Nonlinear stochastic control of self-powered variable-damping vibration control systems

Traditional mechanical damping systems employ dissipative components, such as viscous dampers, to absorb energy from an externally excited vibratory structure, and dissipate this energy as heat. By contrast, self-powered damping systems employ transducers to convert absorbed mechanical energy into electrical energy, which is then managed in an electrical network. This enables the network to intelligently control and adapt the damping properties of the devices in real-time, and to use the extracted energy to power the control intelligence and power electronic subsystems. In order for the system to have energy-autonomy, the transducers must extract sufficient energy to overcome the parasitic losses in the system, and to maintain the static power requirements of the control intelligence. In this paper we present a nonlinear control design technique for this technology, which is based on multi-objective LQG control. The technique has the objective of strongly reducing one performance measure in stochastic response, subject to constraints on other performance measures, as well as on the power generated by the system. The technique is described generically, and is demonstrated on an application with relevance to earthquake engineering.