结构地震响应序列-序列代理建模方法、系统及存储介质
The structural seismic response prediction framework using convolutional CNN+Transformer encoder solves the problems of low computational efficiency and insufficient long-range dependency modeling, and achieves efficient and physically consistent structural seismic response prediction, which is suitable for structural seismic analysis and post-earthquake assessment.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- QINGDAO UNIV OF TECH
- Filing Date
- 2025-11-13
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies suffer from low computational efficiency, insufficient long-range dependency modeling, and poor physical consistency in structural seismic response analysis. Traditional finite element methods involve large computational loads, while deep learning methods struggle to capture long-range time-series responses and exhibit physical inconsistencies.
A structural seismic response prediction framework using convolutional CNN + Transformer encoder is proposed. Local temporal features are extracted through a one-dimensional convolutional stack, and structural information is introduced by combining local residual blocks and skip convolutions. A global model is then performed using a Transformer encoder, and dynamic constraints are introduced to ensure physical consistency.
It achieves efficient long-range dependency modeling, improves computational efficiency, ensures physical consistency and prediction stability, and significantly improves the real-time performance and economy of seismic analysis.
Smart Images
Figure CN121500440B_ABST