铁道科学与工程学报2025，Vol.22Issue(10)：4697-4711,15.DOI:10.19713/j.cnki.43-1423/u.T20242048
基于数字孪生的转辙机健康状态智能分析方法
Intelligent analysis method for the health status of switch machines based on digital twin technology
摘要
Abstract
To address the limitations of traditional switch machine health management methods in achieving real-time assessments,this study introduced a health status analysis approach based on digital twin technology.The proposed method constructed a digital twin-driven assessment framework that integrates multi-source data fusion and data visualization techniques,enabling real-time monitoring and evaluation of switch machine health.This approach enhanced the health monitoring and intelligent maintenance capabilities of railway signal equipment.A digital twin-driven framework for analyzing switch machine health status was developed,incorporating operational data,environmental monitoring data,and maintenance records.An analysis of common faults and their causes in switch machines was conducted,followed by the application of SMOTE and RFECV-RF algorithms for data feature optimization.A feature-enhanced meta-learning(FEML)model was employed for real-time health status assessment,with performance comparisons conducted against single-model approaches.Experimental results demonstrated that the FEML model's evaluation accuracy exceeded that of single models by at least 1.85 percentage points.Even with the introduction of Poisson noise with amplitude magnitudes ranging from 0.5 to 1.5,the accuracy remained above 97%.Validation using actual power data from S700K switch machines confirmed the method's effectiveness in real-time health monitoring.By integrating life-cycle information and on-site expert experience,the approach supports informed maintenance and replacement decisions.Compared to traditional methods,this method achieved a health status assessment accuracy of 99.07%,satisfying the operational requirements of railway field operations.The digital twin-based health status analysis method significantly enhances assessment accuracy,enabling real-time monitoring and intelligent maintenance.The application of digital twin technology offers a novel solution for switch machine health management and provides a new technological pathway for monitoring and maintaining other critical railway signal equipment.The successful implementation of this method is expected to substantially improve the operational safety and maintenance efficiency of railway signal devices.关键词
转辙机/数字孪生/FEML模型/健康监测/元学习Key words
switch machine/digital twins/FEML model/health monitoring/meta learning分类
交通运输引用本文复制引用
张振海,李哲远,鲍盛成,孙岩..基于数字孪生的转辙机健康状态智能分析方法[J].铁道科学与工程学报,2025,22(10):4697-4711,15.基金项目
中央引导地方科技发展资金资助项目(24ZYQA044) （24ZYQA044）
甘肃省重点研发计划项目(22YF7GA141) （22YF7GA141）
兰州交通大学重点研发计划(ZDYF2303) （ZDYF2303）
