.png){kind=icon}
.png){kind=icon}
Return
Reliability-based global sensitivity updating method considering new input observation information
.png){kind=icon}
PRE
AI
X
X
L
DOI：10.1016/j.probengmech.2026.103983.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Reliability-based global sensitivity analysis (GSA) aims to quantify the influence of each model input on structural reliability, thereby laying the foundation for simplifying reliability-based design and optimization models. Existing research on reliability-based GSA often assumes that the stochastic characterization model of the model inputs is deterministic. Consequently, it has not yet considered how sensitivity is affected when new input observation information becomes available. To address this, this paper proposes a Bayesian updating method for reliability-based GSA applicable to scenarios with new input observation information. First, based on Bayesian updating theory, the posterior stochastic characterization model of the model inputs under the condition of new input observation information is constructed, and the posterior failure probability of the structure is further derived. Next, by measuring the average difference between the posterior failure probability and the conditional posterior failure probability across the entire model inputs space, a posterior sensitivity index is defined. Finally, combining a single-loop numerical simulation technique with an active-learning Kriging model, an efficient algorithm for solving this index is investigated. The Bayesian updating framework established in this paper can dynamically update structural sensitivity as new input observation information becomes available. This provides real-time decision-making support for dimensionality reduction in reliability-based design and optimization models.
Journal
IF:
3.5
Papers:
1.7K
Citations:
4.1K
Organization
No organization information available
Cited Papers