Numerical Simulations of Ship Liquid Tank Sloshing Based on Graph Neural Networks

doi:10.16182/j.issn1004731x.joss.25-FZ0629

Abstract

Abstract:

To address the high consumption of computational resources in simulating ship liquid tank sloshing using computational fluid dynamics simulation methods, a data-driven numerical simulation model was proposed based on graph neural networks. An encoder-processor-decoder framework was employed in the proposed model. The encoder extracted features of fluid particles from the first five time steps. The processor learnt latent motion patterns of fluid and updated features, and the decoder predicted features of particles at subsequent time steps. The processor incorporated a self-attention mechanism to enable dynamic adjacency weight allocation and emphasize the influence of irregular tank wall regions. Training data for the model were generated through the moving particle semi-implicit (MPS) method. Numerical simulations were conducted for rolling conditions of scaled ballast tanks. The results have demonstrated that the proposed model achieves over 50% improvement in simulation accuracy measured by MAE compared to conventional graph neural networks, while maintaining two orders of magnitude higher computational efficiency than the MPS method. This approach provides a novel numerical simulation for ship tank sloshing analysis that effectively balances precision and efficiency.

Key words: ship liquid tank sloshing, numerical simulation, self-attention mechanism, graph neural network, moving particle semi-implicit (MPS) method

CLC Number:

U663.85

Zhang Wenkang, Sun Xiaofeng, Zhong Yiping, Yin Yong. Numerical Simulations of Ship Liquid Tank Sloshing Based on Graph Neural Networks[J]. Journal of System Simulation, 2025, 37(12): 3087-3099.

