arrow
Return

A machine learning approach for automatic operational modal analysis

delete2022-05-01
delete45
PRE
AI
L
Luca Zanotti Fragonara *
M
Marco Civera
DOI:10.1016/j.ymssp.2022.108813delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En 中文
One of the major applications of Structural Dynamics in Civil, Mechanical, or Aerospace Engineering regards the dynamic characterisation of man-made structures and components. Yet, traditional Experimental Modal Analysis (EMA) needs dedicated setups which may not be always available where and when needed. For these and other reasons, output-only Operational Modal Analysis (OMA) is regarded as a more practical and convenient alternative. Many OMA algorithms have been reported in the scientific literature during the last twenty and more years. In this study, an Automatic OMA method is presented. The proposed algorithm is completely independent of the user experience, fully objective, and based on statistical principles and a Machine Learning (ML) clustering approach. The AOMA code is firstly applied to a numerical case study, to test all the parameters which control the process. An Airbus H135 helicopter blade is then analysed to verify the performance of the algorithm experimentally.
Keywords:
Operational modal analysis
AOMA
Stabilization diagram
Machine learning
Ambient vibration
Clustering analysis
Helicopter blade
Structural Health Monitoring

Journal

Mechanical Systems and Signal Processing cover
Mechanical Systems and Signal Processing
IF:
8.9
Papers:
1.3W
Citations:
6.6W

Organization

P
Polytechnic University of Turin
Scholars:
1.3W
Papers: 1.3W
Citations: 1.3W
C
cranfield university
Scholars:
6.3K
Papers: 6.6K
Citations: 1