.png){kind=icon}
.png){kind=icon}
Return
Adaptive optimization deep neural network framework of reliability estimation for engineering structures
.png){kind=icon}
PRE
AI
DOI：10.1016/j.istruc.2024.106621.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
To improve the accuracy and efficiency of reliability estimation for engineering structures, the adaptive optimization deep neural network framework (DNN-AO) is proposed by introducing the coati optimization algorithm-based disturbance strategy, momentum gradient descent, and weighted Bayesian optimization technology into the deep neural network model. In this framework, the deep neural network model is employed to establish the relationship between input variables and output response; the coati optimization algorithm-based disturbance strategy is adopted to obtain the initial weights and thresholds; the momentum gradient descent is utilized to solve the final initial weight and threshold; the weighted Bayesian optimization strategy is used to determine the optimal neural network topology. In addition, the nonlinear function approximation and aeroengine turbine blisk strain reliability analysis cases validate the effectiveness of the proposed DNN-AO with multiple methods. The reliability level of aero-engine turbine blisk strain is 0.9985 when the allowable value of strain is 5.298 x 10 -4 . The results show that the DNN-AO exhibits the advantages of modeling properties (i.e. modeling accuracy and efficiency) and simulation performances (i.e., simulation precision and efficiency) in various methods. The research work in this paper can enrich the reliability design theory of engineering structures and further guide the optimization design of aero-engine turbine blisk.
Keywords:
Adaptive optimization deep neural network
framework
Engineering structures
Reliability estimation
Coati optimization algorithm-based distur
bance strategy
Weighted Bayesian optimization technology
Journal
IF:
4.3
Papers:
1.2W
Citations:
2.7W