为实现船舶系统或设备的实时状态评估，本文采用核Fisher判别分析法，以船舶中央冷却器为例，选择合适的核函数及核参数，利用其正常数据和异常数据建立状态评估模型，即最佳投影方向，并利用过程数据验证其有效性。结果表明，核Fisher判别分析法无需深入分析中央冷却器的结构与原理即可有效识别中央冷却器的正常工况和异常工况，同时能够通过投影值准确描述过程工况的变化过程。在故障发展初期，根据运行参数投影值的变化趋势，可判断船舶系统或设备状态的发展趋势，为早期发现船舶系统或设备的重复性故障提供有效手段。对于船舶系统或设备而言，具有重要的工程实际应用意义。
To achieve real-time condition assessment of ship systems or equipment， this paper adopts the kernel Fisher discriminant analysis method, and takes the ship central cooler as an example, and select appropriate kernel function and parameters, and establishes a state evaluation model using its normal and abnormal data, that is, the optimal projection direction, and uses the process data to verify its effectiveness. The evaluation results show that the kernel Fisher discriminant analysis can effectively identify the normal and abnormal working conditions without in-depth analysis of the structure and principle of the central cooler, and can accurately describe the change process of the working conditions through the projected value. At the initial stage of fault development, the development trend of ship systems and equipment status can be judged according to the change trend of the projected values of operating parameters, which provides an effective means for early detection of repetitive faults of ship systems and equipment. For Marine systems and equipment, it has important engineering practical application significance.
2025,47(2): 185-189 收稿日期：2024-3-7
DOI：10.3404/j.issn.1672-7649.2025.02.030
分类号：U664.5+3
作者简介：吴小豪（1991 – ），男，硕士，工程师，研究方向为智能船舶
参考文献：
[1] 吴小豪. 基于超球支持向量机的船舶高温淡水系统状态评估[D]. 大连: 大连海事大学, 2017.
[2] 章浩伟, 高燕妮, 苑成梅等. 基于Fisher判别分析的双相障碍识别研究[J]. 中国医学物理学杂志, 2015, 32(1): 134-138.
ZHANG H W, GAO Y N, YUAN C M, et al. Bipolar disorder recognition research based on Fisher discriminant analysis[J]. Chinese Journal of Medical Physics, 2015, 32(1): 134-138.
[3] 梁亮, 袁洪芳, 曹晰. 基于费舍尔判别分析法的故障诊断[J]. 计算机工程与设计, 2008(11): 2897-2900.
LIANG L, YUAN H F, CAO X. Fault diagnosis based on Fisher's discriminant analysis[J]. Computer Engineering and Design, 2008(11): 2897-2900.
[4] 于健, 谷洪彪, 王贺, 等. 基于核Fisher判别分析方法对北京温榆河及部分支流水质综合评价[J]. 数学的实践与认识, 2020, 50(22): 103-110.
YU J, GU H B, WANG H, et al. Comprehensive evaluation of water quality of Wenyu River and its tributaries in beijing based on kernel fisher model[J]. Mathematics in Practice and Theory, 2020, 50(22): 103-110.
[5] 祝芹芹. 基于核Fisher的多机牵引道岔故障诊断技术研究[D]. 兰州: 兰州交通大学, 2019.
[6] 王鹏, 胡向阳, 魏水建. 基于核Fisher判别的复杂储层岩性识别[J]. 科学技术与工程, 2017, 17(31): 29-34.
WANG P, HU X Y, WEI S J. Lithologic identification of complex reservoirs based on kernel Fisher discriminant analysis[J]. Science Technology and Engineering, 2017, 17(31): 29-34.
[7] 杨东海, 胡凌, 钱莹. 基于核Fisher判别分析的高职学生考试成绩预测[J]. 中国教育信息化, 2016(16): 76-79.
YANG D H, HU L, QIAN Y. Prediction of exam scores for vocational college students based on kernel Fisher discriminant analysis[J]. The Chinese Journal of ICT in Education, 2016(16): 76-79.
[8] 周欣, 吴瑛. 核Fisher判别分析在数字信号分类中的应用[J]. 北京邮电大学学报, 2011, 34(1): 35-39.
ZHOU X, WU Y. Application of kernel fisher discriminant analysis to digital signal classification[J]. Journal of Beijing University of Posts and Telecommunications, 2011, 34(1): 35-39.
[9] 叶和军. 基于随机Fisher判别分析的故障诊断方法研究[D]. 杭州: 浙江理工大学, 2023.
[10] 邓乃扬, 田英杰. 支持向量机理论、算法与拓展[M]. 北京: 科学出版社, 2009.
[11] 钟世勇. 基于费舍尔判别分析的半监督故障分类方法研究[D]. 杭州: 浙江大学, 2015.
