基于代数网络和混沌参数的碰摩声发射源定位方法研究

针对时差定位法受不同模式波速度差异及波形传播畸变等因素影响的问题,将神经网络技术应用到声发射源定位中.为了克服传统BP算法训练时间长和精度不够高的缺点,提出代数神经网络概念,在网络训练阶段引入代数算法.将复杂的非线性优化问题转化为简单的代数方程组求解问题,直接获得最优点,大大缩短了网络的学习时间.同时作为定位特征研究分析了转子碰摩声发射信号的非线性动力学特性,提出了关联维数、最大Lyapunov指数和Kolmogoroy熵等声发射源的非线性动力学新特征,并将其作为神经网络的输入定位特征.实验结果表明,利用这些声发射源的非线性动力学特征和神经网络结合能较好地解决了碰摩声发射源定位问题,为转子碰摩故障诊断提供依据,具有良好的应用前景和进一步研究的价值.

Research on localization of acoustic emission source based on algebraic neural network and chaotic features

Due to defects of time-difference of arrival localization which influenced by speed differences of various model waveforms and waveform distortion in transmitting process,a neural network technique was introduced to calculate localization of the Acoustic Emission(AE) source.In order to overcome the shortcomings of the traditional BP algorithm such as long training time and low accuracy,we propose the concept of algebraic neural network and introduce the algebraic algorithm in the network training phase whic...