自适应最优化窄带分解方法及其应用

Adaptive optimization narrow-band decomposition method and its application

提出了自适应最优化窄带分解(Adaptive Optimization Narrow-Band Decomposition,AONBD)方法。该方法将信号分解转化为对滤波器参数的优化问题,以得到信号的最优化解为优化目标,在优化过程中将信号自适应的分解成多个内禀窄带分量(Intrinsic Narrow-Band Components,INBC)。AONBD分为两步,首先通过优化得到最优的滤波器,然后使用该滤波器对信号进行滤波以得到信号的最优化解。阐述了AONBD的基本原理及分解步骤。采用仿真信号将AONBD方法与自适应最优化时频分析(Adaptive Sparsest Time-Frequency Analysis,ASTFA)方法及经验模态分解(Empirical Mode Decomposition,EMD)方法进行对比。结果表明,AONBD在抑制端点效应和模态混淆、抗噪声性能、提高分量的正交性和准确性等方面具有一定的优越性。对转子振动信号的分析结果表明,AONBD能有效应用于机械故障诊断。

The adaptive optimization narrow-band decomposition （AONBD ） method was proposed. Signal decomposition was converted into optimizing parameters of a filter.The optimization objective was to obtain the optimal solution of signals.An original signal was adaptively decomposed into several intrinsic narrow-band components（INBC） via optimization.AONBD method had two steps.Firstly,the optimal filter was obtained with optimization.Secondly,the optimal solution was derived by filtering the original signal using the optimal filter.The basic theory and decomposition steps of AONBD were described.Comparisons were made among AONBD,the adaptive sparsest time-frequency analysis （ASTFA）and the empirical mode decomposition （EMD）by utilizing a simulated signal.The results showed that the AONBD method is superior to the other two methods in restraining end effects and mode mixing,anti-noise performance, and improving the orthogonality and accuracy of components.The AONBD method was applied to analyze vibration signals of a rotor.The results indicated that AONBD can be effectively applied to diagnose mechanical faults.