Jang, Jinwoo

This dissertation focuses on the development of data-driven and physics-based modeling for two distinct significant structural engineering applications: time-varying response variables estimation and unwanted lateral vibration control. In the first part, I propose a machine learning (ML)-based surrogate modeling to directly predict dynamic responses over an entire mechanical system during operations. Any mechanical system design, as well as structural health monitoring systems, require transient vibration analysis. However, traditional methods and modeling calculations are time- and resource-consuming. The use of ML approaches is particularly promising in scientific and engineering challenges containing processes that are not completely understood, or where it is computationally infeasible to run numerical or analytical models at desired resolutions in space and time. In this research, an ML-based surrogate for the FEA approach is developed to forecast the time-varying response, i.e., displacement of a two-dimensional truss structure. Various ML regression algorithms including decision trees and deep neural networks are developed to predict movement over a truss structure, and their efficiencies are investigated. ML algorithms have been combined with FEA in preliminary attempts to address issues in static mechanical systems.

The purpose of online parameter learning and modeling is to validate and restore the properties of a structure based on legitimate observations. Online parameter learning assists in determining the unidentified characteristics of a structure by offering enhanced predictions of the vibration responses of the system. From the utilization of modeling, the predicted outcomes can be produced with a minimal amount of given measurements, which can be compared to the true response of the system. In this simulation study, the Kalman filter technique is used to produce sets of predictions and to infer the stiffness parameter based on noisy measurement. From this, the performance of online parameter identification can be tested with respect to different noise levels. This research is based on simulation work showcasing how effective the Kalman filtering techniques are in dealing with analytical uncertainties of data.

This experiment investigated the effect of infill pattern shape on structural stiffness for 3D printed components made out of carbon fiber reinforced nylon. In order to determine the natural frequency of each specimen, nondestructive vibrational testing was conducted and processed using data acquisition software. After obtaining the acceleration information of each component, in response to ambient vibrational conditions and excitation, frequency response functions were generated. These functions provided the natural frequency of each component, making it possible to calculate their respective stiffness values. The four infill patterns investigated in this experiment were: Zig Zag, Tri-Hex, Triangle, and Concentric.
Results of the experiment showed that changing the infill pattern of a 3D printed component, while maintaining a constant geometry and density, could increase mechanical stiffness properties by a factor of two. Comprehensively, the experiment showed that infill pattern geometry directly attributes to the mechanical stiffness of 3D printed components.