1 Introduction

This document is part of OpenTURNS' documentation. Its aim is to introduce the global methodology for the quantification of uncertainties by a model-based approach, and the methods proposed by OpenTURNS to carry out the different steps of such a study. Indeed, even if each industrial study exhibits some particularities, a common framework composed of four steps is proposed.

The first part of this document briefly presents each step, and illustrates the key points through a realistic – even though simplified – example.

In the second part of the document, the focus is placed on the methods that an analyst may use in OpenTURNS to carry out steps B, C and C'. For each method, a synthetic form is given to highlight:

For the practical side i.e. the use of these functions in OpenTURNS Textual User Interface, the User is refered to the [TUI Use Cases Guide] .

1.1 Presentation of the flood example

Suppose that the industrial study concerns a dyke built along a river to protect an industrial facility from floods. For the industry that runs this facility, a risk exists: even if the dyke would have been high enough to contain the major floods of the last century, one does not know if the protection will be sufficient to face the next flood. For instance, meteorological events vary from year to year and are thus considered random, at least given our current knowledge in this scientific field. Therefore, an uncertainty study becomes valuable to ensure risk control (the dyke should be high enough to limit the risk of inundation) and economical optimization (the construction and maintenance cost increase with the dyke height, which should consequently not be either under or over-dimensioned).

Table of contents
Global methodology of an uncertainty study