【作者】 李娜；

【导师】 邬再新；

【作者基本信息】 兰州理工大学 ， 机械制造及其自动化， 2008， 硕士

【摘要】 本文对国内外刀具监控技术进行了研究,着重研究了人工神经网络在刀具监控技术中的应用。神经网络因其极强的非线性映射能力,特别适合于复杂模式识别,所以成为刀具状态识别广泛而强有力的工具。本文结合检测系统的三大部分,即信号检测、特征提取、模式识别对刀具状态检测进行了具体的研究。本文的主要研究工作如下:·首先分析了刀具切削状态监控技术的发展及国内外研究的现状,确定了本课题研究的意义。·通过对刀具磨损监测方法的研究,确定了间接监测信号方法的实施,并对刀具加工过程中的各种传感器进行研究,对各种检测信号进行比较,最后选择了电机电流信号,为成功研究刀具的状态奠定了坚实的基础。·对所选信号进行分析并提取信号特征,分别在时域,频域及时频域中对信号进行分析,提出用小波理论分析电机电流信号,有效的提取了刀具的信号特征,为刀具状态的诊断提供可靠依据。·结合了人工智能神经网络对BP网、RBF网和小波神经网络进行研究,并在Matlab中进行仿真,比较了各种网络在刀具状态模式识别决策中的应用。最后结合本课题的结论,提出了改进意见和方案。更多还原

【Abstract】 In this thesis,the tool monitoring and control technology of both oversea and domestic has been studied,more focus on the application of the Artificial Neural Networks in the tool monitoring and control technology.Due to the extremely high non-linear mapping capability of Neural Networks,it is especially suitable for the identification of the complex pattern recognition,which has therefore become a widely used and effective measure for identifying the status of the tools.Hereinafter,the tool monitoring has been investigated based on the three main parts of the detecting system, i.e.signal detecting,character extract and pattern recognition.The main content of the thesis is as follows:·Analyse the development of tool condition monitoring and its status on native and overseas,illustrate research significance of this title.·Do research on tool wear monitoring and select indirect methods.Study all kinds of sensors adopted during tool machining.By comparing all sorts of detecting signals,the current signal of the motor was extracted.It has laid a solid foundation for successfully study the status of the tools.·Analyze the selected signals and extract the signal characters,analyzing the signals in the time domain,frequency domain and the time frequency domain.It has been raised that the wavelet theory should be adopted for analyzing the current signals.It enables extracting the signal characters of the tools effectively and provides the reliable reference for diagnosing the status of the tools.·Study BP net,RBF net and wavelet Neural Networks with the artificial intelligence networks and do simulation with Matlab,comparing the different networks which have been applied for the recognition of the status patterns of tools.At last,the recommendation and plan for improvement has been raised based on the conclusion of the thesis.更多还原