基于智能模糊推理系统的船型概念方案快速生成研究  

Rapid ship hull conceptual scheme design based on intelligent fuzzy inference system

作　　者：杨萌 龚俊斌 曹晋 周塔 YANG Meng;GONG Junbin;CAO Jin;ZHOU Ta(Hanjiang National Laboratory,Wuhan 430060,China;School of Mechanical Engineering,Jiangsu University of Science and Technology,Zhenjiang 212114,China;School of Computer,Jiangsu University of Science and Technology,Zhenjiang 212114,China)

机构地区：[1]汉江国家实验室,湖北武汉430060 [2]江苏科技大学机械工程学院,江苏镇江212114 [3]江苏科技大学计算机学院,江苏镇江212114

出　　处：《中国舰船研究》2024年第6期45-55,共11页Chinese Journal of Ship Research

基　　金：国防基础科研计划资助项目(JCKY2020206B037);国家春晖计划项目(HZKY20220133);江苏省自然科学基金面上项目资助(BK20191200,BK20241831)。

摘　　要：[目的]针对现有船舶概念方案设计阶段中,船型概念方案生成技术对具有模糊特性的用户需求难以建模的问题,研究一种基于模糊推理系统的船型概念方案快速生成系统,以快速推出符合用户要求的船舶性能数值,并生成与之匹配的船型设计参数作为概念方案。[方法]首先,收集并整理现有船型的数据作为先验知识,用于在用户需求建模过程中超参数初始化。其次,基于模糊集理论将用户需求定量地映射至模糊空间,利用整理的先验知识以可解释的方式推理出符合用户需求的船舶性能数值。最后,对符合用户需求的每个船舶性能数值匹配出最佳的船型设计参数,并通过模糊集理论中的“且”原理,进一步推理出符合用户对全部船舶性能要求的船型设计参数,将其作为船型概念方案。[结果]实验结果显示,智能模糊推理系统能够于30 s内根据模糊的用户需求快速推理出多条偏差在3.5%以内的船型概念方案。[结论]智能模糊系统可量化用户需求中的模糊性,经过两个阶段的推理,高效地生成符合用户各项需求的船舶概念方案,研究成果可为智能化、快速化的船舶概念设计的方案生成提供有益的参考。[Objective]Aiming at the problem that it is difficult to model user requirements with fuzzy characteristics in existing ship conceptual scheme design(SCSD),this study focuses on a rapid SCSD system based on a fuzzy inference system(FIS) that can quickly obtain ship performance values that meet the user requirements and generate the corresponding ship hull design parameter values.[Methods]First,the existing ship hull parameter data is collected and sorted as prior knowledge for hyper-parameter initialization in the process of user requirement modeling.Second,based on the fuzzy set theory,the user requirements are quantitatively mapped to the fuzzy space,and the sorted prior knowledge is used to infer the ship performance values that meet the user requirements in an interpretable way.Finally,the best ship hull design parameters are matched for each ship performance value that meets the user requirements,and through the "AND" principle in fuzzy set theory,the ship design parameters that meet all the user requirements for ship performance are further inferred and taken as the SCSD.[Results]The experimental results show that the intelligent FIS can quickly infer multiple SCSDs with deviations within 3.5% according to the fuzzy user requirements within 30seconds.[ Conclusions] The intelligent FIS quantifies the fuzziness in user requirements,and after two stages of inference,the SCSD that meets the various user requirements is efficiently generated.The results of this study can provide useful references for rapid intelligent SCSD generation.

关 键 词：船舶设计 人工智能 船舶性能要求 船型设计参数 船型概念方案生成 模糊集理论 模糊规则 

分 类 号：U662.2[交通运输工程—船舶及航道工程] U662.9[交通运输工程—船舶与海洋工程]