Research On Parameter Identification Of Mechanical Model Of Magnetorheological Damper

Magnetorheological damper is one of the semi-active control devices with the most potential for development in structural vibration control.This low-energy-consuming and fast-response device has caused huge impact in the field of civil engineering vibration control due to its superior performance.However,this device is highly nonlinear due to its own magnetorheological effect in the application process,which also makes its application in semi-active control extremely difficult.Therefore,establishing a simple and effective MR damper mechanical model is an important condition for maintaining stability and effectiveness during the control process.In the mechanical model,the determination of unknown parameters is crucial.Therefore,this paper mainly focuses on the parameter identification of the magnetorheological damper mechanical model as follows:(1)The methods used mainly in the process of parameter identification and the common mechanical models of magnetorheological dampers are reviewed.In view of the mechanical model,the advantages and disadvantages of the common models are analyzed.Compared with other mechanical models,the Bouc-Wen model can better describe the characteristic curves between damping force and displacement,or velocity of MR damper,and the number of unknown parameters and the number of differential equations in the model equation are relatively few,which is convenient for programming.Therefore,this study selects the Bouc-Wen model as the main object of parameter identification,while the Bingham model and the modified Bouc-Wen model are the auxiliary identification objects.(2)The ant colony optimization algorithm is applied to the parameter identification of the mechanical model of magnetorheological damper.First,the characteristics and steps of the ant colony algorithm based on grid partition are introduced.At the same time,the influence of the selection of parameters in the ant colony optimization algorithm on the recognition accuracy is analyzed.The test results show that the number of ants affects the size of the solution vector space;pheromone evaporation coefficient,total amount,equal interval parameters and the range of parameters affect the convergence speed and solution quality;the combination of the differentparameters on the final form of recognition accuracy is greatly affected.Therefore,in order to identify the accuracy of the model,the values of the parameters in each model is analyzed.This method is used to identify the Bouc-Wen model and the Bingham model of the magnetorheological damper under different working conditions.The simulation test shows that the recognition accuracy of the Bingham model is the highest,the Bouc-Wen model is the next.And from the feature curve,the ant colony optimization algorithm can solve the problem of parameter identification of mechanical model of the magnetorheological damper to a certain extent.(3)The Unscented Kalman filter algorithm is applied to the parameter identification of mechanical model of the magnetorheological damper.First,the basic contents and steps of the Unscented Kalman filter algorithm are summarized,and the application of the algorithm in the parameter identification is explained.The Unscented Kalman filter algorithm does not need to approximate the nonlinear state equation compared to the Extended Kalman filter algorithm.It uses the unscented transform method to obtain sampling points,avoiding the tedious procedures for deriving the Jacobian matrix.And the recognition accuracy is also high.This paper mainly uses this algorithm to identify the Bouc-Wen model,Bingham model and modified Bouc-Wen model of magnetorheological damper.Simulation test results show that the Unscented Kalman filter algorithm has high recognition accuracy of parameters and mechanical characteristic curves in these three models;the required optimization time will be different with the complexity of the model,but compared with the ant colony optimization algorithm,it takes less time.(4)In order to illustrate the advantages and disadvantages of the above two algorithms,this paper compares the recognition accuracy of the grid-based ant colony algorithm and the unscented Kalman filter algorithm from the global and local aspects respectively.In the comparison process,the Unscented Kalman filter algorithm is affected by the initial value,but as long as a good initial value is determined,the Unscented Kalman filter algorithm has better recognition effect;and the ant colony optimization algorithm can identify a wide range,the accuracy of the identified parameters is not high,but can provide a certain reference.Therefore,the true value of the parameters in the actual process is unknown.A method of identifying the unknown parameters in the mechanical model of the magnetorheological damper is proposed by combining the ant colony optimization algorithm and the Unscented Kalman filter algorithm.First,the ant colony optimization algorithm is used to provide reliable initial value conditions for the Unscented Kalman filter algorithm.Then,the Unscented Kalman filter algorithm is used to identify the parameters of model,and the final identified parameters are of high accuracy.