MULTI-OBJECTIVE OPTIMISATION AND ITS APPLICATION TO CONCEPT DEVELOPMENT IN FLOATING OFFSHORE SYSTEMS

摘要

This paper addresses the need for a rapid, multi-disciplined and rational approach to floating system concept developmentand selection during the very earliest stages of project definition. It describes the implementation of a modified multiobjectiveGenetic Algorithm for this purpose. A formulation of the NSGA-II algorithm is combined with additional TargetFunctions to reduce otherwise large multi-disciplined problems to more tractable solution using tools commonly availablein the design office. It also provides a rational basis for the comparison of different design solutions each of which arePareto Optimal with respect to the technical and economic performance of each underlying concept. A specific exampleof marginal field development using a novel FPSO concept is presented. Starting with just the oil field location and reservesestimate, the algorithm provides the means to define preliminary hull form and production facility capacities, matchperformance to payload, and give preliminary indicators of likely investment performance. The method may also beapplied more generally in preliminary ship design, particularly where it is possible to model economic performancealongside efficiency, safety and key technical factors in hydrodynamics and structures.