Abstract
This research illustrates semi-active structural control of a three-story nonlinear building using magnetorheological dampers and a neuro-fuzzy algorithm. The structure being studied was developed for a third generation benchmark problem that is being investigated on a worldwide basis. A magnetorheological (MR) damper is tested experimentally and employed as a control device. An MRD-300kN-240 MR damper, which is manufactured by the Sanwa Tekki Corporation, is tested in a controlled laboratory environment to obtain a set of experimental data. A model of the damper is trained and validated via an adaptive neuro-fuzzy (ANFIS) algorithm. Next, an array of trained and validated MR damper models is inserted between the ground and the first floor of the benchmark building and numerically simulated under seismic excitation. To control response of the benchmark structure, a neuro-fuzzy control (NEFCON) strategy is applied with acceleration feedback. Training of a NEFCON controller is accomplished through use of an artificial earthquake. Next, performance of a fuzzy controller is validated by investigating time histories of the absolute acceleration response and the overall performance indices when the structure is subjected to a set of actual earthquakes. Numerical results indicate that the proposed semi-active control strategy satisfactorily mitigates absolute accelerations in comparison with response of the uncontrolled structure. In addition, performance indices indicate reduction in response of the structure in comparison with some passive-off and passive-on states of the control devices.
Likhitruangsilp, Visit (2002). Neuro-fuzzy control of a multi-degree of freedom structure with semi-active magnetorheological dampers. Master's thesis, Texas A&M University. Available electronically from
https : / /hdl .handle .net /1969 .1 /ETD -TAMU -2002 -THESIS -L55.