Nonlinear black-box models and force-sensorless damping control for damping systems using magneto-rheological fluid dampers

In vibration control field, magneto-rheological (MR) fluid dampers are semi-active control devices that have recently begun to receive more attention. This paper presents a nonlinear black-box model (BBM) and an inverse black-box model (IBBM) for the identification of a MR fluid damper and their application to design a novel force-sensorless control method for any damping system using that damper. The nonlinear model named ‘black-box’ is a simple direct modeling method which was designed based on fuzzy-neural technique. Characteristics of the damper in study are directly estimated through a fuzzy mapping system. In order to improve the model accuracy, neural network technique including back-propagation and gradient descent method were used to train the fuzzy parameters to minimize the modeling error function. The inverse model, IBBM with self-learning ability, was then derived based on the BBM with optimized parameters and neural network technique. Consequently, the designed BBM and IBBM models can be used as a ‘virtual’ force sensor and an adaptive force controller, respectively, to perform a closed-loop feedback force-sensorless control for any damping system which uses the corresponding MR fluid damper. Effectiveness of the proposed models for modeling as well as force-sensorless damping control technique has been investigated through a series of simulations and real-time experiments on two vibrating systems employing the same MR fluid damper. The simulation and experimental results show that the suggested BBM could describe well the MR fluid damper behavior and could be combined with the IBBM for the force-sensorless damping control system