MICROSTRUCTURE OPTIMISATION
algorithmic design of space frames

date: 2003 - present
(in collaboration with Siavash Haroun Mahdavi)

The focus of this research is on structural design. In collaboration with Siavash Haroun Mahdavi (also at UCL), this work has yielded a method for optimising the microstructure of objects made by stereolithography, a rapid prototyping technology. This method maximises the strength of an object while minimising its weight, by placing material in the areas required to best withstand the external forces applied, similar to the organic microstructures in natural wood or bone.

A series of linear structural members acting in either tension or compression traverse the volume of the object to be made and meet at node points, much like a 3D space frame, only rather than consisting of identical members at fixed angles from one another, their position and orientation vary continuously. A genetic algorithm is used to determine a module's topology as defined by structural members and their connections to one another, and a deterministic process finds the ultimate shape. This underlying module is evolved to suit the material and machine: its topology is evolved in response to the specific fabrication process of stereolithography and the range of forces, and its shape is modified to suit the object and specific stress, flexibility or other required properties as they change from point to point. The module maintains a constant topology and as such connects seamlessly to its neighbours while changing shape. An item that would have been composed of many jointed parts by any other process may be produced as a seamless whole.

There are several principle advantages of the structural optimisation method that make it of genuine practical interest: it is based on the constraints of a real fabrication process, taking material constraints into account in the algorithm, and it is designed to be easily scalable to objects of large size and complexity. Many other optimisation methods call for drastically increased computation time as the complexity of a structure increases, making many real world problems infeasible, but our method scales linearly with the size of the objects, so the structure can be designed as fast as it can be fabricated.

For a further description of the methods and results of the research, please see the following papers:

Hanna, S and Haroun Mahdavi, S. (2004) Modularity and Flexibility at the Small Scale: Evolving Continuous Material Variation with Stereolithography. In Beesley P. Cheng W. and Williamson R Eds. Fabrication: examining the digital practice of architecture. University of Waterloo School of Architecture Press, Toronto.
HTML | PDF

Haroun Mahdavi, S and Hanna, S. (2004) Blurring the Boundaries between Actuator and Structure: Investigating the use of Stereolithography to build Adaptive Robots, Proceedings of ICARCV 2004.
HTML | PDF

Haroun Mahdavi, S and Hanna, S. (2004) Optimising Continuous Microstructures: A Comparison of Gradient-Based and Stochastic Methods, Proceedings of SCIS & ISIS 2004, The Joint 2nd International Conference on Soft Computing and Intelligent Systems and 5th International Symposium on Advanced Intelligent Systems.
HTML | PDF

Haroun Mahdavi, S and Hanna, S. (2003) An Evolutionary approach to microstructure optimisation of stereolithographic models. Proceedings of CEC203. The Congress on Evolutionary Computation.
HTML | PDF