Study on Inversion of Damage Incentives of High Pile Wharf in Inland River Based on SEResNet

Abstract

: Based on SEResNet neural network algorithm, the inversion model of damage incentives of inland high-piled wharf is constructed. The stress data of pile foundation under the action of damage incentives of high-piled wharf are obtained by using solid element finite element model calculation and indoor model test methods. The parameterized finite element calculation model of high-piled wharf is established by using subprocess in Python program to call MANSYS module, and verified with solid element model. The parameterized simplified finite element model meets the needs of inversion calculation. Based on the stress data samples of the pile foundation of the high-piled wharf obtained from the model test, the inversion analysis of single and multiple damage incentives is carried out. The model can identify the location, size and type of injury causative agent with good generalization ability.