Novel risk assessment and decision support techniques for safety management systems

Abstract

A brief review of the current status of marine and offshore safety regulations in the UK is outlined. Several issues pertaining to the problems encountered in applying the existing reliability and safety analysis methods in quantitative safety appraisal studies, especially in the early concept design stage of marine and offshore engineering products, are discussed. In this paper, four novel risk assessment and decision support frameworks are presented. These include a design trade-off approach using the Taguchi method, a safety-based decision support system based on artificial neural network techniques, a fuzzy-logic-based synthesis incorporating the Dempster-Shafer approach for multiple attribute decision-making, and an integration of approximate reasoning approach and evidential reasoning method for design evaluation. Four illustrative examples are used to demonstrate the novel tools, together with the discussion on the conditions under which each approach may be applied effectively. Finally, recommendations on further development in subjective safety modelling, decisionmaking techniques and their integration into a safety management system are suggested.