:Accounting for Uncertainty Caused by Lack of Data and Conflicting Knowledge. Robust Reliability via Uncertainty Quantification
Speaker: Roberto Rocchetta, Research Scholar, NIA
Date: Wednesday, March 13, 2019
Location: NIA, Room 101
Abstract: Classical probabilistic methods are commonly employed to handle reliability analysis and uncertainty quantification tasks. However, classical probabilistic methods often rely upon a considerable body of empirical evidence, large volumes of data, and good quality information. When the data is limited, information vague or inconsistent (i.e. in the presence of both aleatory and epistemic uncertainty), classic probabilistic approaches may lead to misleading conclusions and a false sense of confidence which may not fully represent the quality of the available information. This talk introduces a generalized probabilistic framework for reliability assessment and uncertainty quantification under a lack of data. The developed computational tool allows the effect of epistemic uncertainty to be quantified and is used to assess the reliability of an electronic circuit and power grid test cases. Strength and weakness of the proposed approach are illustrated by comparison to traditional probabilistic approaches.
Speaker Bio: Roberto Rocchetta, Research Scholar involved in a collaborative research effort with NIA and the Dynamic Systems & Control Branch at NASA Langley. Roberto holds a Master degree in Energy Engineering from the University of Bologna in Italy, a Master of Research in Decision-making under risk and uncertainty and Ph.D. in Reliability Engineering from the University of Liverpool, United Kingdom. Prior to joining NIA and during his Ph.D., Roberto was a visiting researcher at ETH Zurich in Switzerland and at Polytechnic of Milano in Italy. His research primarily focused on the development of methods for stochastic optimization, reliability-based design, and on the analysis of complex systems and critical infrastructures in uncertain environments.