TY - JOUR
T1 - Seismic topology optimization considering first-passage probability by incorporating probability density evolution method and bi-directional evolutionary structural optimization
AU - Yang, Jia Shu
AU - Chen, Jian Bing
AU - Beer, Michael
N1 - Publisher Copyright:
© 2024 Elsevier Ltd
PY - 2024/9/1
Y1 - 2024/9/1
N2 - This contribution focuses on addressing the challenging problem of dynamic-reliability-based topology optimization (DRBTO) of engineering structures involving uncertainties by synthesizing the probability density evolution method (PDEM) and the bi-directional evolutionary structural optimization (BESO) approach. The considered optimization problem aims at minimizing the first-passage probability under the constraint of material volume. Generally, the double-loop essence of DRBTO involving dynamic reliability evaluation and topology searching makes the computational efforts prohibitively large. To this end, the PDEM is adopted to efficiently assess the first-passage probability of structures under earthquake actions. In particular, by reformulating the first-passage probability under the framework of the PDEM, the sensitivity of the first-passage probability is derived. To further improve the efficiency, a strategy taking advantage of important representative points (IRPs) is employed to achieve a robust estimate of sensitivity of the first-passage probability. The adjoint variable method (AVM) for the sensitivity analysis of transient response considering given modal damping ratios is incorporated to considerably improve the computational efficiency when the reliability sensitivity analysis in terms of multiple design variables is needed. To drive the topology towards the optimum, the above highly efficient reliability assessment and sensitivity analysis are embedded into BESO. Finally, numerical examples are presented to demonstrate the effectiveness of the proposed method, illustrating significant improvement in computational efficiency compared to direct implementation. Additionally, the necessity of introducing seismic reliability in topology optimization is also discussed based on the numerical results.
AB - This contribution focuses on addressing the challenging problem of dynamic-reliability-based topology optimization (DRBTO) of engineering structures involving uncertainties by synthesizing the probability density evolution method (PDEM) and the bi-directional evolutionary structural optimization (BESO) approach. The considered optimization problem aims at minimizing the first-passage probability under the constraint of material volume. Generally, the double-loop essence of DRBTO involving dynamic reliability evaluation and topology searching makes the computational efforts prohibitively large. To this end, the PDEM is adopted to efficiently assess the first-passage probability of structures under earthquake actions. In particular, by reformulating the first-passage probability under the framework of the PDEM, the sensitivity of the first-passage probability is derived. To further improve the efficiency, a strategy taking advantage of important representative points (IRPs) is employed to achieve a robust estimate of sensitivity of the first-passage probability. The adjoint variable method (AVM) for the sensitivity analysis of transient response considering given modal damping ratios is incorporated to considerably improve the computational efficiency when the reliability sensitivity analysis in terms of multiple design variables is needed. To drive the topology towards the optimum, the above highly efficient reliability assessment and sensitivity analysis are embedded into BESO. Finally, numerical examples are presented to demonstrate the effectiveness of the proposed method, illustrating significant improvement in computational efficiency compared to direct implementation. Additionally, the necessity of introducing seismic reliability in topology optimization is also discussed based on the numerical results.
KW - Bi-directional evolutionary structural optimization
KW - First-passage probability
KW - Probability density evolution method
KW - Sensitivity analysis
KW - Topology optimization
UR - http://www.scopus.com/inward/record.url?scp=85195799339&partnerID=8YFLogxK
U2 - 10.1016/j.engstruct.2024.118382
DO - 10.1016/j.engstruct.2024.118382
M3 - Article
AN - SCOPUS:85195799339
VL - 314
JO - Engineering structures
JF - Engineering structures
SN - 0141-0296
M1 - 118382
ER -