Topological design of structures and composite materials with multiobjectives

N. Kruijf, de, Q. Zhou, Q. Li, Y.W. Mai

Research output: Contribution to journalArticleAcademicpeer-review

146 Citations (Scopus)
1 Downloads (Pure)


This paper studies the influence of heat conduction in both structural and material designs in two dimensions. The former attempts to find the optimal structures with the maximum stiffness and minimum resistance to heat dissipation and the latter to tailor composite materials with effective thermal conductivity and bulk modulus attaining their upper limits like Hashin–Shtrikman and Lurie–Cherkaev bounds. In the part of structural topology optimization of this paper solid material and void are considered respectively. While in the part of material design, two-phase ill-ordered base materials (i.e. one has a higher Young’s modulus, but lower thermal conductivity while another has a lower Young’s modulus but higher conductivity) are assumed in order to observe competition in the phase distribution defined by stiffness and conduction. The effective properties are derived from the homogenization method with periodic boundary conditions within a representative element (base cell). All the issues are transformed to the minimization problems subject to volume and symmetry constraints mathematically and solved by the method of moving asymptote (MMA), which is guided by the sensitivities with respect to the design variables. To regularize the problem the SIMP model is explored with the nonlinear diffusion techniques to create edge-preserving and checkerboard-free results. The illustrative examples show how to generate Pareto fronts by means of linear weighting functions, which provide an in-depth understanding how these objectives compete in the topologies.
Original languageEnglish
Pages (from-to)7092-7109
JournalInternational Journal of Solids and Structures
Issue number22-23
Publication statusPublished - 2007


Dive into the research topics of 'Topological design of structures and composite materials with multiobjectives'. Together they form a unique fingerprint.

Cite this