BP neural network vibration control with time delay for offshore platforms under wave forces

Abstract

This paper designs a vibration controller based on back propagation (BP) neural network for jacket-type offshore platforms with time delay. With an exosystem designed, we can describe the irregular wave forces in the context of Morison equation and wave theory. The original vibration control can be seen as an optimal control problem. This linear discrete-time system is affected by wave forces as external disturbances. A BP neural network controller is designed based on a feedforward and feedback optimal control law with memory (FFOCLM). Neural network has been widely applied in controlling with strong capability of approximation and adaption. The tuning rule for updating weights and thresholds of the network is derived to approximate the dynamics of the control system. Data of the FFOCLM are sampled to train the BP neural network. The simulation results prove the feasibility and efficiency of the BP neural network controller designed in paper for attenuating vibrations of offshore platforms.