EMD方法基于径向基神经网络预测的数据延拓与应用

APPLICATION OF EMD METHOD WITH DATA EXTENSION TECHNIQUE BASED ON RBF NEURAL NETWORK TO TIME-FREQUENCY ANALYSIS

把基于径向基神经网络(radbas function,RBF)预测的数据延拓技术引入经验模态分解(empirical mode decompo-sition,EMD)时频分析领域,论述基于RBF神经网络预测的数据延拓技术原理,通过对非线性仿真信号基于RBF神经网络预测延拓研究表明,该延拓技术是有效的,并且把该延拓技术应用于转子横向裂纹的时频分析,获得良好的效果。该研究成果能广泛用于信号时频分析领域。

The data extension technique based on the radbas function（RBF） neural network prediction is introduced to time-frequency analysis by the empirical mode decomposition（EMD） method. The data extension technique applied to a simulation signal is explained in detail. EMD time-frequency analysis with the data extension technique on vibration signal of a deeply cracked rotor is introduced and the results show that the method is helpful to detect the phase-modulation caused by torsional vibration of the rotor. The suggested method can be applied widely to time-frequency analysis field.