The newly proposed modal distribution method statistically quantifies overall differences between measured time-histories. In this method, power spectra of measured structural response are interpreted as collections of independent modal responses. Each modal response is isolated, re-scaled, and interpreted as a statistical distribution. Data-sets (windows in a measured time-history) are then compared using standard statistical methods, resulting in a quantitative significance level of the differences between power spectra. An example is presented to validate the new method and to quantify how long a time-history is required for the new method to meet confidence level requirements. The modal distribution method is found to be very effective at detecting subtle changes of mean modal frequencies, which may be used to infer changes in structural condition. The method is general and may find a broad variety of applications, but seems particularly well suited for structural health monitoring because it can be used to infer changes in structural condition from measured response data with only limited knowledge of the excitation.
[1]
J. Hoffman.
Numerical Methods for Engineers and Scientists
,
2018
.
[2]
William Cyrus Navidi,et al.
Statistics for Engineers and Scientists
,
2004
.
[3]
Andrew W. Smyth,et al.
Surveillance of Mechanical Systems on the Basis of Vibration Signature Analysis
,
2000
.
[4]
J. Kim Vandiver,et al.
Detection of Structural Failure on Fixed Platforms By Measurement of Dynamic Response
,
1977
.
[5]
Cyril M. Harris,et al.
Shock and vibration handbook
,
1976
.
[6]
Hoon Sohn,et al.
A review of structural health monitoring literature 1996-2001
,
2002
.
[7]
Benjamin S. Duran,et al.
Statistical Methods for Engineers and Scientists
,
1985
.
[8]
Lambros S. Katafygiotis,et al.
A probabilistic approach to structural health monitoring using dynamic data
,
1997
.
[9]
Siegmund Brandt,et al.
Statistical and Computational Methods in Data Analysis
,
1971
.