TY - JOUR
T1 - Decomposed-coordinated framework with intelligent extremum network for operational reliability analysis of complex system
AU - Jia-Qi, Liu
AU - Yun-Wen, Feng
AU - Cheng, Lu
AU - Wei-Huang, Pan
N1 - Publisher Copyright:
© 2023
PY - 2024/2
Y1 - 2024/2
N2 - The analysis of operational reliability in complex systems, which involve numerous subsystems and multiple disciplines, presents significant computational challenges due to their highly nonlinear, transient nature and the presence of many hyperparameters. Although reliability analysis models have made progress, they are still inadequate for accurately modeling composite functions with multiple sublayers and sub-functions. To improve the performance of modeling composite functions, the decomposed-coordinated intelligent extremum network model (DC-IENM) is proposed in this paper. The present study employs the decomposed-coordinated (DC) strategy as a means to effectively address the coordination relationship among multiple analysis objectives. To assess the efficacy of the proposed approach, two illustrative examples are considered: (1) the approximate and probabilistic analysis of a nonlinear function with multiple responses, and (2) the reliability analysis of civil aircraft brake system temperature. These examples serve to demonstrate the effectiveness of the developed DC-IENM. Furthermore, the modeling and simulation properties are rigorously examined by means of a comparative analysis involving various methodologies. The obtained results unequivocally indicate that the proposed DC-IENM exhibits distinct advantages in terms of both computational efficiency and precision.
AB - The analysis of operational reliability in complex systems, which involve numerous subsystems and multiple disciplines, presents significant computational challenges due to their highly nonlinear, transient nature and the presence of many hyperparameters. Although reliability analysis models have made progress, they are still inadequate for accurately modeling composite functions with multiple sublayers and sub-functions. To improve the performance of modeling composite functions, the decomposed-coordinated intelligent extremum network model (DC-IENM) is proposed in this paper. The present study employs the decomposed-coordinated (DC) strategy as a means to effectively address the coordination relationship among multiple analysis objectives. To assess the efficacy of the proposed approach, two illustrative examples are considered: (1) the approximate and probabilistic analysis of a nonlinear function with multiple responses, and (2) the reliability analysis of civil aircraft brake system temperature. These examples serve to demonstrate the effectiveness of the developed DC-IENM. Furthermore, the modeling and simulation properties are rigorously examined by means of a comparative analysis involving various methodologies. The obtained results unequivocally indicate that the proposed DC-IENM exhibits distinct advantages in terms of both computational efficiency and precision.
KW - Approximate modeling
KW - Decomposed-coordinated surrogate model
KW - Machine learning method
KW - Multi-failure mode
KW - System operational reliability
UR - http://www.scopus.com/inward/record.url?scp=85175248118&partnerID=8YFLogxK
U2 - 10.1016/j.ress.2023.109752
DO - 10.1016/j.ress.2023.109752
M3 - 文章
AN - SCOPUS:85175248118
SN - 0951-8320
VL - 242
JO - Reliability Engineering and System Safety
JF - Reliability Engineering and System Safety
M1 - 109752
ER -