TY - JOUR
T1 - Combined sizing, shaping and topology optimization of truss structure using evolutionary structure optimization
AU - Liu, Tao
AU - Deng, Zichen
PY - 2004/12
Y1 - 2004/12
N2 - Ref. 8 did sizing and shaping optimization of truss structure using evolutionary structure optimization. We believe that in addition to sizing and shaping, topology should also be included in optimization process as such addition can get much better optimization results. In this paper, we present an evolutionary optimization method to combine sizing, shaping and topology design of truss structure with aim of weight minimum. The structure is subject to stress, local buckling and displacement constraints. We optimize three types of design variables: (1) as regards sizing/shaping-variables optimization, we do what Ref. 8 did; a full stressed design (FSD) and scaling techniques are applied to sizing variables; the evolutionary node shift method is applied to shaping variables; the two variables are optimized together; (2) the evolutionary killing method, proposed, explained and named by us, is applied to topology variables. An alternating procedure is utilized to couple the three types of design variables and combine the results. The optimum solution is the smallest weight structure in the whole evolutionary process. A 12-element truss, as numerical example, confirms the validity of the method. This example also shows that the optimum solution is not usually the structure with the least elements.
AB - Ref. 8 did sizing and shaping optimization of truss structure using evolutionary structure optimization. We believe that in addition to sizing and shaping, topology should also be included in optimization process as such addition can get much better optimization results. In this paper, we present an evolutionary optimization method to combine sizing, shaping and topology design of truss structure with aim of weight minimum. The structure is subject to stress, local buckling and displacement constraints. We optimize three types of design variables: (1) as regards sizing/shaping-variables optimization, we do what Ref. 8 did; a full stressed design (FSD) and scaling techniques are applied to sizing variables; the evolutionary node shift method is applied to shaping variables; the two variables are optimized together; (2) the evolutionary killing method, proposed, explained and named by us, is applied to topology variables. An alternating procedure is utilized to couple the three types of design variables and combine the results. The optimum solution is the smallest weight structure in the whole evolutionary process. A 12-element truss, as numerical example, confirms the validity of the method. This example also shows that the optimum solution is not usually the structure with the least elements.
KW - Sizing/shaping optimization
KW - Topology optimization
KW - Truss structure
UR - http://www.scopus.com/inward/record.url?scp=14844364009&partnerID=8YFLogxK
M3 - 文章
AN - SCOPUS:14844364009
SN - 1000-2758
VL - 22
SP - 739
EP - 743
JO - Xibei Gongye Daxue Xuebao/Journal of Northwestern Polytechnical University
JF - Xibei Gongye Daxue Xuebao/Journal of Northwestern Polytechnical University
IS - 6
ER -