本文针对船舶快速性能高效预报中存在的难度,以33艘油船、43艘散货船水池模型试验数据为基础,应用机器学习算法建立船舶快速性能预报的代理模型。研究了不同组合方式的船体有量纲与无量纲参数、代理模型种类对船型快速性能预报精度的影响,对比了代理模型在相近船型预报和同类船型预报2种应用场景中的适用性。结果显示,有量纲与无量纲混合的参数组合可以提高模型预报精度,船舶快性能代理模型预报误差约为3%~7%。
Aiming at the difficulties in fast performance prediction of ships, based on the test data of 33 oil tankers and 43 bulk carriers, a surrogate model for fast performance prediction of ships is established by using machine learning algorithm. This paper studies the influence of different combinations of hull dimensional and non dimensional parameters, surrogate model types on the accuracy of ship form fast performance prediction, and compares the applicability of surrogate model in two scenarios of similar ship form prediction and similar ship form prediction. The results show that the combination of dimensional and dimensionless parameters can improve the prediction accuracy of the model, and the prediction error of the ship fast performance surrogate model is about 3%~7%.
2022,44(4): 126-131 收稿日期:2021-02-24
DOI:10.3404/j.issn.1672-7649.2022.04.026
分类号:U663.2
作者简介:陈小平(1983-),男,硕士,高级工程师,研究方向为船体综合性能研究
参考文献:
[1] 曾帅. 典型海况下的船舶阻力预报研究[D]. 大连: 大连海事大学, 2020.
[2] 黄超. 肥大型船舶阻力快速预报方法研究[D]. 哈尔滨: 哈尔滨工程大学, 2014.
[3] 盛振邦, 刘应中. 船舶原理[M]. 上海: 上海交通大学出版社, 2004.
[4] 何清, 李宁, 罗文娟, 等. 大数据下的机器学习算法综述[J]. 模式识别与人工智能, 2014, 27(4): 327–336
[5] HADDARA M, R, WANG Y. Parametric Identification of manoeuvering models for ships[J]. International Shipbuilding Progress, 1999, 46(445): 5–27
[6] HESS D., FALLER W. Simulation of ship maneuvers using recursive neural networks[C]//23th Symposium on Naval Hydrodynamics, Washington D. C. (USA): National Academies Press, 2000: 223−242.
[7] CHIU F. C., CHANG T. L., GO J., et al. A Recursive Neural Networks Model for Ship Maneuverability Prediction[C]//Proceeding Oceans MTS/IEEE Techno-Ocean, Kobe(Japan): IEEE, 2004: 1211−1218.
[8] 张恒, 詹成胜, 刘祖源, 等. 基于船舶阻力性能的船型主尺度参数敏感度分析[J]. 船舶工程, 2015(6): 11–14
[9] 刁峰, 周伟新, 魏锦芳, 等. 基于模型试验的船舶最小推进功率研究[J]. 中国造船, 2017, 58(4): 31–37
[10] 廖兴涛, 基于代理模型的汽车碰撞安全性仿真优化研究[D], 长沙: 湖南大学, 2006.
[11] 肖振业, 冯佰威, 刘祖源, 等. 基于支撑向量机的船舶阻力近似模型[J]. 计算机辅助工程, 2015, 24(4): 20–23
[12] BESNARD E, SCHMITZ A, HEFAZI H, et al. Constructive neural networks and their application to ship multi-disciplinary design optimization[J]. Journal of Ship Research, 2007.