The Pantograph-catenary system is crucial for transferring electrical power from catenary lines to electrified train and the occurrence of arcing could damage railway operations, thus, it is important to detect arcing. Detecting arcing in complex scenes and detecting different size and shape of arcing is still a challenge. To overcome these issues, a robust image-based semantic segmentation model named ArcMSF is proposed for arcing detection of pantograph-catenary, which designs a novel hybrid multi-scale feature fusion model that aggregates Transformer with CNN to realize arcing pixel segmentation. A down-top decoder for combining low-level features with high-level features is designed to achieve multi-scale level arcing feature detecting in complex scenes. Inspired by the arcing image properties that arcing is always the brightest, the global max features and global threshold features are designed to augment the arcing features. Experiments on dataset IVAIS-PCA2021 and comparative experiments are conducted to demonstrate the effectiveness of the ArcMSF, which can achieve 89.13% segmentation accuracy and the fast inference speed of 23.84ms. Moreover, the detection results have a clear depiction of edge details.

Stochastic process models have been widely used in the field of remaining useful life (RUL) prediction due to their notable generality and high interpretability. Traditional methods typically employ a single degradation function to describe the degradation of an equipment. However, the degradation process of a tool exhibits multi-stage characteristics. A single degradation model ignores this characteristic and is challenging to accurately track the tool wear process. To address this issue, this paper proposes a multi-source state space based tool RUL prediction method considering multi-stage degradation characteristics. First, the mapping between multi-source monitoring signals and tool degradation states is established based on the multi-source state space model. Then, a multi-stage degradation model based on the Wiener process is established to describe the tool wear degradation. The method fully integrates the historical wear information and the real-time multi-source monitoring signals during tool machining to achieve the online RUL prediction. Through a detailed numerical simulation analysis and a tool wear experiment the proposed method is verified with better comprehensive performance. The average cumulative relative accuracy of the prediction results is 0.789, which outperforms the commonly used tool RUL prediction method.

In view of the serious sliding electrical contact performance caused by external vibration and environmental contaminant, a study on the tribological characteristic and contact resistance of Cu alloy was conducted using a self-developed micro-load reciprocating electric contact device. Various glue concentrations (0%, 10%, 30%, and 50%) were prepared with anhydrous ethanol and deposited on the surface of a pure copper block via the deposition method. An external vibration source was installed on the sliding module to achieve vertical vibration. The results indicate that the final contact resistance and coefficient of friction (COF) in direct metal contact are about 0.01 Ω and 0.3, respectively. At this time, the wear volume is 2 to 3 orders of magnitude higher than the condition with glue residual. As glue concentration is above 10%, residual glue on the surface of Cu alloy hinders efficient contact between friction pairs, resulting in higher contact resistance. Glue exhibits lubrication, anti-wear, and insulation properties. External vibration causes friction pairs to briefly separate, leading to a lower glue removal capacity than that under non-vibration conditions. The contact resistance with glue addition under vibration conditions is higher than that under non-vibration conditions at 3 × 10⁴ cycles. The dominant oxide product is CuO, which has a limited effect on contact resistance.