Using Backpropagation Neural Network To Correct Bridge Finite Element Model

As China's infrastructure capacity has steadily improved,bridge projects are springing up.They can be seen in the mountains and rivers across the country.Today,in the area of bridge engineering,researchers have been paying attention to static and dynamic response during the bridge's life span.The combination of mathematics and computer development has made the finite element method widely used in structural analysis.The numerical analysis tool of the finite element method is low in cost and speed.When the bridge finite element model is established based on the design data,the error between the model-calculated value and the measured value is large because the model is simplified and assumed.Therefore,it is necessary to correct the finite element model to improve its accuracy.Based on the monitored engineering structural parameters,this paper uses ABAQUS finite element software to establish the bridge finite element initial model and uses the finite element model modification technology based on BP neural network algorithm to modify and optimize the relevant parameters of the bridge model.The modified finite element model approximates the real system,so that the finite element model has a clear physical meaning.In order to achieve this process,this article has mainly done the following work:(1)The rapid development of the finite element method leads to the background and significance of the correction of the finite element model,proposes the research background and significance of this article,and introduces the overview of the finite element model correction technology.(2)Introduce the principles and processes involved in model revision.In addition,focus on the important steps involved in the process to focus on elaboration,and finally,put forward the key research content of this article,artificial neural network algorithm in the application of finite element model correction and do a final summary.(3)Introduce the important algorithm artificial neural network algorithm used in this paper,first briefly describe its basic theory and working principle,and clarify why artificial neural network can be used for finite element model correction.Next,it focuses on the artificial neural network algorithm.The BP algorithm describes the detailed steps of application of the BP algorithm and lays a foundation for subsequent analysis of engineering algorithms.(4)Using MATLAB as a platform,use the correction algorithm proposed in this paper to practice an example.For practical problems,the steps of structural design of BP neural network and the homogenization of sample data introduced in detail.Finally,using a MATLAB platform to perform a programmatic prediction of an example,the prediction results obtained,which are consistent with the real data.(5)The finite element model modification method based on BP neural network algorithm applied to a restressed concrete bridge,and the finite element model was corrected according to the abovementioned method and operation steps,and the control error was made to meet the engineering requirements and the purpose of correction is accomplished.