Date of Award
5-2025
Document Type
Thesis
Degree Name
Master of Science (MS)
Department
Mechanical and Civil Engineering
First Advisor
Anand Balu Nellippallil
Second Advisor
Ilya Mingareev
Third Advisor
Karen Holness
Fourth Advisor
Troy V Nguyen
Abstract
The design of complex engineered systems often relies on simulation-based models, which inherently introduce uncertainties. Managing these uncertainties is critical to ensure robust system performance. From a decision-based design (DBD) perspective, robust design metrics such as the Design Capability Index (DCI) and Error Margin Index (EMI) are employed to formulate and solve problems under uncertainty. However, existing formulations of these metrics rely on first-order Taylor series approximations for variance estimation, which are often inadequate in the presence of highly nonlinear or multimodal performance functions.
This thesis proposes alternate formulations of DCI and EMI, which use the multiple-point method of variance estimation to ensure robustness in the presence of highly non-linear performance functions. This approach uses the weighted average of three points for each design variable to obtain the correct variance at the local optimum. An updated robust Concept Exploration Framework (rCEF) is proposed by introducing the alternate DCI and EMI metrics in the compromise Decision Support Problem(cDSP) of the rCEF and using a machine learning-based visualization tool, the interpretable Self-Organizing Maps (iSOM) technique, to visualize and explore the high-dimensional robust solution space in two dimensions.
The proposed framework is validated using two examples: a simple non-linear polynomial function and a complex hot rod rolling problem. Results demonstrate the improved robustness and applicability of the updated metrics and framework in managing uncertainty in complex engineering systems.
Recommended Citation
Mukundan, Pooja, "Alternative Robust Design Formulation and Exploration Strategies for Managing Uncertainties in the Realization of Complex Engineering Systems" (2025). Theses and Dissertations. 1572.
https://repository.fit.edu/etd/1572
