An evolutionary approach for simultaneous optimization of material property distribution and topology of FG structures. Issue 2 (20th April 2015)

• Record Type: Journal Article
• Authors: Maleki Jebeli, Saeed; Shariat Panahi, Masoud
• Is Part Of: Engineering computations. Volume 32:Issue 2(2015)
• Page Start: 234
• Page End: 257
• Publication Date: 2015-04-20
• Subjects: BESO -- Element-wise material volume fraction -- Finite element method -- Functionally graded structures -- GA -- Topology optimization; Computer-aided engineering -- Periodicals; Computer graphics -- Periodicals; 620.00285
• DOI: 10.1108/EC-07-2013-0188
• Languages: English
• ISSNs: 0264-4401

Abstract

Abstract : Purpose: – The purpose of this paper is to introduce an evolutionary heuristic method for simultaneous optimization of topology and material property distribution of functionally graded (FG) structures under a prescribed loading condition. Design/methodology/approach: – The proposed procedure is inspired by heuristic nature of bi-directional evolutionary structural optimization (BESO) and genetic algorithm (GA). The optimization algorithm is developed in the context of minimum compliance (maximum stiffness) design problem. The problem is modeled by means of finite element method (FEM). The element-wise material volume fractions and elements' status (i.e. existence or nonexistence in FE model) are introduced as design variables. After FE analysis, sensitivities are obtain and filtered according to BESO. Having determined sensitivities, by means of a heuristic scheme combined by GA, topology and material property distribution for the next cycle of optimization are determined by updating design variables. Findings: – The adopted method has been tested by means of several examples previously reported in literature. The comparison showed the superiority of the proposed method against its rival in terms of relative reduction in compliance, smoother material property distribution and computational cost. Originality/value: – The value of the described method lies in its simple (yet efficient) nature. In contrast with its only rival in literature which more relied on mathematical approach, proposed method uses a series of logic-based heuristic ideas to drive reasonable solutions but with far less computational cost.