Development of a constraint non-causal wave energy control algorithm based on artificial intelligence. (March 2021)

• Record Type: Journal Article
• Authors: Li, L.; Gao, Y.; Ning, D.Z.; Yuan, Z.M.
• Is Part Of: Renewable & sustainable energy reviews. Volume 138(2021)
• Publication Date: 2021-03
• Subjects: Wave energy control -- Non-causality -- Wave force prediction -- Artificial intelligence -- Power extraction -- Control constraint; Renewable energy sources -- Periodicals; Power resources -- Periodicals; Énergies renouvelables -- Périodiques; Ressources énergétiques -- Périodiques; 333.794
• DOI: 10.1016/j.rser.2020.110519
• Languages: English
• ISSNs: 1364-0321

Abstract

Abstract: The real-time implementation of wave energy control leads to non-causality as the wave load that comes in the next few seconds is used to optimize the control command. The present work tackles non-causality through online forecasting of future wave force using artificial intelligence technique. The past free surface elevation is used to forecast the incoming wave load. A feedforward artificial neural network is developed for the forecasting, which learns to establish the intrinsic link between past free surface elevation and future wave force through machine learning algorithm. With the implementation of the developed online wave force prediction algorithm, a real-time discrete control algorithm taking constraint on response amplitude into account is developed and implemented to a bi-oscillator wave energy converter in the present research. The dynamic response and the wave power extraction are simulated using a state-space hydrodynamic model. It is shown that the developed real-time control algorithm enhances the power capture substantially whereas the motion of the system is hardly increased. The prediction error effect on power extraction is investigated. The reduction of power extraction is mainly caused by phase error, whilst the amplitude error has minimal influence. A link between the power capture efficiency and the constraint on control is also identified. Highlights: A real-time declutching wave energy control based on model predictive control strategy is developed. Artificial intelligence is used to tackle the non-causality. Forecasting error effect on control efficiency is investigated. Constraint effect on control efficiency is investigated.