Figures/Tables 26

Fig. 1

Table 1

Fig. 2

Fig. 3

Fig. 4

Fig. 5

Fig. 6

Fig. 7

Fig. 8

Table 2

Table 3

Table 4

Fig. 9

Table 5

Table 6

Fig. 10

Fig. 11

Fig. 12

Fig. 13

Table 7

Table 8

Fig. 14

Fig. 15

Fig. 16

Fig. 17

Fig. 18

References 24

[1]	Ma Bole, Yue Baozeng, Yan Sen, et al. Numerical and Experimental Study on the Nonlinear Liquid Sloshing in Cassini Tank[J]. Journal of Spacecraft and Rockets, 2025, 62(2): 466-476.
[2]	Belakroum R, Kadja M, Mai T H, et al. An Efficient Passive Technique for Reducing Sloshing in Rectangular Tanks Partially Filled with Liquid[J]. Mechanics Research Communications, 2010, 37(3): 341-346.
[3]	Qi Yan, Söding Heinrich, Stöcker Jasmin, et al. Prediction of Waves in Tanks Excited by Ship Motion and Moving Walls with a Three-dimensional Fully Linear Finite Difference Approach[J]. Ocean Engineering, 2024, 310, Part 1: 118648.
[4]	Buldakov E. Lagrangian Modelling of Fluid Sloshing in Moving Tanks[J]. Journal of Fluids and Structures, 2014, 45: 1-14.
[5]	朱仁庆, 侯玲. LNG船液舱晃荡数值模拟[J]. 江苏科技大学学报(自然科学版), 2010, 24(1): 1-6.
	Zhu Renqing, Hou Ling. Numerical Simulation of Liquid Sloshing in the Tanks of LNG Carrier[J]. Journal of Jiangsu University of Science and Technology(Natural Science Edition), 2010, 24(1): 1-6.
[6]	Chen Y F, Huang C, Yan W H, et al. Numerical Study on the Sloshing Behaviors of Dual Liquid Tanks with Gas Inflow[J]. Journal of Applied Fluid Mechanics, 2024, 17(9): 1885-1895.
[7]	Jiao Jialong, Ding Shang, Zhao Mingming, et al. Simulation of LNG Ship's Motions Coupled with Tank Sloshing in Regular Waves by DualSPHysics[J]. Ocean Engineering, 2024, 312, Part 2: 119148.
[8]	张之凡, 王越美, 李昌良, 等. 基于SPH方法的二维液舱晃荡特性研究[J]. 船舶力学, 2024, 28(5): 676-688.
	Zhang Zhifan, Wang Yuemei, Li Changliang, et al. Sloshing Characteristics of a 2D Rectangular Tank Based on SPH Method[J]. Journal of Ship Mechanics, 2024, 28(5): 676-688.
[9]	Sun Xiaofeng, Zhong Yiping, Bian Feng, et al. Numerical Computation of Sloshing-Induced Force in Complex Ship Tanks Under the Excitation of Ship Rolling Motion Based on the MPS Method[J]. Applied Sciences, 2022, 12(10): 5130.
[10]	Huang Congyi, Wang Jifei, Zhao Weiwen, et al. Numerical Simulation of Liquid Sloshing in a Spherical Tank by MPS Method[J]. Journal of Hydrodynamics, 2024, 36(2): 232-240.
[11]	金鑫, 王宇圣, 张福贵, 等. 基于神经网络的流体晃荡波高和压强的预测研究[J]. 船舶力学, 2025, 29(3): 388-399.
	Jin Xin, Wang Yusheng, Zhang Fugui, et al. Predictions of Wave Height and Pressure Induced by Liquid Sloshing Based on Neural Network[J]. Journal of Ship Mechanics, 2025, 29(3): 388-399.
[12]	Li Huan, Zhang Xinshuo, Yang Xiufeng. Numerical Study of Liquid Sloshing Using Smoothed Particle Hydrodynamics with Adaptive Spatial Resolution[J]. Engineering Analysis with Boundary Elements, 2024, 159: 272-287.
[13]	Hossein Goudarzvand Chegini, Zarepour Gholamreza. Utilizing Artificial Neural Network for Load Prediction Caused by Fluid Sloshing in Tanks[J]. Geofluids, 2021, 2021(1): 3537542.
[14]	Hossein Goudarzvand Chegini, Zarepour Gholamreza. Numerical Study of Rectangular Tank with Sloshing Fluid and Simulation of the Model Using a Machine Learning Method[J]. Geofluids, 2022, 2022(1): 4121956.
[15]	Zhang Ningbo, Yan Shiqiang, Ma Qingwei, et al. A CNN-supported Lagrangian ISPH Model for Free Surface Flow[J]. Applied Ocean Research, 2023, 136: 103587.
[16]	Jong Kim Ki, Kim Daegyoum. Prediction of Sloshing Pressure Using Image-based Deep Learning[J]. Ocean Engineering, 2024, 303: 117718.
[17]	Zhang Chongwei, Tan Jie, Ning Dezhi. Machine Learning Strategy for Viscous Calibration of Fully-nonlinear Liquid Sloshing Simulation in FLNG Tanks[J]. Applied Ocean Research, 2021, 114: 102737.
[18]	Zhang Songtao, Li Chenyang, Zhao Peng, et al. LSTM-based Prediction Control of Free-surface Tank[C]//2023 International Conference on Marine Equipment & Technology and Sustainable Development. Singapore: Springer Nature Singapore, 2023: 1013-1028.
[19]	Pfaff T, Fortunato M, Sanchez-Gonzalez A, et al. Learning Mesh-Based Simulation with Graph Networks[EB/OL]. (2021-06-18) [2025-05-18]. .
[20]	祝晓, 王子豪, 罗文清, 等. 基于数据驱动的液舱晃荡的低维表达方法研究[J]. 江苏船舶, 2024, 41(5): 12-15.
	Zhu Xiao, Wang Zihao, Luo Wenqing, et al. Research on Low Dimensional Expression Method of Liquid Tank Oscillation Based on a Data-driven Approach[J]. Jiangsu Ship, 2024, 41(5): 12-15.
[21]	Sanchez-Gonzalez A, Godwin J, Pfaff T, et al. Learning to Simulate Complex Physics with Graph Networks[C]//Proceedings of the 37th International Conference on Machine Learning. Chia Laguna Resort, Italy: PMLR, 2020: 8459-8468.
[22]	Li Yunzhu, Wu Jiajun, Tedrake R, et al. Learning Particle Dynamics for Manipulating Rigid Bodies, Deformable Objects, and Fluids[EB/OL]. (2019-04-18) [2025-05-26]. .
[23]	Deng Yangyu, Zhang Di, Cao Ze, et al. Spatio-temporal Water Height Prediction for Dam Break Flows Using Deep Learning[J]. Ocean Engineering, 2024, 302: 117567.
[24]	SDARI. Final Loading Manual of 400,000 Dwt Ore Carrier "Pacific Vision" [R]. Shanghai: Shanghai Merchant Ship Design and Research Institute, 2019.

时间/s	位移/(°)	角速度/(rad/s)	角加速度/(rad/s²)
0.6	-7.82	0.976 8	10.994 0
1.2	-9.75	-0.348 6	13.709 2
1.8	-4.34	-0.075 7	6.101 2

液舱参数	数值
液舱识别码	R2.03AP
液舱容积/m³	8 477.14
艉端肋位	117(222.24 m)
艏端肋位	123(238.38 m)
最低点距基线高/m	0
最高点距基线高/m	30.97
液舱宽度/m	20.18

模型参数	数值
缩尺模型y方向尺寸	403.60
缩尺模型z方向尺寸	619.48

算例	横摇角/(°)	横摇周期/s
Ⅰ	10	1.45
Ⅱ	10	1.80
Ⅲ	20	1.45
Ⅳ	20	1.80

时间/s	模型	CD	R²	RMSE	MAE	ΔMAE/%
0.6	GNN	1.309 7	0.966 5	0.007 9	0.006 5	72.3
0.6	GSANN	0.805 8	0.996 3	0.002 5	0.001 8	72.3
1.2	GNN	1.402 2	0.921 8	0.011 1	0.008 5	54.1
1.2	GSANN	1.183 1	0.941 6	0.004 9	0.003 9	54.1
1.8	GNN	1.563 0	0.816 1	0.028 0	0.023 8	58.8
1.8	GSANN	1.176 0	0.920 5	0.008 7	0.009 8	58.8