连续参数小波网络在机械状态预报中的应用

Continuous Parameter Wavelet Networks Associated with Mechanical Operating Condition

研究了连续参数小波网络的原理及其构造，通过仿真数据的比较分析，说明了连续小波网络较传统ＢＰ网络有更快的收敛速度和更好的非线性逼近能力．同时，在用连续参数小波网络对实际机组振动峰峰值数据的预报中，收到了满意的效果．

The principle and construction of continuous parameter wavelet networks are studied. Convergence is improved for nonlinear approach as compared with the traditional back propagation networks. Satisfactory results are obtained for the vibratory forecasting of machine sets.