A SDNN-MPC method for power distribution of COGAG propulsion system. (1st September 2022)

• Record Type: Journal Article
• Authors: Li, Jian; Wang, Zhitao; Li, Shuying; Ming, Liang
• Is Part Of: Energy. Volume 254:Part B(2022)
• Publication Date: 2022-09-01
• Subjects: COGAG propulsion System -- Model predictive control -- Simplified dual neural network -- Power distribution; Power resources -- Periodicals; Power (Mechanics) -- Periodicals; Energy consumption -- Periodicals; 333.7905
• DOI: 10.1016/j.energy.2022.124310
• Languages: English
• ISSNs: 0360-5442

Abstract

Abstract: In order to solve the problems of power distribution taking a long time and large fluctuation of propeller speed in the process of COGAG propulsion system switching operating patterns, model predictive control (MPC) is introduced to control power distribution and a SDNN-MPC method is proposed. Discrete incremental LPV model is established as the prediction model and MPC for power distribution process is designed. In order to make MPC suitable for COGAG, which is a nonlinear time-varying system with short sampling period, a simplified dual neural network (SDNN) with simplified constraints is used to solve the quadratic programming in MPC. A large number of simulation results show that the established LPV model has high prediction accuracy and SDNN is helpful to shorten the execution time of MPC. The softening factor and weight factors in MPC can affect the performance of power distribution control. Two principles are summarized to choose weight factors and the recommended range of the softening factor is 0.026–0.034. Compared with the PI controller, based on the SDNN-MPC the time of power distribution process is shortened and the fluctuation of propeller speed is reduced. The maneuverability and stability of the vessel are improved. Highlights: A SDNN-MPC algorithm based on LPV model is proposed for the power distribution process of COGAG propulsion system. The designed control method can significantly improve the maneuverability and stability of COGAG propulsion system. The robustness and good computing efficiency of the SDNN-MPC are verified.