Observer-based controller for floor vibration control with optimization algorithms

Abstract

This study presents the results of vibration suppression of a walkway bridge structure with a single actuator and sensor pair by using a proportional-integral (PI) controller and observer-based pole-placement controllers. From the results of experimental modal analysis, reduced-order models of the walkway are identified. These are used for the design of a PI controller as well as for state estimation procedures that are necessary for the development of reduced-order observer controllers. The respective orders of the latter are dependent on the number of plant modes used for their designs. They are formulated from plant and observer feedback gains that are obtained from the specification of desired floor closed-loop eigenvalues and observer eigenvalues. There are numerous solutions possible with the observer-based controller design procedures whereas the PI controller defaults to a particular solution. There is also the flexibility for isolation and control of target vibration modes with the observer-based controllers for higher controller orders from a purely single-input single-output controller scheme as demonstrated in the analytical and experimental studies presented. Further, in this work, a design space of potential feedback gains is specified, where only a single plant mode has been used for the observer-based controller design process, and a multi-objective genetic algorithm optimization scheme is used to search for an optimal solution within some pre-defined constraint conditions. The best solution here is regarded as one that offers the greatest vibration mitigation performance amongst the solutions identified.