Adaptive estimation approaches for on-line estimation and identification of vibrating hysteretic systems under arbitrary dynamic environments are crucial for the on-line control and monitoring of time-varying structural systems. The available adaptive estimation/identification techniques suffer from two drawbacks: they assume that (1) the internal restoring forces applied to the system's elements are available for measurement and that (2) the nonlinear differential equation driving these restoring forces can be parametrized as a linear combination of unknown constant parameters and known nonlinear terms. In this paper, a new approach is presented which completely eliminates the above two restrictions. Specifically, the proposed method solves the problem of estimating/identifying the restoring forces without assuming that the restoring forces are available for measurement, and without invoking any assumptions concerning the nature of the restoring forces dynamics. The new approach uses appropriate adaptive filtering and estimation techniques and also makes use of the Volterra/Wienner Neural Network (VWNN) which is capable of learning input/output nonlinear dynamical behaviors. Simulations performed on a representative structural system, as well as physical tests on a steel frame and on a reinforced concrete assembly undergoing severe hysteretic behavior, demonstrate the utility and verify the efficiency of the proposed technique.
展开▼
