为了更加准确地提取高炉煤气余压透平发电机组(TRT)振动故障信息,提出了基于局部均值分解(LMD)的全矢包络技术。该方法采集TRT转子某一截面上互相垂直的2个振动信号,利用LMD分别将2个信号分解为若干乘积函数(PF)分量。对2个信号的PF分量的包络函数分别进行融合,得到全矢包络谱,利用其对TRT进行故障诊断。仿真和实际算例分析结果表明,较之单源信息分析方法,所提方法获取的故障特征更全面、准确。
The full vector envelope technique based on LMD(Local Mean Decomposition) is proposed to more accurately extract the vibration fault information of TRT(Top gas pressure Recovery Turbine unit),which acquires two mutually perpendicular signals from a section of TRT rotor,applies LMD to decompose each signal into multiple PF(Product Function) components,fuses the PF envelope functions of two signals to get a full vector envelope spectrum for each PF component,and diagnoses the TRT vibration faults accordingly.The results of simulation and case analysis show that,the fault information extracted by the proposed method is more comprehensive and accurate than that by the single-signal information analysis.