In this paper, dynamic characteristics of electrorheological dampers are identified using an effective procedure including dynamic tests, modeling, and combination of least squares and radial basis function neural networks techniques. A two step computationally efficient and experimentally affordable approach is chosen to achieve this objective. At first step, the responses of the investigated damper to harmonic excitations for limited number of applied command voltages to the damper are determined and the damper parameters are identified based on Bingham viscoplastic model and using least squares technique. In the identification procedure using least squares technique, the Euclidean distance between the experimental value of force -velocity data and the one predicted force-velocity data by the model is considered as an objective function. Since the number of experiments is limited and conducting tests for all possible values of command voltage is a formidable task, therefore, in the second step, radial basis function neural networks (RBFNN) technique is implemented to identify the parameters of the model for any arbitrary value of the command voltage in the experimentally investigated range. The RBFNN is trained with the data obtained from the identification procedure in the previous stage. In the present work, command voltages are considered as training sets inputs and their corresponding parameters for Bingham model are considered as targets for training sets. In this study, all the inputs are chosen as centers, i.e., the number of neurons at hidden layer is equal to the training sets.&nbsp;Comparison of predicted responses of the investigated damper with experimental counterparts in the first step and examining of the generalization capability of the radial basis function neural networks in the second step verified the ability of the whole procedure in successful identification of the model parameters for any arbitrary applied command voltages.&nbsp;Furthermore, with using the identified parameters for the investigated ER damper, a preliminary estimate of the damper behaviour under drop mass impacting condition is presented. It is roughly assumed that the damper has linear behaviour during high-speed motion due to impact, in other words the value for flow index number, n, in Herschel–Bulkley constitutive equation is considered equal 1. In modeling the impact dynamics, the impact is divided into 3 phases. In the phase 1, the dropping mass with the polyurethane impactor (to prevent ringing due to metal-to-metal contact) attached to it () after releasing from height h touches the load cell attached to the top of the damper rod and moves down with decreasing speed while the polyurethane impactor is compressed and still the spring force in this elastic impactor is less than the combination of the accumulator and yield force of the ER damper. When the elastic force of the impactor exceeds the sum of the accumulator and the ER fluid yield forces, the phase 2 of the impact begins where, the drop mass continues decelerating and reducing speed while the moving part of the accelerates and its speed increases from zero until reaches the speed of the drop mass .In the phase 3 of the impact, the dropping mass and the ER damper piston decelerate with same speed until either their speed decrease to zero or touch the damper protection barrier against piston bottom out. This preliminary estimation of the damper’s impact behaviour can be helpful in quickly finding appropriate test conditions such as height and mass of the dropping mass as well as command voltages for identifying the actual flow index number using the drop tower test in the next research.

Identification of Dynamic Characteristics of Electrorheological Dampers Using a Combination of Least Squares and Radial Basis Function Neural Networks Techniques

Paper Proceedings: https://asmedigitalcollection.asme.org/IMECE/IMECE2021/volume/85543

ASME 2021 International Mechanical Engineering Congress and Exposition (IMECE2021)

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.