美国Case Western Reserve Univisity的轴承数据EMD分析出现问题

oggilim

最近想用EMD分析这网站上轴承实验数据，是IR007_1内圈数据，用MATLAB仿真，却出现这错误
1IMFs have been obtained.
2IMFs have been obtained.
3IMFs have been obtained.
4IMFs have been obtained.
5IMFs have been obtained.
6IMFs have been obtained.
7IMFs have been obtained.
8IMFs have been obtained.
9IMFs have been obtained.
10IMFs have been obtained.
11IMFs have been obtained.
12IMFs have been obtained.
13IMFs have been obtained.
14IMFs have been obtained.
15IMFs have been obtained.
16IMFs have been obtained.
17IMFs have been obtained.
??? Error using ==> interp1
Y must have length(X) rows.

Error in ==> Emd at 40
        Min_Env=interp1([LMinEtd t(IndMin) RMinEtd],[hlMaxEtd h(IndMin) hrMinEtd],t,'spline');

Error in ==> Untitled at 2
[Imf]=emd(sig);
我个人认为既然能分析出前17个IMF，语句肯定是没问题，那么是否意味着在计算过程中残余部分不满足三次样条插值呢？问题有点棘手，向各位大虾请教了

gczhang

前面5-6个IMF足够你用来分析轴承故障了，后面的EMD错误可以不理，估计是因为端点效应出的问题，你换组数据可能就没有了。

lxllwb

回复 1 # oggilim 的帖子

matlab程序问题不大，关键是不知道数据是如何采集的，仿真后分析没用，必须与现场的实际结合啊，

oggilim

谢谢各位朋友热心的帮助，虽然问题没解决，我想用IMF计算能量熵，所以希望能得到全面的IMF。