Design and Research on Modification Method of Finite Element Dynamic Model of Concrete Beam Based on Convolutional Neural Network

Abstract

Because of the uncertainty of the measurement results, the dynamic inverse problem equation is called the stochastic model correction equation. In order to make this equation or method can be used smoothly in actual engineering, it is aimed at low modal orders and small samples in actual engineering. Condition, this paper proposes a concrete beam finite element dynamic model correction method based on convolutional neural network technology. This method trains the convolutional neural network algorithm by expanding the data set, so that the accuracy of the finite element dynamic model correction method is obtained. Improve, and the method is feasible in actual engineering.