摘要
针对传统电梯故障检测依赖专家知识、易丢失传感器特征信息等问题,笔者提出一种深度自编码器模型,用于从原始传感器数据中自动提取深度特征。同时,基于随机森林算法对提取的深度特征进行分类,进而实现故障检测。实验结果表明,与现有特征检测方法相比,该方法能够有效避免误报,有助于实现电梯系统自动化故障检测。
Aiming at the problems that traditional elevator fault detection depends on expert knowledge and is easy to lose sensor feature information, a depth self encoder model is proposed to automatically extract depth features from the original sensor data. At the same time, the extracted depth features are classified based on the random forest algorithm, and then the fault detection is realized. The experimental results show that compared with the existing feature detection methods, this method can effectively avoid false positives and is helpful to realize automatic fault detection of elevator system.
作者
高鹏
GAO Peng(Beijing Institute of Special Equipment Inspection and Testing,Beijing 100029,China)
出处
《信息与电脑》
2022年第8期105-107,共3页
Information & Computer
