摘要
为提高分形维数的计算精度和效率,减少分形无标度区确定中的主观因素,基于分形维数计算中的关联积分算法,综合利用具体时间序列和拟合点数等信息,建立故障诊断中分形无标度区的一种自动识别方法。运用该方法对所测得的两台航空发动机在不同工况、不同转速下的振动数据进行分析计算,给出分形无标度区的上、下限和关联维数,并与一般方法的计算结果进行对比。计算结果表明,无论实测时间序列无标度区跨度大小如何变化,应用本方法均能很好地识别,区间内曲线的线性度较好,且该自动识别方法简单、高效、准确、适用性强。需指出的是,该方法确定的无标度区附近的值有可能属于无标度区,故它对无标度区的选择趋于保守。
In order to improve the precision and efficiency for fractal dimension calculation and reduce manmade influence, a technique is proposed for fractal scaleless band automatic identification. The method is based on the study of the correlation integral algorithm and the information of time series. The number of fit points was also considered. Firstly the fractal scaleless＇s probable range is estimated.Then it is exactly confirmed. Finally regression significance test of Inr-InCq（r） curve is made.The method was applied to analyze two aero engine＇s vibration time series, which were tested in different operation conditions and different rotate speeds. Meanwhile the result was compared with the general calculation results. It shows that the method is effectual no matter how much scaleless band is and the linearity of curves is better in the scaleless band. The method is more objective, precise, simple and applicable, however sometimes the fractal scaleless band may be shrunk by it.
出处
《机械工程学报》
EI
CAS
CSCD
北大核心
2006年第12期106-109,共4页
Journal of Mechanical Engineering
基金
学院博士研究基金。
关键词
分形无标度区
关联积分
相空间重构
故障诊断
Fractal scaleless band Correlation integral Phase space reconstruction Fault diagnose
