Scholarlink

<p>Vehicle-mounted relays are crucial components in railway control systems, and their operational status has a significant impact on the normal functioning of the entire railway system. Addressing the current issues of insufficient extraction and utilization of feature parameters for predicting the remaining useful life (RUL) of vehicle-mounted relays and the inability to predict RUL accurately, this paper proposes a prediction method based on an NRBO-CNN-xLSTM model. First, degradation data throughout the entire life cycle is acquired via a full-life test platform for vehicle-mounted relays. Feature parameters reflecting their operational status are extracted from this data and used as input samples for the prediction model. The Newton–Raphson-Based Optimizer (NRBO) is employed to optimize the hyperparameters of the Convolutional Neural Network (CNN) and the extended Long Short-Term Memory network (xLSTM). Finally, the RUL prediction for the railway vehicle-mounted relay is accomplished using the CNN-xLSTM model. Experimental results demonstrate that compared to a single CNN model, a transformer model, a xLSTM model, a CNN-xLSTM model, and a PSO-CNN-xLSTM model, the proposed RUL prediction model can fully utilize the contextual information from full-life-cycle time-series data more effectively, significantly improving prediction accuracy and greatly enhancing the precision of remaining life prediction.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Remaining useful life prediction of railway vehicle-mounted relays based on NRBO-CNN-xLSTM model

  • 期刊论文

  • 发表日期：

    2026年3月20日

查看全文
  • Wang Yao,
  • Shuixin Liu,
  • Yankai Li,
  • Zhibin Ma,
  • Yufan Chai

摘要

Vehicle-mounted relays are crucial components in railway control systems, and their operational status has a significant impact on the normal functioning of the entire railway system. Addressing the current issues of insufficient extraction and utilization of feature parameters for predicting the remaining useful life (RUL) of vehicle-mounted relays and the inability to predict RUL accurately, this paper proposes a prediction method based on an NRBO-CNN-xLSTM model. First, degradation data throughout the entire life cycle is acquired via a full-life test platform for vehicle-mounted relays. Feature parameters reflecting their operational status are extracted from this data and used as input samples for the prediction model. The Newton–Raphson-Based Optimizer (NRBO) is employed to optimize the hyperparameters of the Convolutional Neural Network (CNN) and the extended Long Short-Term Memory network (xLSTM). Finally, the RUL prediction for the railway vehicle-mounted relay is accomplished using the CNN-xLSTM model. Experimental results demonstrate that compared to a single CNN model, a transformer model, a xLSTM model, a CNN-xLSTM model, and a PSO-CNN-xLSTM model, the proposed RUL prediction model can fully utilize the contextual information from full-life-cycle time-series data more effectively, significantly improving prediction accuracy and greatly enhancing the precision of remaining life prediction.