基于数据挖掘的实验室设备故障在线检测方法  

On-line Detection Method of Laboratory Equipment Fault Based on Data Mining

作　　者：赵常军 常贤龙 ZHAO Changjun;CHANG Xianlong(Shaanxi Weifeng Nuclear Electronics Co.,Ltd.,Xi’an 710199)

机构地区：[1]陕西卫峰核电子有限公司,西安710199

出　　处：《现代制造技术与装备》2024年第11期82-84,88,共4页Modern Manufacturing Technology and Equipment

摘　　要：针对实验室机械设备的主要故障类型,提出一种基于数据挖掘的在线故障检测方法。该方法通过融合多源传感器数据,提取设备状态特征,利用卷积神经网络算法识别异常模式,结合知识图谱进行故障诊断推理。实验结果表明,该方法在轴承故障诊断任务上取得了95.5%的故障分类准确率,故障原因和位置诊断准确率超过90%,为实现实验室设备的智能运维提供了新思路。Aiming at the main fault types of laboratory machinery equipment,an online fault detection method based on data mining is proposed.The method fuses multi-source sensor data to extract device state features,uses convolutional neural network algorithm to identify abnormal patterns,and combines knowledge graph to perform fault diagnosis reasoning.The experimental results show that the fault classification accuracy of the proposed method is 95.5%in the task of bearing fault diagnosis,and the fault cause and location diagnosis accuracy is more than 90%,which provides a new idea for realizing intelligent operation and maintenance of laboratory equipment.

关 键 词：实验室设备 故障 在线检测 数据挖掘 

分 类 号：TH17[机械工程—机械制造及自动化] TP311.13[自动化与计算机技术—计算机软件与理论]