Real-Time Computing for a Parameterized Feedback Control Problem of Boussinesq Equations by POD and Deep Learning

Abstract

An efficient real-time computational method for a feedback control problem of the Boussinesq equations is studied. We consider a simple and effective feedback control law based on the relationship between the control and adjoint variables in the optimality system. We investigate a closure type modeling in reduced order model (ROM) of this problem for real-time computing. In order to improve the existing well-known POD-ROM method, the deep learning technique, which is currently being actively researched, is studied and applied. Computational results presented show that the suggested methods work well.