Author ORCID Identifier

https://orcid.org/0009-0009-8561-1882

Semester

Summer

Date of Graduation

2024

Document Type

Dissertation

Degree Type

PhD

College

Statler College of Engineering and Mineral Resources

Department

Mechanical and Aerospace Engineering

Committee Chair

Nithi T. Sivaneri, Ph.D.

Committee Co-Chair

Osama M. Mukdadi, Ph.D.

Committee Member

Christopher D. Griffin, Ph.D.

Committee Member

Marcello R. Napolitano, Ph.D.

Committee Member

Adam M. Halasz, Ph.D.

Abstract

Knowledge of excitation loads that structures experience during their service life is pivotal in different engineering fields, not only from a structural design optimization point of view but also as prevention of possible damages to the structures themselves. However, in case of dynamic events such as tornadoes, structures subjected to impulsive load due to their vicinity of explosions, the excitation load cannot be directly determined through direct measurements. In these scenarios, the inverse problem is used, and it is called load identification. This type of problem tries to determine the excitation load knowing the system response through a series of sensors placed on the structure. Most of the time, the number of sensors and their locations are randomly selected causing errors in the load estimation. The proposed mathematical method provides the appropriate number of sensors and their optimal locations. It combines the Craig-Brampton condensation method with the normal-mode method and D-optimal design technique. This method is employed and verified on simple structural members such as beams and plates and then it is applied on more complex structures such as a wind-turbine tower, wind-turbine blade, and an aircraft wing, which lead to more complexity in the procedures. Finite element models of these simple and complex structures are made in the general-purpose software ABAQUS. Then, a free-vibration analysis is carried out on each structure and the natural frequencies and mode shapes are extracted. Different load shapes are applied on the structures at different frequencies and noise conditions. The dynamic load is reconstructed by measuring the transient response of the structure at the optimum sensor locations. The results reveal that the proposed method can reconstruct a dynamic load with a high level of accuracy. Furthermore, the implementation of different parameters, such as noise effects, does not cause amplified errors in the final load estimation making the proposed method more robust.

Share

COinS