Domain Adversarial Network for Bridge Damage Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current bridge damage identification methods face challenges due to uncertainties such as modeling errors and measurement noise, leading to inaccuracies in damage detection between finite element models and real bridge structures.
Innovation Solution
A domain adversarial network is used to find common features between bridge finite element models and real bridge structures, allowing for damage identification without requiring damage labels for the real structure and implicitly addressing uncertainty factors, using only acceleration response signals for efficient damage extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If model-based damage identification methods are used with finite element models, then damage identification can be performed, but modeling errors and measurement noise cause large variations in damage indicator accuracy
Solution Approach 1:
The patent creates a virtual copy of the real bridge structure through finite element modeling. The domain adversarial network learns to map features from the virtual model domain to the real structure domain, enabling damage identification on the real structure using data generated from the virtual model without requiring direct labeled data from the real structure.
Solution Approach 2:
The domain adversarial network serves as an intermediary between the finite element model and the real bridge structure. It translates features from the virtual domain to the real domain, bridging the gap caused by modeling errors and measurement noise, and enabling accurate damage identification without direct correspondence between model and reality.
2Measurement precision
If all uncertainty factors in the real bridge structure are modeled, then comprehensive damage identification is achieved, but the complexity of the finite element model becomes unmanageable
Solution Approach 1:
The patent extracts only the essential features needed for damage identification from the complex reality, rather than attempting to model all uncertainty factors. The domain adversarial network learns to identify and extract discriminative features that are sufficient for damage detection, ignoring unnecessary complexities in the finite element model.
Data Source
AI summary
A bridge damage identification method considering uncertainty is used for damage identification based on a convolutional neural network. A domain classifier is added to form a domain adversarial transfer network, a finite element model of a bridge and a time domain acceleration signal of a real structure serve as input, and parameters in a feature extractor are continuously updated in an adversarial process of the domain classifier and the feature extractor, so as to design a brand-new feature extractor, and to achieve a purpose that extracted features are only sensitive to damage. The bridge damage identification method can solve the problem that model-based methods for bridge damage identification are influenced by environment uncertainty or modeling error to generate a difference between the finite element model and the real structure, resulting in reduction in damage identification performance of the method in practical application.