Towards Failure-free Design: Reducing Dimensionality in Function-failure Similarity Analysis for Large Databases
Abstract
When designing products, it is crucial to assure failure and risk-free operation in the intended operating environment. Failures are typically studied and eliminated as much as possible during the early stages of design. The few failures that go undetected result in unacceptable damage and losses in high-risk applications where public safety is of concern. Published NASA and NTSB accident reports point to a variety of components identified as sources of failures in the reported cases. In previous work, data from these reports were processed and placed in matrix form for all the system components and failure modes encountered, and then manipulated using matrix methods to determine similarities between the different components and failure modes. In this paper, these matrices are represented in the form of a linear combination of failures modes, mathematically formed using Principal Components Analysis (PCA) decomposition. The PCA decomposition results in a low-dimensionality representation of all failure modes and components of interest, represented in a transformed coordinate system. Such a representation opens the way for efficient pattern analysis and prediction of failure modes with highest potential risks on the final product, rather than making decisions based on the large space of component and failure mode data. The mathematics of the proposed method are explained first using a simple example problem. The method is then applied to component failure data gathered from helicopter accident reports to demonstrate its potential. Copyright © 2002 by ASME.
Recommended Citation
I. Y. Tumer et al., "Towards Failure-free Design: Reducing Dimensionality in Function-failure Similarity Analysis for Large Databases," ASME International Mechanical Engineering Congress and Exposition, Proceedings, pp. 631 - 638, American Society of Mechanical Engineers, Jan 2002.
The definitive version is available at https://doi.org/10.1115/IMECE2002-33473
Department(s)
Mechanical and Aerospace Engineering
Keywords and Phrases
Failure prevention; Failure-free product development; Function-failure similarity analysis; Principal components analysis; Risk-Based design
International Standard Book Number (ISBN)
978-079183628-6
Document Type
Article - Conference proceedings
Document Version
Citation
File Type
text
Language(s)
English
Rights
© 2024 American Society of Mechanical Engineers, All rights reserved.
Publication Date
01 Jan 2002
