一种基于双驱动策略的焊接结构疲劳寿命预测方法及系统
By combining a deep convolutional neural network and a physical constraint dual-drive strategy, the problem of insufficient generalization ability in fatigue life prediction of welded structures is solved, and accurate fatigue life prediction of engineering welded structures under various service conditions is achieved. It is applicable to fatigue performance analysis of welded structures under various complex working conditions.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- TIANJIN UNIV
- Filing Date
- 2024-09-02
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies have poor generalization ability in predicting the fatigue life of welded structures, making it difficult to guarantee the reliability of welded structures. In particular, the data set of fatigue behavior fluctuates greatly under multiple changing conditions, making it difficult to build a unified fatigue life prediction model.
A dual-drive strategy based on deep convolutional neural networks (DCNN) is adopted, combining physical constraints and data-driven methods. By introducing an objective function model, physical constraints such as material strength, average stress, weld geometry, load conditions, and size effects are generated to construct a fatigue life prediction model. Data augmentation and feature engineering techniques are used to optimize the dataset.
It enables accurate and stable fatigue life prediction of welded structures under different service conditions, reduces the resource consumption of fatigue tests, provides more reliable fatigue design support, and is applicable to fatigue performance analysis of welded structures under various complex working conditions.
Smart Images
Figure CN119294223B_ABST