In electric railways, the interaction performance of the pantograph–catenary systems (PCS) is crucial for maintaining a stable current supply. Establishing high-fidelity numerical models based on the finite-element method (FEM) is a common practice, but it also has substantial computational complexity. The Koopman operator, a promising candidate for data-driven modeling, provides a global linear representation of nonlinear dynamic systems. In this article, we develop a novel generalized Koopman neural operator (GKNO) implemented by an autoencoder and an improved Transformer for modeling complex nonlinear dynamic systems with large-scale degrees of freedom. It consists of an observable function, an evolution function, and an invertible observable function. First, the encoder, as the embedding model, maps the state variables of the original system into observable space with linear dynamics. Then, an improved Transformer model is proposed to learn the evolution function in the embedding space based on an autoregressive task. Finally, the decoder reconstructs the state variables of the original system from the embedding space. Experimental results on the PCS model demonstrate that GKNO can capture the intrinsic evolution patterns to represent high-dimensional and nonlinear PCS dynamics, significantly reducing computational complexity and solution time. Comparative experiments show that GKNO achieved considerable solution accuracy with negligible consumption of computing resources, providing a promising potential for parameter optimization and pantograph hardware-in-the-loop (HIL) test rigs.

As an important component of the catenary system in electrified railways, the operating status of the catenary support components (CSC) is related to the operational safety of the railway. They are usually detected by a camera on the inspection equipment using computer vision techniques. However, current CSC detection research faces the following challenges. 1) The components in the captured images have issues of multi-scale and multiple categories; 2) Some components are covered because of the limited shooting angle; 3) The existing detectors based on CNN and Transformer are difficult to associate features between long-interval regions. To solve these problems, a novel graph-guided Mamba-DETR detector (Graph-MDETR) is proposed for detecting CSC. Firstly, a new backbone network (DAMamba-Adapter) is proposed to enhance the model’s ability to associate features between long-interval regions, and multi-scale training (MST) is used for solving multi-scale problems. Secondly, the semantic graph-guided module (SGM) and positional graph-guided module (PGM) are designed to alleviate the problems of occlusion and multi-category components by extracting prior category relationships with global semantic features and component geometric information. Finally, a CSC dataset taken by a UAV is established, and various experiments are conducted to verify the effectiveness and superiority of the proposed method.

The existing drone-based detection systems for electrified railway catenary support components face the following challenges: 1) the background of aerial images of the catenary support component is complex; 2) a unified model struggles to handle anomaly detection for various component types; and 3) there is inconsistency in the normal region feature information between training and testing images. To address these issues, this article proposed a novel adaptive anomaly detection framework for catenary support component an anomaly detection model based on invariant normal region prototype extraction (INRP-Ader). First, we proposed a new segmentation model (CSC-SAM) that embeds key catenary component location information to extract foreground images. Next, we design an anomaly detection model that directly extracts invariant normal region prototype (INRP) features from the test images. This model includes an INRP extractor constrained by INRP smoothness loss and an INRP-guided decoder, aimed at solving the problem of inconsistency in normal region features between training and testing images and the challenge of adapting a unified model to multiple component types. In addition, a soft mining loss is introduced to further optimize the training process of the multiclass anomaly detection model. Finally, we established a real-world catenary support component dataset, catenary system component UAV dataset (CSCUD), collected by drones, and achieved detection performance of 99.1/98.2/98.2 in image-level metrics (I-AUROC/I-AP/I-F1) and 96.4/37.1/55.7 in pixel-level metrics (P-AUROC/P-AP/P-F1). The proposed method outperforms traditional methods by approximately 0.4%–16.6% across various metrics.

The core task of intelligent detection in electrified railway catenaries is to detect their supporting components. However, due to the large number and variety of components in catenary inspection images, labeled catenary data is often limited, and few studies have focused on leveraging large amounts of unlabeled datasets in this field. This article proposes a novel self-supervised pretraining model catenary support rod masking-based masked image modeling (CSRM-MIM) for detecting catenary support components (CSCs), which effectively utilizes the valuable information in unlabeled catenary data. Specifically, a new semantic catenary support rod area masking (CSRA masking) strategy based on the CSRA features is proposed to guide the model in learning meaningful catenary semantic information during the pretraining process. Additionally, a new Siamese pretrained network framework is designed, incorporating an information interaction enhancement module (IIEM) and a dual reconstruction network to perform dual reconstruction on the semantic mask, thereby extracting global features in the catenary domain. Finally, a multiscale knowledge distillation (MSKD) strategy optimized by the cross-layer fusion module cross-layer fusion decoding (CLFD) is introduced to assist the self-supervised model in acquiring specific-general representations for the catenary field and transferring multiscale information, which benefits subsequent detection tasks. Experimental results demonstrate that this method significantly enhances the detector’s performance in identifying CSCs, owing to the retained self-supervised learning (SSL) strategy.

In electrified railway systems, the localization performance of catenary support components (CSCs) in service status monitoring systems is crucial for downstream tasks such as anomaly identification and defect detection. However, CSCs detection faces challenges such as multiple categories, multi-scales, and tiny objects, which make it difficult for general object detection algorithms to improve their detection performance effectively. To address these challenges, this paper proposes a novel Anchor-Free Interactive Detector Based on Multi-Receptive Field Mamba (MRFM-IFCOS) to accurately and effectively detect all CSCs. First, a novel lightweight Multi-Receptive Field Mamba (MRFMamba) backbone network is proposed to extract high-frequency edge details and improve feature extraction efficiency by capturing dependencies between long-range regions in images and integrating multi-scale information from different receptive fields, achieving a compelling balance between detection performance and computational efficiency. Then, an Interactive Global Knowledge Distillation (IGKD) strategy is presented to foster more effective knowledge transfer and interaction within student-teacher learning paradigms, enhancing the student network’s feature representation and localization capabilities and reducing computing resource consumption. Finally, a Cross-layer Fusion Feature Pyramid Network (CLF-FPN) is developed to substantially improve multi-level feature fusion and capture fine-grained, high-frequency details crucial for small objects. Experimental results show that the proposed detection framework can achieve a SOTA detection accuracy of 52.52% (mAP) and 19.89% (AP_s) on the CSCs dataset while maintaining low computational complexity.

As an essential subsystem of electrified railway operation and maintenance, intelligent detection of catenary support components still faces several critical challenges: (1) the number of abnormal (negative) samples for components is severely limited; (2) component anomalies are highly diverse and exhibit heterogeneous visual characteristics; and (3) existing models generally show unsatisfactory detection performance when confronted with previously unseen anomaly types. To address these issues, this paper proposes a novel few-shot anomaly detection model for catenary components, termed BCLIP-ADer, built upon a Bayesian prompt contrastive vision–language pretraining framework. Specifically, a Bayesian prompt flow module (PFM) is designed to regularize the text prompt space via the jointly learned image-specific feature distribution (ISFD) and image-agnostic feature distribution (IAFD), thereby mitigating the degradation in detection performance on unseen component anomalies. Monte Carlo sampling over these learned distributions is further employed to generate diverse text prompts, leading to more comprehensive coverage of the prompt space. In addition, a cross-modal feature refinement module (CFRM) is designed to more effectively align dynamic text embeddings with fine-grained image features, thus enhancing anomaly detection at the component level. Finally, extensive experiments conducted on a UAV-based catenary dataset (CSCUD) demonstrate the effectiveness and superiority of the proposed approach. Specifically, the proposed method achieves I-AUROC/I-AP/I-F1_max scores of 94.2/93.2/93.1 under few-shot conditions.