以齿轮箱为研究对象,采用以DSP-TMS320F28335为主控芯片的嵌入式故障诊断中心模块和以ZigBeeCC2530为主控芯片的无线通信模块,开发了一种实时在线嵌入式远程故障诊断系统。该远程故障诊断系统中神经网络模型采用基于Matlab模型的设计,对采集的样本信号进行预处理(滤波和和FFT变换),并对经过预处理的信号提取特征值。利用Matlab神经网络工具箱进行网络训练,提取阈值和权值,在Simulink中搭建神经网络模型,并将该模型嵌入到DSP的嵌入式系统,对故障进行诊断识别分类。DSP嵌入式故障诊断中心模块和ZigBee无线通信模块两部分是相互独立的,通过串口接口SCI进行数据发送。该远程故障诊断系统具有信号采集、数据存储、故障识别分类以及无线通信等功能。
Seen the gear box as the research object,based on DSP-TMS320F28335 as main control chip embedded fault diagnosis center module and ZigBee- CC2530 as main control chip wireless communication module,an online real-time embedded remote fault diagnosis system was developed. The neural network model of the remote fault diagnosis system was based on Matlab model design,and the collected samples of the signal was preprocessed( filter and FFT transform),and the characteristic value was extracted from the pre-processed signal. Taking advantage of the Matlab neural network toolbox to train network,extract threshold and weight,neural network model was built in Simulink and it was embedded in DSP embedded system to classify and recognise fault diagnosis. The DSP embedded fault diagnosis center module and the Zig Bee wireless communication modules are independent each other,they send data through a serial port interface SCI. The remote fault diagnosis system has the functions of signal acquisition,data storage,fault identification,fault classification,and wireless communications.