New functionalities, higher comfort and increasing performance
requirements are often be solved by adding new technologies to
existing (passive) solutions. Monitoring and control approaches
uses additional sensors and actuators, new materials, microprocessors and new devices realizing new and improved functionalities. Two effects are becoming more and more interesting:
(1) the lifetime of new actuators/materials strongly depends on the usage-history, (2) the functionality of the new composed systems depends on the fully functionality of all elements. In the consequence, the availability of such new systems is decreased by the number of elements and depends strongly on the use. These effects are known and act against new developments improving performance behavior also in mechanical engineering, automotive systems etc. This will be also the case for multifunctional composite or compound systems such as piezomaterials, magnetostrictive alloys or smart memory alloys (SMA) and is actually within the focus of the Structural-Health-Monitoring (SHM)-community. This contribution explains a new and systematically structured methodological approach to avoid and eliminate failures in mechatronical systems in an integrated and intelligent way to achieve a desirable or required amount of utilization in compliance with a defined failure rate. The result is an enhancement of the dependability of such a system.
In this paper, a model-based approach for fault detection and vibration control of flexible structures is proposed and applied to 3D-structures. Faults like cracks or impacts acting on a flexible structure are considered as unknown inputs acting on the structure. The Proportional-Integral-Observer (PI-Observer) is used to estimate the system states as well as unknown inputs acting on a system. Also the effects of structural changes are understood as external effects (related to the unchanged structure) and are considered as fictitious external forces or moments. The paper deals with the design of the PI-Observer for practical applications when measurement noise and model uncertainties are present and shows its performance in experimental results. As examples, impacts acting upon a one side clamped elastic beam and on a thin plate structure are estimated using displacement or strain measurements. To control the vibration of the flexible plate, two piezoelectric patches bonded
on the structure are used as actuators. The control algorithm introduced in this contribution contains a state feedback control
and additionally a disturbance rejection. The disturbances are
estimated using the PI-Observer. Experimental results show the
performance and the robustness properties of the control strategy
for the vibration control of a very thin plate.