论文部分内容阅读
提出了一种基于小波神经网络的切削刀具故障监测方法，即提取反映刀具磨损状态的多源特征参数，利用小波神经网络的非线性模型和学习机制，实现在线状态监测；同时针对故障诊断的多输入输出问题带来的网络规模增大、收敛速度慢等问题，提出一种网络优化算法，即采用尺度参数的自适应调整法及平移参数的寻优搜索法，寻找最优小波基元，从而简化小波网络并加快收敛，仿真实例证明了该方法的有效性。
A method of cutting tool fault monitoring based on wavelet neural network is proposed, that is to extract multi-source feature parameters that reflect tool wear status, and to realize online condition monitoring by using wavelet neural network’s nonlinear model and learning mechanism. Input and output problems brought about by the network size increases, the convergence speed is slow and so on, a network optimization algorithm, that is, the use of adaptive scaling method and the translation parameter search algorithm to find the optimal wavelet primitive Simplify the wavelet network and speed up the convergence. The simulation example proves the effectiveness of the method.