Research On Multi-Parameter Optimization Control Of Shaking Table Based On BP Neural Network

Seismic simulation shaking table experiment is the most direct method to study seismic response and failure mechanism of structures in laboratory.Based on the basic performance parameters of the unidirectional seismic simulation shaking table,the frequency response characteristics of the shaking table under the three-parameter control and the multi-parameter control are analyzed and simulated respectively.In order to explore the influence of multi-parameter control parameters on the performance of shaking table system,the influence of each control parameter on the frequency response characteristic curve of shaking table system is analyzed under multi-parameter control.In order to improve the waveform reproduction accuracy of shaking table system,BP neural network algorithm is introduced to identify the multi-parameter control parameters and analyze the system simulation.The main research contents of this paper are as follows:(1)In this paper,based on the basic performance parameters of the seismic simulation shaking table in our university,the frequency response characteristics and modeling simulation analysis of the shaking table under three parameter control and multi-parameter control are respectively carried out.The simulation results show that,compared with the three-parameter control,the oil column formant can be weakened under the multi-parameter control.Taking El-Centro(NS)wave as input signal,the waveform correlation coefficient of shaking table system is increased by 6.8% and the maximum peak error is reduced by 17% after acceleration control is introduced.The introduction of acceleration feedforward and feedback can effectively weaken the formant of the oil column,improve the accuracy of waveform recurrence,and improve the overall stability of the system.(2)Under the multi-parameter control,the influence of each control parameter on the frequency response characteristic curve of the shaking table system is analyzed one by one by using the control variable method,and the multi-parameter control parameters are adjusted manually under the adjustment rules of the multi-parameter control parameters,and the system frequency response characteristic analysis and simulation analysis are carried out.The simulation results show that compared with the theoretical control parameters,the effective use band of the system under the control of manually adjusted parameters is expanded by 18 Hz,the oil column formant caused by the characteristics of the servo valve is effectively eliminated,and the waveform reproduction accuracy of the shaking table system is improved.(3)In order to improve the waveform reproduction accuracy of shaking table system,BP neural network algorithm is introduced to identify the multi-parameter control parameters of shaking table and analyze the system simulation.The simulation results show that,compared with the theoretical control parameters,the effective use frequency band of the shaking table system is expanded by 25 Hz and the formant is effectively eliminated after the introduction of BP neural network algorithm.Four kinds of seismic signals are used as input signals and simulation analysis of shaking table system is carried out.The simulation results show that the waveform correlation coefficients of the shaking table system under the BP neural network algorithm are all above 0.98,and the maximum peak errors are all below 0.0298.Compared with the multi-parameter theoretical control parameters,the waveform correlation coefficients of the shaking table system are increased by about 35% after the introduction of BP neural network algorithm,and the maximum peak errors are reduced by about 88%,which verifies the effectiveness of the BP neural network algorithm in this paper.