Delivering the Next Generation of Probabilistic Safety Analyses

Simply put, XFTA is a powerful and efficient calculation engine for fault trees and related models.

- XFTA has a high expressive power, as it provides a full-fledged object-oriented language to design models: S2ML+SBE.
- XFTA implements state-of-the-art assessment algorithms and a wide range of calculations of risk indicators, which cover most, if not all of, the needs of probabilistic safety analyses.
- XFTA is fully documented as it comes with the XFTA Book, which provides an in-depth presentation of mathematical and algorithmic frameworks of probabilistic safety analyses, a complete description of XFTA commands and numerous application examples.
- XFTA is free of use, including for commercial purposes.

XFTA is a calculation engine: it reads a model into one or several text files, performs a number of calculations on this model, and prints out results of these calculations into one or several text files. The calculations to be performed are described by means of scripts, i.e. text files containing sequences of commands. There are commands to load models and scripts, extract minimal cutsets, calculate values of qualitative and probabilistic indicators (including probability of the top-event, importance measures, failure intensity and safety integrity levels), print out various data, in a word to perform all operations involved in probabilistic safety analyses.

The syntax for equations that describe basic and intermediate events is quite intuitive. If you know what a fault tree is, you should have no problem to write down your models. The syntax for commands is slightly more specific. However, the same scripts are used over and over. In a word, even a non experienced analyst should be able to make a simple use of XFTA quickly.

The above presentation is however quite reductive because XFTA can be used for much more advanced purposes.

XFTA provides a full-fledged object-oriented language to design models: S2ML+SBE. S2ML+SBE is the combination of S2ML, which stands for system structure modeling language and SBE, which stands for stochastic Boolean equations. Systems of stochastic Boolean equations are the underlying mathematical framework of fault trees and reliability block diagrams: any fault tree or reliability block diagram is eventually a system of stochastic Boolean equations. S2ML is a coherent and versatile set of object-oriented constructs that help to design and to structure models. Indeed, you are not forced to use S2ML constructs if you do not want to, i.e. you can just write down systems of stochastic Boolean equations. Nevertheless, object-oriented constructs provide a significant additional expressive power and induce no additional cost, as S2ML+SBE models are automatically (and efficiently) translated into systems of stochastic Boolean equations.

S2ML+SBE is a textual modeling language. In addition to this textual format, XFTA supports the Open-PSA format, which is a XML grammar systems of stochastic Boolean equations (your can find here the specification of the original format). The Open-PSA format has been extended to support S2ML constructs. The two formats, S2ML+SBE and (extended) Open-PSA are thus equivalent.

XFTA implements state-of-the-art algorithms and data structures to assess fault trees. There are actually two main approaches to assess fault trees. The first one consists in extracting directly the minimal cutsets of the top event, then in performing calculations of probabilistic indicators from these minimal cutsets. The second one consists in building a binary decision diagram encoding the structure function of the top event, then in performing calculations of probabilistic indicators from this binary decision diagram. In this second approach, minimal cutsets can also be extracted, but starting from the binary decision diagram rather than from the model itself.

XFTA algorithms work for both coherent and non-coherent models. XFTA implements also probabilistic assessments:

- Calculation of the top event probability for different mission times.
- Calculation of importance factors for basic events (Birnbaum/Marginal Importance Factor, Critical Importance Factor, Diagnosis Importance Factor, Risk Reduction Worth, Risk Achievement Worth, Differential Importance Measure, Barlow-Proschan Factor).
- Sensitivity analyses by means of Monte-Carlo simulation.
- Calculation of safety integrity levels for low demand mode and high/continuous demand mode safety related system as required by Safety Standard IEC 61508 and daughters.

Calculation of safety integrity levels require the calculation of system availability, average availability, system failure intensity, approximation of system reliability, probability of failure on demand, probability of failure per hour…

The XFTA book provides an in-depth presentation of mathematical and algorithmic frameworks of probabilistic safety analyses, a complete description of XFTA commands and numerous application examples. You can download a free pdf version here. We are currently looking for an editor to distribute a paper version. |

XFTA is provided free of charge. Even though there is no license fee, there are conditions for using XFTA. Downloading the software means you accept with no restriction the terms of the License Agreement.

The current version of XFTA is version 2.0.1.

- XFTA installer for Windows (64 bits)
- XFTA for Windows (64 bits) coming as zip archive
- XFTA for Linux (64 bits) coming as gzipped tar archive
- XFTA for MacOS (64 bits) coming as zip tar archive

You can still downloaded version 1.3.1. The user manual for this version is included in the zip archive.

XFTA is developed by Prof. Antoine B. Rauzy.

Norwegian University of Science and Technology Department of Mechanical and Industrial Engineering S. P. Andersens veg 3, Valgrinda 7491 Trondheim, Norway +47 918 97 151 (mobile phone) |