An investigation of finite element model updating of the Pioneering Spirit

An investigation of finite element model updating of the Pioneering Spirit

Schout, Arthur (TU Delft Mechanical, Maritime and Materials Engineering)

Kaminski, M.L. (mentor)
Bos, R.W. (mentor)
van Woerkom, P.T.L.M. (mentor)
van der Horst, Tim (mentor)
Rodenburg, Freek (mentor)

Delft University of Technology

2018-08-27

Correctly predicting the structural behavior of the Pioneering Spirit is vital for ensuring the structural integrity of the ship and the cargo. A detailed finite element model is used to predict the structural behavior of the Pioneering Spirit. The finite element method is based on fundamental principles in solid mechanics. However, when using finite element models considerable differences between the predicted and observed behavior of a structure can occur, even when best industry practices are used to create such models. Because of these
differences there is a need to validate the detailed finite element model of the Pioneering
Spirit.
Finite element model updating is a method that can validate finite element models. In this method the discrepancy between the measured behavior and the observed behavior is minimized by modifying model assumptions and parameters. Currently a number of sensors is installed on the Pioneering Spirit, which can be used to find the measured behavior. Whether or not the measured behavior is detailed enough to be used in the validation of the finite element model is the subject of this research. To investigate this a simplified finite element model of the Pioneering Spirit was created using beam elements, this model provides the predicted behavior. Then sensitivity-based finite element model updating was implemented and applied to the beam model.
Simulated
measurements were used to show that the beam model can be updated using the current sensor setup. When actual measurements were used to update the beam model it was found that the beam model does not correlate with the measured behavior, making it impossible to update the beam model in a meaningful way.
The detailed model does correlate with measured behavior. By assuming that the method will work similarly for the detailed model as it did for the beam model, it can be concluded that the detailed model can be validated using the current sensor setup for a static case. For a dynamic case this is not possible.

FEMU
Finite element model updating
FEM

http://resolver.tudelft.nl/uuid:277f7b38-3eb2-455d-9d4f-c5afdf349b06

2023-08-27

Student theses

master thesis

© 2018 Arthur Schout