Topology Optimization of Multicell Tubes Under Out-of-Plane Crushing Using a Modified Artificial Bee Colony Algorithm

Publication Type:
Journal Article
Citation:
Journal of Mechanical Design, Transactions of the ASME, 2017, 139 (7)
Issue Date:
2017-07-01
Filename Description Size
MD-16-1115.pdfAccepted Manuscript Version3.8 MB
Adobe PDF
Full metadata record
Copyright © 2017 by ASME. Multicell tubal structures have generated increasing interest in engineering design for their excellent energy-absorbing characteristics when crushed through severe plastic deformation. To make more efficient use of the material, topology optimization was introduced to design multicell tubes under normal crushing. The design problem was formulated to maximize the energy absorption while constraining the structural mass. In this research, the presence or absence of inner walls were taken as design variables. To deal with such a highly nonlinear problem, a heuristic design methodology was proposed based on a modified artificial bee colony (ABC) algorithm, in which a constraint-driven mechanism was introduced to determine adjacent food sources for scout bees and neighborhood sources for employed and onlooker bees. The fitness function was customized according to the violation or the satisfaction of the constraints. This modified ABC algorithm was first verified by a square tube with seven design variables and then applied to four other examples with more design variables. The results demonstrated that the proposed heuristic algorithm is capable of handling the topology optimization of multicell tubes under out-of-plane crushing. They also confirmed that the optimized topological designs tend to allocate the material at the corners and around the outer walls. Moreover, the modified ABC algorithm was found to perform better than a genetic algorithm (GA) and traditional ABC in terms of best, worst, and average designs and the probability of obtaining the true optimal topological configuration.
Please use this identifier to cite or link to this item: