Application Of Adaptive Inverse Control In The Elecrohydraulic Shaking Tables

Multi-axis electrohydraulic shaking table(EST)is a basic strcutral testing device.It can duplicate the desired waveform to simulate the real environment of mechanics.Thus the fatigue life and reliabilty of the spicements can be tested.It has been widely used in civil,transportation,military,aerospace and precision instruments industries.How to make the multi-axis EST produce highly correlatively time-domain acceleration waveform to the desired signal is the prensent plume in Engineering fields.Whereas adaptive invese control(AIC)has been ackonwleged as most suitble technology to solve the refered issue as an adanced control technology.This reserch has been sponsored by the National Natural Science Foundation of China as parts of the "Inner Force Coupling and Its Rejection of Seismic Simulator with Hyper-Redundant Parallel Actuation and Long Stroke".Aimimg at improving the accuracy of time-domain acceleration waveform duplication of multi-DOF on multi-axis EST via the AIC,this research has been conducted.This research has studied the servo control system which is mainly based on the three variable controller(TVC),in addition with the DOF synthesis and decomposition matrix and the inner-pressure stabilization controller.Considering the nonlinearity of hydraulic actuators mainly the flow,a model of multi-axis EST has been established to represent the actual dynamic response.Based on this model,a rule on adjusting parameters of the TVC has been given.Simulation results have proved the effectiveness of the servo control system.Meanwhile the essential drawbacks have been summarized.All above results make a solid foundation for the further reserch.Introduced as the vibration controller,the AIC has been studied in its forms and the adaptive filtering algorithm which takes the core parts.Where a realizably variable tap-length NLMS adaptive filtering algorithms has been introduced to find the optium tap-length of the system.To improve the perfoemance of the off-line AIC,an adaptive filtering algorithms based on the BP neural network has been applied.The classic on-line AIC denoted to filtered-X algorithms is also been studied.The simulation results of referrenced algorithms have shown the AIC can coordinated phase frequency characteristic and compensated amplitude frequency characteristic of the system.Parts of drawbacks have been solved through the aforementioned results.The complex procedure of adjusting the TVC still hurts the practicability when applying.Thus a servo controller based on off-line adaptive inverse control is proposed to simplify the procedure by representing the fedforward part of TVC.Finally,a combined adaptive inverse control strategy is proposed to sol ve the problem of high accuracy and multi-DOF of acceleration time-domain waveform duplication.The simulation results has strongly support ed.the superiorities of proposed algorithms.All the algorithms referred in this research have been experimented in a DOF of a multi-axis EST.The results has shown that the proposed control strategy is feasible and can meet the requirements of accuracy on time-domain waveform duplication.