Intelligent Approach to Floor Vibration Control
Genetic Algorithms (GAs) have been used to solve a multiple of engineering problems with the civil engineering applications ranging from optimal placement of sensors and actuators on structures to pipeline layouts. GAs are especially useful in finding optimal solutions to problems that have many parameters with complex search spaces and a high level of interaction among the describing parameters. The novel experimental control approach presented in this paper uses a GA and a piezoelectric actuator to control the vibration of an aluminum cantilever beam. This set-up is based on a floor vibration problem, where the human perception of vibration dictates the sensitivities in the cost function of the GA. Lightweight floors can be excited by occupant activities such as walking, jumping and dancing. Humans are especially sensitive to vibrations in the range of 4 to 8 Hz. The occupancy of a floor system — whether the floor is used in an office, a shopping mall, or a ballroom — determines the degree to which humans are annoyed by the vibration. In this paper, the GA based control limits the peak acceleration within a predefined bandwidth. Since the cantilever beam has a higher natural frequency than a lightweight floor system a bandwidth of approximately 1.6–15.9 Hz (10–100 rad/s) is used as the frequency range to control. The control to be designed will be a genetic algorithm-robust controller. The analytical results indicate that this novel approach works well.