Publication detail

Loading condition monitoring on trusses applying a machine learning approach with training data of a finite element model: A study case

Tinoco, H. A., & Cardona, C. I.

Original Title

Loading condition monitoring on trusses applying a machine learning approach with training data of a finite element model: A study case

Type

conference paper

Language

English

Original Abstract

Structural health monitoring (SHM) techniques deal with the changes in the dynamic or static characteristics of the structures that affect its performance during the service [1]. Mainly, these techniques are based on vibrations, and their implementation includes complex integrated systems not addressed from the structural design. Despite the numerous applications of SHM, loading condition monitoring (application place, direction, and magnitude) is not a very implemented strategy in this engineering field. This paper presents a methodology to monitor the application of external forces on structures using a learning machine process and finite element analysis (FEA). In Figure 1a is described the proposed monitoring methodology, which is applied to a truss structure to validate this study. The truss contains nine structural elements, and each one presents a piezoelectric transducer to measure the forces in their links, as illustrated in Figure 1b. The real truss was modeled by means of a FEA (implemented in Matlab with truss elements) considering their mechanical properties and loading conditions, as observed in Figure 1c. To simulate different loading conditions, a force F_s is applied in node 3 varying two parameters, angle β, and magnitude. This is carried out to establish a database using the internal forces (in each bar) obtained by FEA.

Keywords

Machine learning, loading monitoring, Finite element method

Authors

Tinoco, H. A., & Cardona, C. I.

Released

4. 11. 2019

Publisher

University of West Bohemia,

Location

Plzen, Czech Republic

ISBN

978-80-261-0889-4

Book

Proceedings of 35th conference Computational Mechanics 2019

Edition

Vítezslav Adámek

Edition number

35

Pages from

205

Pages to

206

Pages count

3

BibTex

@inproceedings{BUT167362,
  author="Hector Andres {Tinoco Navarro}",
  title="Loading condition monitoring on trusses applying a machine learning approach with training data of a finite element model: A study case",
  booktitle="Proceedings of 35th conference Computational Mechanics 2019",
  year="2019",
  series="Vítezslav Adámek",
  number="35",
  pages="205--206",
  publisher="University of West Bohemia,",
  address="Plzen, Czech Republic",
  isbn="978-80-261-0889-4"
}