Date of Award

Winter 2001

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Engineering Management & Systems Engineering

Program/Concentration

Engineering Management and Systems Engineering

Committee Director

Resit Unal

Committee Member

Paul Kauffmann

Committee Member

David Dryer

Committee Member

Han Bao

Committee Member

Arlene Moore

Abstract

Design of complex, one-of-a-kind systems, such as space transportation systems, is characterized by high uncertainty and, consequently, high risk. It is necessary to account for these uncertainties in the design process to produce systems that are more reliable. Systems designed by including uncertainties and managing them, as well, are more robust and less prone to poor operations as a result of parameter variability.

The quantification, analysis and mitigation of uncertainties are challenging tasks as many systems lack historical data. In such an environment, risk or uncertainty quantification becomes subjective because input data is based on professional judgment. Additionally, there are uncertainties associated with the analysis tools and models. Both the input data and the model uncertainties must be considered for a multi disciplinary systems level risk analysis.

This research synthesizes an integrated approach for developing a method for risk analysis. Expert judgment methodology is employed to quantify external risk. This methodology is then combined with a Latin Hypercube Sampling - Monte Carlo simulation to propagate uncertainties across a multidisciplinary environment for the overall system. Finally, a robust design strategy is employed to mitigate risk during the optimization process. This type of approach to risk analysis is conducive to the examination of quantitative risk factors.

The core of this research methodology is the theoretical framework for uncertainty propagation. The research is divided into three stages or modules. The first two modules include the identification/quantification and propagation of uncertainties. The third module involves the management of uncertainties or response optimization. This final module also incorporates the integration of risk into program decision-making.

The risk analysis methodology, is applied to a launch vehicle conceptual design study at NASA Langley Research Center. The launch vehicle multidisciplinary environment consists of the interface between configuration and sizing analysis outputs and aerodynamic parameter computations. Uncertainties are analyzed for both simulation tools and their associated input parameters. Uncertainties are then propagated across the design environment and a robust design optimization is performed over the range of a critical input parameter.

The results of this research indicate that including uncertainties into design processes may require modification of design constraints previously considered acceptable in deterministic analyses.

DOI

10.25777/yyg4-6c21

ISBN

9780493565095

Download

Share

COinS