RT Journal Article
T1 Model updating of uncertain parameters of carbon/epoxy composite plates from experimental modal data
A1 Cuadrado Sanguino, Manuel
A1 Artero Guerrero, José Alfonso
A1 Pernas Sánchez, Jesús
A1 Varas Doval, David
AB This work presents a methodology to obtain physically-sound models of composite structure laminates using a combination of modal analysis, numerical modelling and parameter updating, avoiding the common uncertainties on the constructions of similar numerical models. Moreover this model establishes the baseline (pristine situation) of the dynamic behaviour of the set of composite plates. Therefore it could be applied for condition assessment or quality manufacturing control of existing structures through a non-destructive Structural Health Monitoring (SHM), and hence it could help to detect degradation or defects of the composite components. The driven data of the methodology were the modal frequencies and shapes of composite plates. To obtain these values an extensive experimental campaign of modal analysis has been performed on a set of carbon/epoxy laminates. A multiple input single output technique has been applied, using a roving hammer exciting the plates at evenly distributed Degrees of Freedom (DoF), and a mono-axial accelerometer attached to a single DoF reference point. The use of a high dense grid of points has allowed to identify a number of natural frequencies greater than usual in similar works, as well as improving the smoothness of the mode shape. Modal characteristics numerically obtained from a Finite Element Method (FEM) model based on manufacturer reference data were compared with experimental results. This baseline model was updated through a gradient based optimization algorithm. Before the process of model updating, a sensitivity analysis has been performed to identify the driven uncertain parameters using a Montecarlo approach. This technique reduces the number of parameters to be optimized to a small set increasing the efficiency of the methodology. As a result of the whole process, a physically more accurate model is obtained on which discrepancies with the corresponding experimentally measured modal parameters are drastically reduced. Analy
PB Elsevier
SN 0022-460X
YR 2019
FD 2019-09-01
LK https://hdl.handle.net/10016/31658
UL https://hdl.handle.net/10016/31658
LA eng
NO This research was done with the financial support of the Spanish Ministry of Economy and Competitiveness under Project reference DPI2013-41094-R, and the Vicerrectorado de Política Científica UC3M (Projects 2014/00006/002 and 2013/00413/003).
DS e-Archivo
RD 30 jun. 2024