About the author:

Resettlement Housing Support Center of the Logistics Support Department of the Central MilitaryCommission, Beijing 100036, China

Abstract:

To address the issue of insufficient accuracy in simulation analysis for structural safetyassessment, a digital twin modeling and analysis method is proposed. Firstly, a structural safetyassessment process based on mechanical responses, such as stress and displacement as evaluationindicators is established. On this basis, a digital twin model for structural safety assessment is formed,effectively integrating physical, data, model, analysis, and decision⁃making dimensions, and clarifyingthe integration mechanism of digital twin and deep learning. To achieve high⁃precision assessment ofstructural safety status, the deep learning algorithm is integrated with the digital twin model to form amodel update method. Under the drive of the genetic algorithm, the error between the measured andsimulated values of mechanical responses is used as the basis for model update. The location and degreeof structural damage are determined by comparing the changes in basic parameters before and after modelupdate. To verify the effectiveness of the proposed method, a steel structure example is taken as theresearch object and the proposed method is applied, achieving accurate and effective assessment of thestructural safety.