Development of a digital twin to simulate the post-buckling behavior of composite stiffened panels
Loading...
Date
2022-10-24
Authors
Βαφείδης, Κωνσταντίνος
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The present work is dedicated to the development of a Digital Twin (DT) of a multi-stringer composite panel subjected to Quasi-Static (QS) loading. Four tests are executed with a maximum load of 250,275 and 300 kN. At 250 kN two tests are carried out, one with the specimen being in pristine state (Before Impact – BI), and one where damage After Impact (AI) is present. The rest two tests involve only the specimen in the damaged state. During all four tests strain measurements are recorded via Fiber-Optic Sensors (FOS) placed on the specimen, while for the BI test the post-buckling behavior is captured via Digital Image Correlation (DIC). The purpose of the DT is a) to simulate the post-buckling behavior of the structure, and b) to be used as the healthy state baseline of a Structural Health Monitoring (SHM) system. First, a Finite Element Model (FEM) of the panel is developed and Non-Linear Buckling (NLB) analyses are run via both implicit and explicit solvers. The implicit analysis simulates accurately the buckling load but appears too stiff in the post-buckling region. On the other hand, the post-buckling behavior of the explicit analysis is validated by the DIC findings, though at several load regions the FEM overestimates the acting loads. Then, Surrogate Models (SMs) are trained upon the FEM results (strains) to provide pristine strains for the SHM system. The trained SMs accurately replicate the FEM presenting low RMSEs. Finally, the SHM methodology is applied where, at start, a load identification process is executed utilizing the measured strains. The surrogate and experimental strain distributions are compared at multiple loads to examine if the induced damage can be detected and localized. The results show a successful detection and a semi-successful localization, as only the location of one impact is found. In order to have an automated system, a Health Indicator (ΗΙ) is defined, and a healthy state threshold is specified. Via this method, the damage detection is achieved for both impacts, however only one of them is localized.
Description
Keywords
Composite stiffened panels, Digital twin, Surrogate modeling, Structural health monitoring, Fiber-optic sensors