Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.511410
Title: Design optimisation of steel portal frames using modified distributed genetic algorithms
Author: Issa, Honar Khoshavi
Awarding Body: Nottingham Trent University
Current Institution: Nottingham Trent University
Date of Award: 2010
Availability of Full Text:
Access from EThOS:
Access from Institution:
Abstract:
Distributed genetic algorithms have been modified in this study to improve their quality, performance, and convergence to the optimum solution for structural steel frames. This was achieved by introducing some novelties of the main algorithm of distributed genetic algorithms and applying them in structural optimisation. Among these are the creation of new mutation schemes, adding a crossover scheme, definition of a penalty function, properties of twins, and definition of the reproduction scheme. Many optimisation problems have been designed to minimise the weight of a steel structure and are well documented in the literature. However, having a frame controlled by displacement will necessitate choosing a different approach for the objective function. In addition to weight minimisation, attempts have been made to investigate displacement maximisation and this also forms part of the novelty of this study. Various steel frames in terms of geometry and loading conditions are considered during the optimisation process and they are assumed to have rigid and/or semi-rigid connections. The design optimisations are conducted according to therequirements of both BS 5950 and EC3 codes of practice. A stiffness matrix is developed for a non-prismatic member that is involved in the analysis process. A program DO-DGA, written in Visual Basic 6.0, has been developed to include all aspects of the modified distributed genetic algorithms as well as the decided terms for the analysis, such as the types of connections, the geometry of members, and the type of design problem; minimisation or maximisation. The performance of the developed algorithm is validated by comparing the optimum solutions obtained with the results published in the literature. The results of tests indicate that the developed algorithm is robust and efficient in seeking the optimum solutions within a reasonable time. They also reveal that the weight minimisation outperforms the displacement maximisation and the haunchedrafter steel portal frame with rigid connections yields a lighter frame than one with semi-rigid connections. In general, the design optimisation according to EC3 Design Optimisation of Steel Portal Frames Using Modified Distributed Genetic Algorithms v demonstrates that a lighter frame can be achieved comparing with the design optimisation according to BS 5950.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.511410  DOI: Not available
Share: