Research On Beam Structure Damage Recognition Based On Curvature Mode And GA-BP Neural Network Algorithm

Reinforced concrete beam structure as an important load-bearing component for basic projects such as road bridges and civil buildings,which are used in the fields of civil construction projects around the world.With the increase in the operation of civil engineering structure and its main load-bearing components,the beam structure usually occurs due to the influence of the construction quality,improper maintenance and the factors such as the maintenance of the external complex load.If can’t find damage and take effective measures to ensure structure security,it will bring huge economic losses and casualties.Therefore,it has significant practical significance to the beam structure in time to damage recognition and put forward reliable safety assessment opinions.This article is based on reinforced concrete simple beams as research objects.detailed research and analysis of the curvature modal and modal curvature of the modality and modal curvature of the modal curvature and the damage recognition method based on the GA-BP neural network algorithm,respectively.The main research content and important conclusions are as follows.(1)Summarize various types of impairment identification based on modal information.relying on the rapid development of computer software program algorithms,various damage recognition research methods based on artificial intelligence algorithms are proposed,and the advantages and deficiencies of these cutting-edge research methods are summarized.(2)The theoretical basis for the identification of beam structural damage recognition based on curvature modal and modal curvature is derived,and use finite element to calculate the curvature modal and modal curvature of the curvature under different injuries in the simple beam.Analysis found that the curvature modality produced a small mutation at the damage unit,which only realizes damage positioning.However,the difference between the modal curvature is high,and the maximum large value is generated at the damage unit,which can not only accurately realize damage positioning,but also realize qualitative quantitative analysis of damage.(3)The basic theory of the GA-BP neural network is introduced,and the algorithm is applied to the damage recognition of the bridge structure.Taking a simple support beam is a numerical example,and the GA-BP neural network core program code is prepared to realize the quantitative analysis of the damage positioning of simple beam.The study found that the degree of damage to the GA-BP neural network algorithm can automatically recognize the damage position.And when the degree of damage is 5% to 20%,the degree of damage to the damage is better,and a good positive linear relationship is presented,and the damage quantitative analysis can be achieved in a certain injury.(4)Finally,based on the reinforced concrete test beam,the reference model is finally based on the reinforced concrete test beam.Test beams are damaged by curvature modal and modal curvature difference and two damage recognition methods of GA-BP neural network algorithm.The experiment found that the same value calculation was the same,the curvature modal identification sensitivity is low,only damage positioning can only be achieved.The difference between modal curvature is high,and the degree of damage to the degree of damage can be achieved within a certain range.The damaged degree factor z obtained by the GA-BP neural network algorithm can automatically identify the damage position,but also better damage quantitative analysis within a certain damage range.