小波分析用于机械系统突发信号在线检测

近年来引起各领域广泛关注的小波分析理论，以其良好的时频局部化功能提供了一种瞬态信号的分析方法，但单纯基于此算法的软件往往缺少对于突发性信号的自适应能力。神经网络则具有良好的自适应性、自组织性及很强的学习功能。本文将此二者结合，以神经网络原理和现代数学小波分析为依据，提出了基于神经网络思想的小波分析；改进原有算法，以自编C程序识别工程技术测量中遇到的突发故障与噪声，并在强噪声环境中机械系统的突发信号实测中获得成功。

Wavelet Analysis for Online Detection of Out-bursting Signals in Mechanical System

Wavelet transformation provides a way to detect out bursting signals, but the software simply based on this algorithm is usually lack of self adaptability when meeting out bursting noise. Neural network has good ability of self adapting, self organizing and learning. On the basis of modern wavelet analysis and the principles of neural network, this paper proposes a method to recognize the type and specification of random signal automatically, so that the original algorithm can be improved to identify the out bursting noise detected in engine ering measurement. The sensor output from the dual cantilever beam apparatus under an out bursting excitation was analyzed successfully by the Turbo C programs.