Structural Damage Identification of Convolutional Neural Networks Based on Self Attention
Traditional damage identification techniques rely on artificial feature engineering and machine learning techniques.The application of deep learning in modern structural damage i-dentification has significantly improved accuracy and efficiency.The article proposes a convolu-tional neural network based on attention mechanism,based on existing deep learning tech-niques.Experimental verification on a three span continuous bridge shows that the error of sin-gle point damage recognition is 2.4％,while the error of multi-point damage recognition is less than 5％.Therefore,this deep learning based method can effectively solve the problem of structural damage identification.
Deep learningBridge structureIdentification of damage severity
NETL
NSTL
万方数据
