SCIENCE@home

Combination of Self‐Powered Hybrid Generator and Hierarchical Hybrid Learning Architecture for Broadband On‐Site Ocean Monitoring

Chenxi Li,Tianyu Li,Kuankuan Wang,Yixin Liu,Fupu Xie,Yongle Zhou,Hongwei Wang,Yongming Yao,Huiying Liu

Published 2025 in Advanced Functional Materials

ABSTRACT

For marine wave monitoring, a dual‐mode energy driven triboelectric‐electromagnetic hybrid generator (DM‐TEHG) is introduced in this study, capable of effectively harvesting large‐area wave energy while incorporating a hierarchical hybrid learning architecture. To increase the energy output, the electromagnetic generator (EMG) applies a bidirectional rotational differential with a planetary gear system. Under a wave frequency of 0.5 Hz and a rotation angle of 50°, the peak output power is 84.2 mW, while at 0.75 Hz and the same angle, it rises to 175.2 mW. Triboelectric nanogenerator (TENG) enhances detection sensitivity and resistance to interference by employing non‐contact triboelectrification along with a speed‐amplifying design. A data processing and transmission circuit (DPTC) is developed and, for the first time, coupled with a hierarchical hybrid learning architecture (HHLA) that fuses traditional machine learning with deep learning to predict wave frequency and amplitude. Within this architecture, 90% of frequency predictions fall within a 10% error margin (< 0.1 Hz), while 92% of amplitude predictions fall within a 4% error margin (< 1.4°). The HHLA model achieves MAEs of 0.07 for frequency and 0.4 for amplitude, both lower than those of the standalone Stacking and CNN‐LSTM‐Attention (CLA) baselines. These improvements significantly enhance prediction accuracy and support the practical monitoring of wave‐health conditions.

PUBLICATION RECORD

CITATION MAP

EXTRACTION MAP

CLAIMS

  • No claims are published for this paper.

CONCEPTS

  • No concepts are published for this paper.

REFERENCES

Showing 1-32 of 32 references · Page 1 of 1

CITED BY

Showing 1-4 of 4 citing papers · Page 1 of 1