Sparse reconstruction of the flow field around a submerged platform in internal solitary waves

With the deep transformation of the global energy structure and the continuous advancement of the maritime power strategy, the development of deep-sea oil and gas and renewable energy is becoming an important direction of the new energy revolution. The complex and variable deep-sea marine environment poses severe challenges to traditional fixed platforms, promoting the rapid development of floating and submerged platform technologies. Submerged platforms, due to their strong resistance to flow erosion, high structural stability, and ability to avoid geological subsidence, are regarded as an important development path for future deep-water oil and gas extraction. However, internal solitary waves (ISWs) widely existing in deep stratified seas, with their propagation process accompanied by intense flow velocity and pressure fluctuations, can generate significant dynamic load impacts on submerged platforms and their auxiliary systems, threatening the safety and stability of deep-sea development. Accurately understanding the flow field distribution patterns around submerged platforms under the action of internal solitary waves is a key scientific issue for deep-sea structural protection design, dynamic response prediction, and safety assessment.

In real marine environments, although monitoring equipment such as Acoustic Doppler Current Profilers (ADCP) and pressure sensors can obtain local flow field information, their observation range is limited, equipment costs are high, and maintenance is difficult, making it challenging to achieve long-term, full-domain monitoring. Traditional numerical simulations can obtain relatively complete flow field distributions but rely on numerous boundary conditions, have high computational costs, and are difficult to meet real-time prediction requirements under complex sea conditions. How to achieve high-precision reconstruction and dynamic prediction of the global flow field around submerged platforms under sparse observation data conditions has become a core bottleneck problem urgently needed to be solved in the current deep-sea engineering field.

Figure 1: Schematic diagram of laboratory-generated internal solitary waves

To address the above major scientific and engineering challenges, this study originally proposed a sparse flow field reconstruction method that integrates Physics-Informed Neural Networks (PINN) with Attention Mechanisms, achieving high-fidelity reconstruction of the global flow field around submerged platforms under extremely sparse observation data conditions (as shown in Figure 2).

This method explicitly embeds the Navier-Stokes equations into the constraint system of deep neural networks, enabling the network to automatically follow fluid dynamics laws while learning flow field distributions, fundamentally improving the model’s physical consistency and cross-scenario generalization capabilities. By introducing multi-scale convolutional structures and bidirectional attention mechanisms, the model can adaptively capture the complex spatiotemporal coupling characteristics of flow fields during internal solitary wave propagation, achieving a breakthrough from local sparse observations to global flow reconstruction.

Experimental results (as shown in Figure 3) demonstrate that this method maintains high accuracy and strong robustness under various sparse sampling scenarios, significantly improving the accuracy and stability of internal solitary wave flow field prediction.

This research not only achieves dual innovation in theory and methodology in the field of sparse flow field reconstruction but also lays a technical foundation for future deep-sea intelligent platform safety perception, data-driven modeling, and autonomous decision-making.

Figure 2: Schematic diagram of the proposed method

Figure 3: Reconstruction effects under different levels of sparse data input

This research was published in the international authoritative journal “Ocean Engineering” in October 2025 under the title “Sparse reconstruction of the flow field around a submerged platform in internal solitary waves”.

Paper Authors: First author is Yi Guangmo, PhD student at Tsinghua University Shenzhen International Graduate School, and corresponding author is Associate Professor Ren Zhengru from the Polar Marine Equipment Technology Team at Shanghai Jiao Tong University.

Paper Link: https://doi.org/10.1016/j.oceaneng.2025.123033

[论文下载]