Researchers from the University of Chinese Academy of Sciences, including Yi Zhou, Li Wang, Hang Su, and Tian Wang, have developed a novel method for estimating significant wave height (Hs) from radar data, which could have practical applications in the energy sector, particularly for offshore wind and wave energy industries.
The team’s approach, called the Spectral Point Transformer (SPT), focuses on analyzing sparse spectral points with strong power, as these points are empirically observed to contribute the most to wave energy. By integrating geometric and spectral characteristics of ocean surface waves, the SPT method uses a transformer-based model to estimate Hs through multi-dimensional feature representation. The researchers found that the learned features of SPT align well with physical dispersion relations, with the contribution-score map of selected points concentrated along dispersion curves.
Compared to conventional vision networks that process image sequences and full spectra, the SPT method demonstrates superior performance in Hs regression while consuming significantly fewer computational resources. The team reported that on a consumer-grade GPU, SPT completed the training of a regression model for 1080 sea clutter image sequences within just four minutes. This efficiency could potentially reduce deployment costs for radar wave-measuring systems, making them more accessible for various applications in the energy industry.
For instance, accurate wave height estimation is crucial for offshore wind farm development and operation, as it helps in designing robust wind turbines, predicting maintenance needs, and ensuring the safety of offshore installations. Additionally, wave energy converters rely on precise wave height data to optimize their performance and maximize energy capture. The SPT method’s efficiency and accuracy could therefore contribute to the advancement of these renewable energy technologies.
The research was published in the journal Remote Sensing of Environment, and the open-source implementation of SPT is available on GitHub for further exploration and development. As the energy sector continues to embrace digitalization and data-driven decision-making, innovative methods like SPT can play a vital role in improving the efficiency and reliability of offshore energy systems.
This article is based on research available at arXiv.