Computational & Technology Resources
an online resource for computational,
engineering & technology publications |
|
Civil-Comp Proceedings
ISSN 1759-3433 CCP: 78
PROCEEDINGS OF THE SEVENTH INTERNATIONAL CONFERENCE ON THE APPLICATION OF ARTIFICIAL INTELLIGENCE TO CIVIL AND STRUCTURAL ENGINEERING Edited by: B.H.V. Topping
Paper 35
Application of Genetic Algorithms for the Automated Design of Offshore Riser Systems N. Cunliffe and T.J. McCarthy
Manchester Centre for Civil and Construction Engineering, UMIST, Manchester, United Kingdom N. Cunliffe, T.J. McCarthy, "Application of Genetic Algorithms for the Automated Design of Offshore Riser Systems", in B.H.V. Topping, (Editor), "Proceedings of the Seventh International Conference on the Application of Artificial Intelligence to Civil and Structural Engineering", Civil-Comp Press, Stirlingshire, UK, Paper 35, 2003. doi:10.4203/ccp.78.35
Keywords: offshore engineering, marine design, catenary riser, optimisation, evolutionary search, genetic algorithms.
Summary
The design of steel catenary riser systems for deepwater floating vessels is a
complicated design problem. The principal design variables include vessel
performance and position, riser length, diameter, material, wall section and
buoyancy. The length of the riser is of the order of kilometres and there are a
number of possible configurations (Figure 35.1). The riser system must fulfil the
requirements of several loadcases in which the floating vessel (and therefore risers)
will change position. The principal constraints are maximum stress and fatigue
limitations together with serviceability requirements.
The large number of design variables and distinct loadcases together with geometric non-linear behaviour result in a design methodology based on intuition and iteration, resulting in a tedious process which may not produce the optimal design. The problem may be formulated as an optimisation problem in which minimum riser system cost is the objective function. Larsen [1] uses a sequential quadratic programming scheme to optimise a simple steel catenary riser under static analysis. In more complex applications were several riser configurations may exist and were the design variables may be real or integer valued, several local minima may exist. In complex problems the search space may be discontinuous. Such search space characteristics render gradient based search impractical. The approach taken in this paper is to determine the values of the design variables by using Genetic Algorithms (GAs) which are capable of distinguishing global optima within a non-continuous search space. The developed software system (Figure 35.2) uses a commercially available analysis program, OrcaFlex [2], for design evaluation and riser/optimisation model development. The GA capability is provided by a library of GA objects called GAlib [3]. Genome fitness is based on riser system cost using a bill of materials type approach. The developed system is both powerful and flexible; it can be used for the development of a whole riser system or as an assistant for the optimisation of specific design variables, performing optimisations using both static and dynamic analysis. A comparison is presented between the results of the current system and Larsen's [1] optimised solution for a simple catenary riser under static analysis. Reasonable agreement is achieved. Where as Larsen's approach is highly dependent on the starting position of the solution search, the system presented here finds the global optimum from any starting position.
References
purchase the full-text of this paper (price £20)
go to the previous paper |
|