Improved Adaptive Delay Compensation Methods For Real-time Hybrid Testing

Developed from pseudo dynamic testing,the Real-time Hybrid Testing(RHT)fully utilizes the high efficiency of the numerical simulation and the real and reliable advantages of the physical test and as a method for testing structural systems,closely coupling numerical simulation and experimental testing to obtain the complete response of a structure is especially suitable for the study on control devices with velocitydependent property,such as MR dampers.In RHT,the key to the success of the experiment is to ensure the interface boundary compatible of the two parts which refer to numerical substructure(NS)and experimental substructure(ES).However,this requires the actuator to track the desired commands calculated by the NS accurately.Thus,time delay which lead to inaccuracies and potential instabilities in RHT and its compensation are the key problems for RHT.In RHT with MR dampers as ES,the strong nonlinearity of the MR damper causes a time-varying delay of the loading system,for which conventional compensation methods mostly based on constant delay assumption are no longer suitable.To resolve nonlinear control and time-varying delay compensation in RHT for MR dampers as specimens,an adaptive time delay compensation method should be used cautiously,making the research of great significance to the hybrid simulation and seismic performance evaluation.Studies on a novel adaptive time delay compensation method has been done to resolve the time-varying delay in RHT.The main work of this study is summarized as follows: Firstly,the adaptive time delay compensation method,proposed by Li Qiang,has some impediments in RHT of a small control step size.In view of this problem,a modified adaptive feedforward delay compensation method in which a more suitable compensation controller consisting of the desired displacement rather than measured displacement is used,is introduced for the sake of convenience in RHT.And a finite windows LS method with less disturbance of the controller parameters and better robustness is used for parameter estimation.Numerical simulation and real tests for RHT with MR dampers are then performed to demonstrate the effectiveness of the adaptive compensation method.Secondly,additional PI feedback and MCS feedback control are introduced to improve the performance of the adaptive feedforward time delay compensation method.Subsequently,the Simulink model of servo-hydraulic actuators,developed by Zhao,and the Simulink model of MR dampers,developed by Weber,are employed in the simulation of the RHT.All simulations demonstrate that in RHT with improved methods small values of mean square root,maximum displacement error and standard error are obtained,which means better performance than the time delay compensation method based on polynomial extrapolation and the adaptive feedforward method.Meanwhile,the adaptive method with PI control is suitable for different loading cases and show better robustness than the method with MCS control.Thirdly,an empirical tuning method for the feedback parameters of the outer loop is given.The efficacy of the proposed approach,including feedforward-feedback which links to achieve accurate tracking of the desired displacement,is demonstrated through RTH of a three-story steel building employing a magnetorheological(MR)damper as the rate-dependent experimental specimen.According to the test experience,the parameters of the PI control is easier tuning than the parameters of the MCS method.Meanwhile the adaptive method with PI control showed the best accuracy in RHT.