Analysis of Seismic Elastic Steel Structures Using a New Hybrid Optimization Deep Learning Model

In response to the insufficient performance of traditional seismic elastic analysis models for steel structures, this study proposes a deep layered feature learning network model in a mixed optimization point set metric space by combining cross attention mechanism, graph neural network, and random s...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE access Jg. 13; S. 1
Hauptverfasser: Li, Ping, Liu, Wanfeng, Yan, Tiecheng
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Piscataway IEEE 01.01.2025
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Schlagworte:
Accuracy
Adaptation models
Analytical models
Computational modeling
Cross attention
Deep learning
Deformable models
Deformation
Elastic analysis
Elasticity testing
Feature extraction
Graph neural networks
Machine learning
Matching
Metric space
Noise reduction
Noise sensitivity
Occlusion
Optimization
Peak values
Point cloud compression
PointNet
Random sampling
RANSAC
Real time
Seismic analysis
Semantic segmentation
Steel
Steel pipes
Steel structure
Steel structures
ISSN:2169-3536, 2169-3536
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:In response to the insufficient performance of traditional seismic elastic analysis models for steel structures, this study proposes a deep layered feature learning network model in a mixed optimization point set metric space by combining cross attention mechanism, graph neural network, and random sampling consistency algorithm. The experiment showed that the average F1 score of this model in cloud matching tasks was 0.919, which was 16.5%-22.6% higher than traditional methods. Moreover, the matching success rate (55.43%) was still higher than the comparison method by more than 15% at a 50% occlusion rate. In semantic segmentation tasks, the intersection to union ratio (0.798) and Dice coefficient (0.879) of steel pipe components were higher than traditional methods. In actual scenario testing, the model achieved sub-millimeter level deformation detection (the minimum deformation detection amount for steel pipes and welded balls was 0.44mm and 0.31mm). The generalization error (0.06mm) and memory peak value (11.3GB at 10K points) were reduced by 76.9% and 53.9% compared to the comparison method, which is suitable for real-time monitoring needs of edge devices. In summary, the proposed model addresses the shortcomings of traditional methods in dynamic interaction, noise suppression, and deformation sensitivity through multimodal feature fusion and hybrid optimization strategies. This provides a high-precision and low-energy solution for seismic analysis of complex steel structures.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:2169-3536
2169-3536
DOI:10.1109/ACCESS.2025.3601409