Seismic Safety Assessment Method Of Building Structures Based On Low-cost Inter-story Displacement Monitoring Methods

Earthquake disasters do great harm to urban buildings and always cause a large number of casualties and incalculable economic losses.Whether in the building structure design and vulnerability assessment before an earthquake,structure monitoring during an earthquake,or damage assessment after an earthquake,inter-story displacement(or inter-story drift ratio)is an important structural damage measure.If the inter-story displacement time history of each building can be recorded in real time under earthquakes,the efficiency of post-earthquake safety assessment of building structures can be greatly improved.Besides,buildings’ damage states under new seismic events could be predicted through analyzing the accumulated building seismic damage data.However,due to the high cost of structural health monitoring systems at present,they are only installed in the few important buildings.As a result,it still needs manual investigation and statistics to evaluate the damage states of most buildings after earthquakes,which costs a lot of manpower and resources,and the efficiency is low.In view of this situation,how to evaluate the vulnerability of a large number of building structures before earthquakes,how to monitor the inter-story displacement at a low-cost during earthquakes,and how to predict the damage states and maximum inter-story drift of structures after earthquakes are the main lines of this paper.The research is carried out around an important engineering parameter:inter-story displacement of building structure.At last,a fast seismic safety assessment method of building structures based on low-cost inter-story displacement monitoring methods is formed.The main contents of this paper are as follows:(1)Based on the Multi-story Concentrated-mass Shear(MCS)model,using incremental dynamic analysis to analyze the vulnerability of the common types of urban buildings.Use incremental dynamic analysis to establish the MCS models of 132 buildings with different structure types,stories,and code versions.The seismic responses of each model on five fields with different seismic intensities are calculated.As a result,the vulnerability curves of different types of buildings are obtained.Besides,two vulnerability visualization methods are designed to visualize the vulnerability level of regional buildings.In addition,this paper investigates the buildings in a small region of Dalian city,and assess the seismic vulnerability of buildings in this region using the vulnerability results of incremental dynamic analysis.(2)A low-cost inter-story displacement monitoring method based on smartphone and a laser device is proposed.In this method,a low-cost laser device fixed in the ceiling is used to emit signals,and the laser point is illuminated on the projection board fixed on the floor.The laser point is photographed by a smartphone.The movement of the laser point is calculated by the application D-viewer integrated with computer vision algorithms,and the inter-story displacement time history could be calculated.A total of 19 experiments including harmonic excitation experiment,seismic excitation experiment and night experiment are conducted to verify the robustness of the method and analyze the accuracy of the method.The results show that the method has good accuracy,small percentage error in a real story height.(3)A simple method of inter-story displacement monitoring using a smartphone based on feature point matching algorithm is proposed.To monitor the inter-story displacement more conveniently,a method of calculating the inter-story displacement of the structure is designed by pasting the smartphone on the floor and photographing the ceiling objects with the front camera.The computer vision algorithm for video processing is feature point matching algorithm,which can obtain the movement of feature points of ceiling objects.Through 13 groups of static,dynamic and steel moment resisting frame experiments,the feasibility and robustness of the method under a real floor height are verified.Besides,the measurement accuracy of the method,the reasons for the interruption of monitoring results under low light conditions and the dynamic characteristics of the steel frame are analyzed in turn.The results show that the method has advantages of low cost,convenient operation,and good accuracy.(4)Based on principal component analysis(PCA)and artificial neural network(ANN),a fast prediction method for post-earthquake damage states of regional buildings is proposed.In order to more comprehensively characterize the ground motions,58 ground motion intensity measures(IM)of two horizontal directions of seismic events are selected.The seismic IMs are visualized after PCA.Besides,the dimension of 58 seismic IMs is reduced to 20 by PCA,which retains most of the original information and reduces the correlation.The dimension reduced seismic IMs,building structure characteristics,and building damage states are imported into an ANN for training.The damage states of all kinds of buildings under new ground motions are predicted using ANN.The results show that the method can predict the damage states of regional buildings after a new earthquake very well.Finally,this method is used to predict the damage states of buildings in a small region of Dalian under a new ground motion.(5)Based on inter-story drift spectrum and convolutional neural network,a method of predicting the maximum inter-story drift ratio(MIDR)of structures is proposed.A data-driven method integrating the inter-story drift spectrum is proposed.The inter-story drift spectrum is taken as the first estimation of the MIDR of the structure.A convolutional neural network which is similar to an autoencoder is designed.The network,which is trained with the structure seismic response data,could adjust the first estimation to generate a more accurate second estimation.Two structural models are used to verify the method.The results show that this method is superior to ANN and support vector machine models and give a reasonable prediction of the MIDR.(6)Based on the above five parts,the seismic safety evaluation method of urban building structures is formed.The method includes three parts:the first part evaluates the earthquake vulnerability of urban buildings before the earthquake;the second part monitors the inter-story displacement time history of each building during the earthquake;the third part predicts and evaluates the damage state of each building after the earthquake.Besides,relationships among the three parts are analyzed,and the following problems are analyzed according to the actual application scenarios:explain the data collection method of buildings’earthquake damage database;use the vulnerability analysis results in the first part and K-Means to provide the monitoring system layout suggestions of regional buildings;analyze the Miranda’s composite beam model to provide the monitoring system layout suggestions of a single building;introduce the data processing methods of main earthquake and aftershock sequence.