Vibration Control in Meta-Structures Using Reinforcement Learning

Document Type

Conference Proceeding

Publication Date

1-1-2022

Department

Department of Mechanical Engineering-Engineering Mechanics

Abstract

This chapter considers using reinforcement learning (RL) to adaptively tune frequency response functions of meta-structures. RL algorithm tunes the stiffness of the spring of the lumped multi-DOF system, as the lumped mass is varied. As some of the lumped masses are modified by 10%, the spring’s stiffness is tuned to maintain the original bandgap. A Q-Learning algorithm is used for RL, wherein the Q-value is updated based on Bellman’s equation. The results compare the frequency response functions of the terminal masses of the baseline and varied mass structure.

Publication Title

Conference Proceedings of the Society for Experimental Mechanics Series

ISBN

9783030759131

Share

COinS