【摘要】 曳引系统是电梯的核心部件之一，若其在运行过程中产生故障可能会导致重大生命财产损失，文中针对该问题提出了融合感知层、边缘处理层和云服务层的监测系统和基于长短期记忆网络（LSTM）的故障诊断方法，实现了电梯曳引系统的在线故障监测和诊断。通过对某服役电梯进行现场测试，结果显示该技术方法对电梯曳引系统的诊断响应时间平均约0.03 s，准确率高达95.89%，较传统的人工巡检方式具有更好的及时性和诊断准确率，为电梯这类特种设备的实时在线健康管理提供了技术基础。更多还原

【Abstract】 The traction system is a critical component of an elevator, and its failure during operation can result in significant loss of life and property. To address this, a comprehensive monitoring system that integrates a perception layer, edge processing layer, and cloud service layer, is proposed, along with a fault diagnosis method based on the Long Short-Term Memory network(LSTM). This system enables real-time fault monitoring and diagnosis of the elevator traction system.Field test results from an in-service elevator demonstrate that the average diagnostic response time for the traction system is approximately 0.03 seconds, with an accuracy rate of up to 95.89%. These figures exceed the efficiency and accuracy of traditional manual inspection methods, providing a technical basis for the real-time online health management of critical equipment like elevators.更多还原