Research On Earthquake Damage Of Frame Structure Crossing Ground Crack Based On Neural Network

Ground fissures,as a phenomenon of superficial geological disasters,have a long formation time and a wide distribution range,affect the normal function of buildings,and have a huge impact on human production and life.However,at present,there is a lack of research data on the seismic damage of the ground fissure site structure.Therefore,it is of great theoretical significance and practical engineering value to carry out research on seismic damage assessment across ground fissure structures,and to use artificial intelligence technology to predict structural damage status,eliminate hidden dangers and formulate effective reinforcement measures,make full use of land resources,and relieve urban planning constraints.With the continuous rise of interdisciplinary,neural network technology is widely used in the field of building structure,because of its strong nonlinear mapping ability,associative guessing ability and robustness.Therefore,it provides a new approach to the detection and identification of structural damage,and to predict the further development of damage.In this paper,a two-parameter damage model is used to study the damage of the whole frame and components across the ground fissure.Based on the shaking table test and numerical simulation results,the neural network technology is used to predict and analyze the structural damage.The main research work is as follows:First of all,starting from the theory of structural dynamics,the structural damage indicators based on frequency,mode,deformation and energy are systematically introduced,the development process of the weighted combination method and weighting coefficient for evaluating the overall structural damage is briefly described,and the seismic damage at various levels is described.Correspondence between its damage index.At the same time,three common prediction models of neural network,Markov and gray GM(1,1)are compared,and a neural network prediction model suitable for identifying structural damage is obtained.Secondly,this paper carried out a shaking table test of a frame structure that crosses a ground fissure,describes the test phenomenon,and analyzes the earthquake damage level.At the same time,the ABAQUS finite element software was used to establish the corresponding cross-fracture structure and soil model,and the numerical simulation results were compared with the test data to analyze the acceleration and interlayer displacement of the upper frame structure of the ground fissure site under different working conditions.The response law verified the correctness of the finite element model.Then,the establishment process of BP neural network model,parameter setting and data normalization and the method of adding noise are summarized in detail.A PFPB prediction algorithm based on BP neural network is proposed,and the three types of test data are combined with BP neural network.The gradient descent algorithm is used to train and predict the neural network,which verifies the rationality and accuracy of the BP neural network established in this paper.Finally,based on the simulation results of the ABAQUS finite element model and the dual-parameter damage index,the whole process of seismic damage across the ground fissure frame structure is studied from the three levels of component-floor-structure.At the same time,combined with the BP neural network technology,the PDB algorithm is proposed,which uses the inherent frequency to warn whether the damage occurs,uses the two-parameter damage index to determine the location and degree of the damaged component,and trains according to the damage index of the component and the overall structure to predict the further development trend of the structural damage.