Development of a hybrid deep learning model with HHO algorithm for dynamic response prediction of wind-wave integrated floating energy systems

• Published in: Ocean engineering Vol. 340; p. 122394
• Main Authors: Yin, Jiaqing, Fan, Yihong, Bashir, Musa, Nie, Debang, Lai, Yongqing, Ding, Jieyi, Yu, Jie, Li, Chun, Yang, Yang
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 30.11.2025
• Subjects: Floating offshore wind turbine; Harris Hawk optimization; Hybrid energy system; Hybrid neural network; Wave energy converter; Harris Hawk optimization; Floating offshore wind turbine; Wave energy converter; Hybrid neural network; Hybrid energy system
• ISSN: 0029-8018
• DOI: 10.1016/j.oceaneng.2025.122394

Wind-wave integrated floating energy system (IFES) is recognized for reducing the levelized cost of electricity by sharing the support structures. Developing efficient models to accurately predict dynamic responses is crucial for the performance evaluation and optimization design of wind-wave hybrid energy concepts. This study has developed a novel hybrid deep learning model, CLCBA_HHO by incorporating signal processing and Harris Hawk Optimization(HHO) algorithms into the convolutional neural network(CNN) and bi-directional long-short term memory with attention mechanism(Bi-LSTM-AM). The wind-wave IFES concepts, comprising a 10 MW wind turbine and three wave energy converters (WECs) supported by the OO-Star platform, is selected for the case study. The dynamic responses of the wind-wave IFES concept under environmental loadings are calculated using a well-validated numerical tool and used the dataset for training and testing the CLCBA_HHO model. The performance of CLCBA_HHO in predicting wind and wave power, tower loads, and platform motions is demonstrated through a comprehensive comparison with several existing deep learning methods. The accuracy of CLCBA_HHO in predicting wind and wave power can reach over 95 % under various working conditions, significantly outperforming the state-of-the-art similar method. This study verifies the superiority of the proposed method in predicting dynamic responses of wind-wave IFES concepts, offering a reliable and effective performance evaluator for the optimization design of wind-wave hybrid energy systems. •A hybrid deep learning model is proposed for predicting dynamic response of wind-wave integrated floating energy systems.•Reliable simulation data is used for training and testing the prediction model.•HHO algorithm is used for optimizing the hybrid neural network framework.•The predictive performance of the proposed model is better than existing algorithms.