基于STM32的电机轴承故障诊断系统的研究与设计

【作者】 叶丹；

【导师】 姚凯学；

【作者基本信息】 贵州大学 ， 电子信息， 2023， 硕士

【摘要】 电机作为现代工业和社会生产生活中重要的动能设备,其应用十分广泛。滚动轴承作为电机中最关键的零件之一,其故障会严重影响电机的正常运行和使用寿命。目前轴承故障诊断方法存在以下问题:传统基于信号处理技术的轴承故障诊断方法存在先验知识要求高、依赖专家经验等问题;基于深度学习的轴承故障诊断模型存在尺寸大、参数量多等问题;主流轴承故障诊断系统存在数据上传量大、诊断速度慢等问题,无法满足对故障诊断响应要求高的实际工业场景。针对以上问题,本文主要的研究工作和成果如下:(1)研究了轻量化轴承故障诊断模型。该诊断模型基于卷积神经网络,通过使用深度可分离卷积替换标准卷积,并使用全局平均池化层替换全连接层作为模型的输出单元,有效地减少模型的参数量从而降低模型占用的内存,为模型能够移植嵌入式设备提供保障。(2)设计了基于嵌入式设备的轴承故障诊断系统。该系统由STM32微控制器作为主控平台,其上部署了轴承故障诊断模型,振动采集模块采集轴承振动数据,经过数据格式重构后输入诊断模型,在LCD屏显示实时的故障诊断结果,蜂鸣器报警模块实现轴承故障警示功能,Wi Fi无线通信模块通过将轴承故障数据上传至物联网平台,以Web界面的形式实现数据展示和远程监测。实验结果表明,该系统具有较好的轴承故障诊断精度,且诊断速度快。便携式嵌入式设备易部署在工业现场,采集的数据存储在设备,免去数据上传后台或云端过程,满足工业的轴承故障诊断实时性需求。更多还原

【Abstract】 As an important kinetic energy equipment in modern industry and social production and life,electric motor is widely used.Rolling bearing is one of the most critical parts in the motor,its failure will seriously affect the normal operation and service life of the motor.At present,bearing fault diagnosis methods have the following problems: traditional bearing fault diagnosis methods based on signal processing technology have high prior knowledge requirements,dependence on expert experience and other problems;The bearing fault diagnosis model based on deep learning has some problems,such as large size and large number of parameters.The mainstream bearing fault diagnosis system has problems such as large data upload volume and slow diagnosis speed,which cannot meet the requirements of high fault diagnosis response in actual industrial scenarios.In view of the above problems,the main research work and achievements of this thesis are as follows:(1)This thesis studies the lightweight bearing fault diagnosis model.This diagnostic model is based on convolutional neural network.By using depthwise separable convolution to replace standard convolution,and using global average pooling layer to replace the fully connected layer as the output unit of the model,the number of parameters of the model is effectively reduced,thus reducing the memory occupied by the model,and providing guarantee for the model to be able to transplant embedded devices.(2)This thesis designs the bearing fault diagnosis system based on embedded equipment.The bearing fault diagnosis model is deployed on the STM32 Microcontroller Unit as the main control platform.The vibration acquisition module collects the bearing vibration data and inputs the diagnosis model after data format reconstruction.The real-time fault diagnosis results are displayed on the LCD screen,and the buzzer alarm module realizes the bearing fault warning function.The Wi Fi wireless communication module can realize data display and remote monitoring in the form of Web interface by uploading bearing fault data to the Internet of Things platform.The experimental results show that the system has a good bearing fault diagnosis accuracy,and the diagnosis speed is fast.Portable embedded equipment is easy to deploy in the industrial field,and the collected data is stored in the equipment,eliminating the process of data upload in the background or cloud,and meeting the realtime requirements of bearing fault diagnosis in the industry.更多还原