完全ASTFA方法及其在转子碰摩故障诊断中的应用

摘要

在进行自适应最稀疏分析(Adaptive and Sparsest Time-Frequency Analysis,ASTFA)时需要人为地选定分量带宽控制参数λmax,λmax选用不当会引起模态混淆,针对ASTFA的这种缺陷,利用总体平均经验模态分解(Ensemble Empirical Mode Decomposition,EEMD)的自适应性,提出完全自适应最稀疏分析(Complete Adaptive and Sparsest Timefrequency Analysis,CASTFA)方法.CASTFA以EEMD分解出来的分量为参考,自适应地确定λmax.仿真结果结果表明CASTFA能够抑制ASTFA方法的模态混淆现象,且分解效果优于EEMD方法.将CASTFA方法应用于转子碰摩故障诊断,结果表明了该方法的有效性.%λmax parameter of controlling a component's bandwidth,must be determined manually when using adaptive and sparsest time-frequency analysis (ASTFA).An improper λmax will lead to the problem of mode mixing.A complete adaptive and sparsest time-frequency analysis (CASTFA) was proposed to restrain such a drawback of ASTFA.CASTFA uses the intrinsic mode function (IMF) generated by ensemble empirical mode decomposition (EEMD) as reference to adaptively determine the λmax The simulation example shows that the CASTFA can restrain the mode mixing of ASTFA,and performs better than EEMD in restraining mode mixing.The practical rotor rubbing fault diagnosis result proves that CASTFA is efficient.