Cost-based Structural Optimisation

Research Theme: Computational Design

Structural optimisation research has traditionally addressed functionality aspects of design, such as mass minimisation. If cost of manufacture was also considered, then it is likely that there will be a trade-off between cost and structural performance.

This project seeks to establish a methodology by which manufacturing cost can be considered directly within a structural optimisation framework.

Motivation

There is growing concern in the UK over competition from countries with lower manufacturing cost bases. Better product design can, however, achieve cost savings that far outweigh those achieved through other means, since 70-80% of the cost of manufacturing a product is determined by design. The aim of this project is to leverage the potential of structural optimisation in manufacturing industry by fundamentally shifting the focus of optimisation to a cost-based objective, while retaining performance and manufacturing considerations.

Objectives

  • To create comprehensive, realistic cost models for major manufacturing processes.
  • To develop a technique that will automatically recognise cost-contributingfeatures in a design.
  • To integrate these capabilities into a structural optimisation loop.

Method

The first objective entails developing a cost model relating structural details (design features) to their cost of manufacture. In this project injection moulding is used as an exemplar manufacturing process.

Medial object transformation technology is being used to tackle the problem of recognising automatically salient features of a solid model representation of a design.

The integration of these capabilities into a structural optimisation system will then test the viability of this approach for real-world design tasks.

Findings

Early results indicate that this approach can successfully differentiate between candidate designs on the basis of manufacturing cost, thus paving the way for potential automatic cost-based optimisation of designs.

Acknowledgements

Support is provided by TranscenData Europe Ltd.

Support for this project is provided by EPSRC.

Selected Publications

  • YAO, N., PARKS, G.T., CLARKSON, P.J. (2006) Automaticcost estimation of injection-moulded components using medial object transformationtechnology feature recognition, Proceedings of the 6 th ASMO-UK/ISSMOInternational Conference on Engineering Design Optimisation 2006, Oxford, UK.