Abstract:

A dimensional approach to structural optimization is presented_ Different from the usual analysis use of the concept, its synthesis use is developed to acquire the basic feature of artificial intelligence technology in that a problem is solved by matching it with a class of problem of which the solution is known. For physically similar structures expressible in terms of structural indices, a prototype can be optimally proportioned directly by using design charts once the optimal proportion of a corresponding model is known. For a set of recommended input parameters such as strength and loading statistics, and a target system reliability, for a given offshore location which may be characterized by a design wave describable in terms of wave length and wave steepness, the approach is implemented in the generation of practical design charts for truss models of jacket structures. These charts may effectively facilitate layout selections in an early design situation.

1. INTRODUCTION

With the increasing maturity of optimization methods, current research activities seem to be directed toward the development of general, application-oriented optimization systems. Challenged by the complexities involved in shape and lay-out optimization of structural systems, this direction drives the emergence of concepts which integrate mathematical programming, finite element and sensitivity analysis into a computer-aided design environment [Kibsgaard et.al. (1990)]. The trend has actually been observed for some time with the birth of the so- called Computational structures Technology. This recent development can be characterized by the utilization of Artificial Intelligence (AI) capabilities in the form of rule-based expert systems. A biologically motivated trainable network paradigm called Artificial Neural Nets (ANN) can also be included in these AI generics [Berke and Haje1a (1990)]. The shape and lay-out optimization of structural systems seem to have directed the solutions to be obtainable within an increasingly hardware- and software-intensive environment.

This content is only available via PDF.
You can access this article if you purchase or spend a download.