Archived

This content is available here for research, reference, and/or recordkeeping.

Date Available

4-30-2026

Year of Publication

2026

Document Type

Doctoral Dissertation

Degree Name

Doctor of Philosophy (PhD)

College

Engineering

Department/School/Program

Mechanical Engineering

Faculty

Hasan Poonawala

Faculty

Jonathan Wenk

Abstract

This work introduces process-guided learning, a framework in which process characteristics and learning algorithms are combined in the modeling process of engineering applications with the aim of uniting traditional and data-driven modeling and control techniques. Traditional engineering methods using simplified analytic models fail to capture the increasing complexities of engineering problems with sufficient accuracy to achieve modern requirements. More advanced modeling and control techniques using high-fidelity models often produce a computational burden that is infeasible for real-time solutions. Data-driven modeling offers a simple and flexible algorithm for producing real-time capable, high-fidelity models, but naive implementations lack the physical constraints of traditional methods, and the data-hungry nature of these algorithms often make them experimentally expensive or prohibitive. Process-guided learning aims to provide a union between data-driven learning and traditional engineering approaches in order to improve the overall capabilities and efficiencies of the learned models.

This work first defines process-guided learning and introduces the core principles which constitute its substance. It then follows with four different, independent examples by which process-guided learning is implemented for physical systems.

Digital Object Identifier (DOI)

https://doi.org/10.13023/etd.2026.208

Archival?

Archival

Download

Share

COinS