摘要
在键盘击弦机械故障诊断过程中,由于诊断模型的影响,使得故障诊断结果的准确率较低。因此,提出基于BP神经网络的键盘击弦机械故障诊断分析。通过声敏传感器获取键盘击弦机械数据,并利用小波包分解法提取数据特征。基于BP神经网络,构建机械故障诊断模型,并对模型内的各项参数进行计算。最后,以表格的形式呈现出故障诊断分析结果。实验结果表明:在单弦故障诊断时,设计方法与两种常规方法相比,将诊断准确率分别提升了2.57%、5.66%;在多弦故障诊断时,设计方法将机械故障诊断的平均准确率,分别提升了5.76%、7.15%。
In the process of keyboard mechanical fault diagnosis,due to the influence of the diagnosis mode,the fault diagnosis result is not so accurate.Therefore,the fault diagnosis and analysis of keyboard beating machine based on BP neural network is proposed.The mechanical data of keyboard beating is obtained by sensors,based on the neural BP network,the mechanical fault diagnosis model is created and the parameters in the model are calculated.Finally,the results of fault diagnosis and analysis are presented in a table.Research results indicate that:compared to two conventional methods,the proposed method can improve the diagnosis accuracy by 2.57%and 5.66%respectively;In the case of multi string fault diagnosis,the average accuracy of mechanical fault diagnosis is increased by 5.76%and 7.15%respectively.
作者
田甜
杜泽江
TIAN Tian;DU Zejiang(Xianyang Vocational and Technical College,Xianyang Shaanxi 712000,China;School of Information Science&Technology Northwest University,xi’an 710127,China)
出处
《自动化与仪器仪表》
2022年第3期31-35,共5页
Automation & Instrumentation
基金
咸职职业技术学院科研基金项目:“奥尔夫”音乐教学法的应用与研究(2012KB09)。
关键词
BP神经网络
键盘击弦
机械
故障诊断
BP neural network
keyboard string
mechanics
fault diagnosis
