The active simulation and modification of structural frequency response
Environmental or structural testing work often involves measurements on a test specimen which is attached to a larger structure of finite mechanical input impedance. It would be useful to select from a range of load impedances associated with the larger structure without altering it mechanically. This thesis describes the application of active control techniques to such a problem, and presents theoretical and experimental work which attempts to determine the range of load impedances which can be simulated in practice. Two approaches are presented. The first considers modification of the frequency response of a simple test structure using feedback control. Experimental work using a digital control system demonstrates modification of the modal parameters of the fundamental mode of vibration of a cantilever. The relative merits of state estimation/feedback and a more pragmatic proportional plus integral plus derivative (PID) controller are discussed, and the PID-type controller is used for experimental work which demonstrates independent control of the first two modes of a cantilever. The second approach considers simulation of the desired impedance function by electronic means. The input impedance of an electrodynamic shaker can be modified by using a filter which operates on a signal proportional to the acceleration produced by the shaker and which is fed back to the shaker input. A `black-box' model of a simple shaker allows the design of a feedback controller which when implemented leads to the desired closed-loop response. Experimental results show the simulation of fourth and sixth order systems using a second order test system, for which an eighth order controller is required. Extension of the analysis to a two-channel system allows the electronic simulation of a mechanically coupled system using shakers which are not mechanically coupled. Although experimental results are presented, the complexity of the controller matrix is seen as a limiting factor for this approach.