Researsh On Nonlinear Prediction Of Blasting Seismic Effect Of Frame Structure

Along with the continuous development of the city,the subway has become a hot traffic tool for most cities.The blasting excavation of the subway tunnel brings convenience to the construction workers,and its adverse effects on adjacent buildings have also aroused widespread concern.At present,construction structure damage accidents caused by blasting vibration often occur.Therefore,how to predict and reduce the blasting seismic effect of the structure while achieving the purpose of blasting construction has an important guiding role in adopting safe and effective blasting vibration prevention and control measures and the prediction of blasting seismic effect is an important topic for many scholars to study and analyze for a long time.Based on the blasting excavation of a tunnel in Qingdao,this paper aims at predicting the seismic effect of a four-story frame structure near the ground.Based on the acquisition of blasting monitoring data that systematic research and analysis is carried out by using traditional empirical formulas,genetic neural networks,numerical simulations and on-site measurements.The main research contents of this paper are as follows:(1)Through the real-time monitoring of the blasting vibration during the construction of the new tunnel blasting,on the basis of obtaining a large amount of measured data,the empirical formula is linearly fitted by the least squares method to determine the specific values of K and a in the empirical formula and the relationship between blasting seismic effect and proportional dose(Q1/3/R).Finally,using the obtained empirical formula to perform regression prediction on the predicted samples,The average prediction error between the predicted result and the measured result is 23.73%,and the prediction accuracy of the vertical blasting seismic effect is significantly higher than the horizontal direction.(2)In order to achieve the highly accurate prediction effect of blasting seismic effect,in this paper,a neural network improved by genetic algorithm is established to predict and analyze the blasting seismic effect.Comparing the prediction results with the empirical formula prediction results,the prediction of genetic neural network is better than the empirical formula.The scam prediction error of genetic neural network on blasting seismic effect is 9.248%,within the practically acceptable range of the project.Therefore,it can be used as an effective method for estimating the blasting seismic effect of the frame structure.(3)In order to further highlight the reliability and practicability of the genetic neural network prediction model,the ANSYS/LS-DYNA is used to simulate the peak velocity of the frame structure under different blasting parameters.The relative error between the simulation result and the measured result is within 22%,and the RMSE is in the range of 0.09-0.15.Finally,the simulation results are compared with the results of genetic neural network prediction,and the prediction accuracy of genetic neural network is high.It is fully verified that the genetic neural network prediction model has excellent nonlinear processing ability and can fully reflect the nonlinear characteristics of blasting seismic waves.The prediction accuracy of genetic neural network can fully meet the actual requirements of engineering,and it can provide a powerful method for estimating the seismic effect of frame structure blasting,and provide a theoretical basis for controlling the damage of blasting vibration.It has practical promotion and application value.