The application discloses a method for suppressing
transonic buffeting load and response based on agent cooperation, and relates to the technical field of
aeroelasticity, and aims at solving the problem of
transonic buffeting of a wing of an aircraft. The method comprises the following steps: firstly, calculating the elastic wing tip acceleration response based on a one-way fluid-
solid coupling method; secondly, configuring a multi-target
reinforcement learning agent, taking the variable-camber
trailing edge and the telescopic micro-bulge as control mechanisms, taking the maintenance of the
aerodynamic lift-drag ratio and the suppression of the
vibration response of the structure as targets, and letting the agent adopt a proximal policy optimization
algorithm to be trained to obtain an
optimal control law capable of cooperatively driving the telescopic micro-bulge and the variable-camber
trailing edge; and thirdly, deploying the trained cooperative control law on the target wing, and observing the suppression effect of the wing buffeting through quantitative evaluation of the slowing rate. The application trains the agent through a deep
reinforcement learning method, so that the agent can guide the telescopic micro-bulge to suppress the
diffusion in the separation zone, reduce the shock motion amplitude, drive the variable-camber
trailing edge to regulate the flow at the trailing edge and compensate the overall lift fluctuation, and cooperatively suppress the buffeting load and response.