taillight state of the video. Additionally, we introduce a public dataset, ETR-Taillights, to address the current lack of open datasets for vehicle taillight recognition. The dataset contains 28,799 dashcam video clips, making it the largest public taillight recognition dataset. Experiments show that our method achieves a 91.69% F-measure on the ETR-Taillights dataset, surpassing the latest taillight recognition methods, VIF by 6.94% and CNN-LSTM by 10.82%. Additionally, it achieves an inference speed of 45.06 FPS, being 3.6 times faster than VIF. Furthermore, we conduct real-world road tests to demonstrate our method’s robustness and effectiveness in practical scenarios. Our model and dataset are available at https://github.com/yang590/ vehicle-taillight.
12 August 2026 in Lasers in Manufacturing and Materials Processing