Uncertainty in Industrial Practice - A Guide to Quantitative Uncertainty Management

Abstract

Uncertainty is a fascinating subject, spanning a wide range of scientific and practicalknowledge domains. Economics, physics, decision theory and risk assessment are its traditional locations, while more recently it has also become common in modelling, numerical analysis, advanced statistics and computer science. However, the study of uncertainty has come into contact even with such subjects as epistemology, management science, psychology, public debate and theories of democracy.\nThe focus of this book is narrower. As the product of a working group of practitioners representing a large panel of industrial branches, it is dedicated to the better understanding of the industrial constraints and best practices associated with the study of uncertainty. It is\nconcerned with the quantification of uncertainties in the presence of data, model(s) and knowledge about the problem, and aims to make a technical contribution to decision-making processes. It should illustrate the optimal trade-offs between literature-referenced methodologies and the simplified approaches often inevitable in practice, owing to data, time or budget limitations, or simply to the current formulation of regulations and the cultural habits of technical decision-makers.\nThe book is organized as follows.\nPart I introduces the common methodological framework which forms the backbone of the work: this is the essential generic structure unifying the industrial case studies (which are sketched briefly in Part I Chapter 2), notwithstanding the peculiarities of each due to sector,\nphysical or decision-criteria specificities. Part II comprises a number of industrial case studies recently carried out in nuclear, oil, aerospace and mechanical industries or civil structures around Europe. The chapters follow a standard format in order to facilitate comparison, in line\nwith the common methodological framework: they illustrate a number of practical difficulties encountered or solutions found, according to the level of complexity and regulatory or financial constraints.\nPart III is the scientific core of the book: it includes a review of methods which may be used in the main steps and towards the main goals of any uncertainty study. Citing a large and diverse academic literature, critical recommendations are developed from the applied point of\nview. The annexes gather supporting material for the practitioner. First, salient references and a few generic standards are provided in correspondence with the key steps and methods of the common methodological framework (Annex 1). Certain important items of software are then reviewed with a particular focus on their operational characteristics and features (Annex 2): it is obviously a fast-moving reality for which completeness or permanence cannot be fully guaranteed. Once again, the purpose is to link the review to the salient features and key questions identified in the common methodological framework, rather than to reproduce the detailed specifications of each type of software.\nThe chapters of the book may be read independently to a certain extent: readers familiar with the subject may go directly to any case study (in Part II) to find illustrations of industrial practices in a given domain; alternatively they may refer to the more specialized chapters of Part III for the most appropriate methods for a given purpose. It is suggested, however, that the reader begin at Part I for an introduction to the distinctive vocabulary and approach of the book.