ARTS09-jelena-format.pdf (151.78 kB)
0/0

An integrated design optimisation approach for systems with dependencies

Download (151.78 kB)
conference contribution
posted on 23.01.2012 by Jelena Riauke, Lisa Jackson
The design of a safety system is critical if functionality is to be maximised and consequences reduced. There is often a trade off between the performance obtainable and the resources available. To address these balancing issues, which are usually impractical by hand for a designer, multi-objective optimisation techniques can be used. When considering safety systems there is often the situation of dependencies between components, for example with regard to maintenance. To evaluate the system behaviour in these situations an appropriate analysis method is required. The aim of this paper is to present an optimisation approach which integrates traditional methods of system failure evaluation. The combined method uses the fault tree analysis technique to represent the causes of failure on demand of the system, the binary decision diagram and Markov methods for system quantification (for independent and dependent sections of the fault tree respectively), and the Improved Strength Pareto Evolutionary Approach (SPEA2) to find the most optimal design solution. The end product is a mechanism to yield the best design option for safety systems incorporating dependencies. The paper presents the principles of the method and a case study to illustrate how the method is applied. The results produced, along with conclusions are provided.

History

School

  • Aeronautical, Automotive, Chemical and Materials Engineering

Department

  • Aeronautical and Automotive Engineering

Citation

RIAUKE, J. and BARTLETT, L.M., 2009. An integrated design optimisation approach for systems with dependencies. IN: Proceedings of the Advances in Risk and Reliability Technology Symposium (AR2TS), Loughborough, 21-23 April, pp. 455 - 466

Publisher

Loughborough University Department of Aeronautical & Automotive Engineering & Transport Studies

Version

AM (Accepted Manuscript)

Publication date

2009

Notes

This is a conference paper.

ISBN

9780904947632

Language

en

Exports

Logo branding

Keyword(s)

Exports