基于机械故障模拟实验台的多通道故障诊断实验设计  

Experimental design of fault diagnosis based on mechanical equipment simulation platform

作　　者：王刚 张加斯 张晓光 于洪珍 WANG Gang;ZHANG Jiasi;ZHANG Xiaoguang;YU Hongzhen(School of Information and Control Engineering,China University of Ming and Technology,Xuzhou 221000,China)

机构地区：[1]中国矿业大学信息与控制工程学院,江苏徐州221000

出　　处：《实验技术与管理》2022年第12期62-68,共7页Experimental Technology and Management

基　　金：十三五国家重点研发计划(2017YFC0804404);教育部产学合作协同育人项目(202002280003);中国矿业大学教改项目(2021ZX04)。

摘　　要：该文设计了一个利用机械故障模拟实验台和Python开发的故障诊断实验方案。实验台工作时,利用拾音器构成的拾音器阵列采集滚动轴承声学信号并传输到计算机,并用Python语言编码代码来实现多通道滑窗采样、提取多通道声学信号的特征和识别故障类型。通过Transformer的编码器中的多头自注意力机制提取多通道声学信号中隐含的空间特征。实验结果表明:故障识别率达100%。该文详细介绍了实验的基本原理、具体步骤、所设计的诊断算法,以帮助学生掌握故障诊断流程。This paper designs a fault diagnosis experiment project developed by using mechanical fault simulation experiment platform and Python. When the test-bed is working, the acoustic signal of rolling bearing is collected by the pickup array composed of pickup and transmitted to the computer, and the code is encoded in Python language to realize multi-channel sliding window sampling, extract the characteristics of multi-channel acoustic signal and identify the fault type. The spatial features implicit in multi-channel acoustic signals are extracted through the multi head self-attention mechanism in Transformer’s encoder. Experiments show that the fault recognition rate reaches 100%. This paper introduces the basic principle, specific steps and the designed diagnosis algorithm of the experiment in detail to help students master the fault diagnosis process.

关 键 词：机械故障模拟实验台 TRANSFORMER 层归一化 PYTHON 故障诊断 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]