Closed-loop control based system modification for parameter estimation

Abstract

We deal with the multiple modal analysis that perturb the original system variously and use the modal data set of perturbed systems in applications such as parameter estimation or damage detection. Though many approaches about multiple modal analysis have been proposed, no systematic experiment guideline for modification is proposed. For this reason, the first research objective is the systematic development of system perturbation method. Among various approaches like point mass addition, boundary condition change or feed-back control, closed-loop control based perturbation is selected as the basic perturbation tool for its easiness and accuracy in experimental setup and ability of diverse perturbations. Plenty of sensors are assumed available and distributed over the system and a few actuators such as electro-magnetic exciter are assumed available as usually done in the conventional modal test. And static output feedback control scheme is chosen for simplicity. For the selected configuration of transducer and controller type, decision on the actuator``s and sensor``s locations and feedback gain is considered for systematic system perturbation. A criterion for the decision of actuator location based on control power requirement is proposed. For the chosen actuator locations, the feedback gain is obtained minimizing a performance index of a weighted modal power of each mode with the fulfillment of pre-defined modal change of the target mode. As sensors are distributed over the system and exciters are located in a selected few locations, this configuration results in non-collocated sensor and actuator case where diverse changes of system are possible but the instability problem of closed-loop system prevents its actual usage. The stability of closed-loop system can be guaranteed by adjusting the weighting to each mode on-line during the modal testing. However, the modification of feedback gain vector during the experiment is a little cumbersome. Therefore, it is ...