.png){kind=icon}
.png){kind=icon}
Return
Integration-Enhanced Active Learning Method with Composite Parallel Strategy for Structural Reliability Analysis
.png){kind=icon}
PRE
AI
DOI：10.1016/j.probengmech.2026.103952.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Accurate and efficient structural reliability analysis is essential for ensuring the safety, durability, and long-term performance of complex engineering systems throughout service life. However, existing methods often fall short in synergistically optimizing multiple learning criteria, constructing parallel computing schemes, and deeply integrating surrogate models within the analysis framework. To address these gaps, we propose an enhanced active learning reliability analysis approach. A novel composite learning function is developed to integrate multiple learning functions into a unified decision framework. This function improves the informativeness of selected samples. A sorting priority-based parallel computing strategy is then introduced to improve sample selection quality. This strategy selects multiple high-value samples in each iteration, significantly improving computational efficiency. To strengthen surrogate modeling, polynomial chaos–augmented Kriging (PC-Kriging) is employed to enhance global approximation accuracy. We further integrate this surrogate with subset simulation to better capture failure events. Performance evaluation on numerical benchmarks and a real engineering case shows that the proposed method maintains an average error below 1% and achieves an approximately 18-fold improvement in analysis efficiency compared with benchmark. The proposed method combines high accuracy and efficiency with computational scalability, offering a practical tool for reliability assessment and engineering decision support.
Keywords:
Structural reliability analysis
Active learning
Surrogate modeling
Parallel computing
Subset simulation
Journal
IF:
3.5
Papers:
1.7K
Citations:
4.1K
Organization
View more organizations