Applying reinforcement learning to mitigate wake-induced lift fluctuation of a wall-confined circular cylinder in tandem configuration

The flow around two tandem circular cylinders leads to significant lift fluctuation in the downstream cylinder owing periodic vortex shedding. To address such research issues, we present herein a numerical study that uses deep reinforcement learning perform active control (AFC) on with low Reynolds number of 100, where actuator causes rotation cylinder. First, center spacing ratio L* varies from 1.5 9.0, and variation quasi-steady reattachment regime (L*≤3.5) co-shedding (L*≥4.0). fluctuating is maximum when L*=4.5. Next, train an optimal AFC strategy suppresses 75% This approach differs using direct-opposition change vortex-shedding frequency or strength, as reported previous studies. modifies phase difference between fluctuations by delaying merging upstream wake accelerating formation recirculating bubbles after merging. With new difference, effect additional significantly mitigated. results dynamic mode decomposition show vortices surrounding 1 contribute are weakened. best our knowledge, this investigation can provide ideas physical insights into problem under disturbed incoming flow.