为实现船舶系统及设备的实时状态评估，基于船舶实际运行故障数据不易获得、数据结构非线性、数据量巨大以及噪声多等特征，本文采用核主成分分析法，以船舶中央冷却器为例，选择高斯核函数及不同核参数，仅利用高维的正常运行数据，在特征空间中建立相应的核主成分评估模型，并对异常运行数据进行评估分析。评估结果表明，在合适的核参数下，核主成分分析法无需深入分析中央冷却器的结构与原理，即可快速有效地区分其非线性结构的正常运行数据和异常运行数据，其准确率优于常规主成分分析法，且其倒V字型的评估输出特性辨识度高，对微小故障较为敏感，非常适合用于突发性故障的早期识别。对于船舶机械设备而言，具有重要的工程实际应用意义。
In order to achieve real-time status evaluation of ship systems and equipment, based on the characteristics of the actual engineering operation parameter fault data which are not easy to obtain, nonlinear data structure, huge amount of data and much noise, this paper adopts the kernel principal component analysis method, takes the ship central cooler as an example, selects Gaussian kernel function and different kernel parameters, and establishes the corresponding kernel principal component assessment model in the feature space by using only the high-dimensional normal operation data, and evaluates and analyzes the abnormal operation data. The evaluation results show that under the appropriate kernel parameters, the kernel principal component analysis method can quickly and effectively distinguish the normal operation data and abnormal operation data of the nonlinear structure without in-depth analysis of the structure and principle of the central cooler, and its accuracy is better than that of the conventional principal component analysis method, and its inverted V-shaped evaluation output characteristics are highly discriminative and sensitive to small faults, which is very suitable for the early identification of sudden failures. For ship machinery and equipment, it has important significance for practical application in engineering.
2025,47(9): 65-71 收稿日期：2024-6-19
DOI：10.3404/j.issn.1672-7649.2025.09.012
分类号：U664.5+3
作者简介：吴小豪（1991-），男，硕士，研究方向为智能船舶等
参考文献：
[1] 吴小豪, 张跃文, 孙培廷, 等. 基于超球支持向量机的船舶高温淡水系统状态评估[J]. 船舶工程, 2017, 39(2): 49-52+93.
WU X H, ZHANG Y W, SUN P T. State evaluation of ship high-temperature fresh water system based on hypersphere support vector machine[J]. Ship Enmineering, 2017, 39(2): 49-52+93.
[2] 鲍中新, 文成林, 马雪. 一种基于数据变化率的预处理及主元分析故障诊断方法[J]. 电子学报, 2021, 49(11): 2234-2240.
BAO Z H, WEN C L, MA X. Data preprocessing and PCA fault diagnosis method based on rate of change transformation[J]. Acta Electronica Sinica, 2021, 49(11): 2234-2240.
[3] 高亚娟. 基于全矢主成分分析的故障预测方法研究[D]. 郑州: 郑州大学, 2018.
[4] 孟程程, 曾九孙, 李文军. 核主成分分析的高炉故障检测研究[J]. 中国计量学院学报, 2012, 23(4): 332-337.
MENG C C, ZENG J S, LI W J. Blast furnace fault detection based on KPCA[J]. Journal of China University of Metrology, 2012, 23(4): 332-337.
[5] 武凯, 孙彦广, 张琳. 基于核主成分分析的热连轧断带故障诊断[J]. 中国冶金, 2020, 30(11): 60-65.
WU K, SUN Y G, ZHANG L. Fault diagnosis of strip breaking in hot strip rolling based on kernel principal component analysis[J]. China Metallurgy, 2020, 30(11): 60-65.
[6] 吴天昊, 刘韬, 施海宁, 等. 基于核主元分析法的核电厂设备状态监测技术研究[J]. 核动力工程, 2020, 41(5): 132-137.
WU T H, LIU T, SHI H N. Research on condition monitoring technology for nuclear power plant equipment based on kernel principal component analysis[J]. Nuclear Power Engineering, 2020, 41(5): 132-137.
[7] 荣智海, 齐波, 张鹏, 等. 基于核主成分分析的油色谱在线监测装置异常状态快速辨识[J]. 高电压技术, 2019, 45(10): 3308-3316.
RONG Z H, QI B, ZHANG P, et al. Fast identification of abnormal state based on kernel principal component analysis for oil chromatography analysis on-line monitoring Device[J]. High Voltage Engineering, 2019, 45(10): 3308-3316.
[8] 石新发, 贺石中, 谢小鹏, 等. 基于核主成分的船舶柴油机磨损信息特征提取方法研究[J]. 武汉理工大学学报(交通科学与工程版), 2022, 46(6): 1039-1043.
SHI X F, HE S Z, XIE X P, et al. Research on feature extraction method of marine diesel engine wear information based on kernel principal component[J]. Journal of Wuhan University of Technology(Transportation Science & Engineering), 2022, 46(6): 1039-1043.
[9] 孙宜权, 张英堂, 李志宁, 等. 基于核主成分分析的柴油机技术状态评估[J]. 车用发动机, 2012(2): 89-92.
SUN Y Q, ZHANG Y T, LI Z N, et al. Evaluation of diesel engine technical state based on KPCA[J]. Vehicle Engine, 2012(2): 89-92.
[10] 张珂, 宋文丽, 石怀涛, 等. 基于改进核主元分析的故障检测方法研究[J]. 控制工程, 2017, 24(2): 418-424.
ZHANG K, SONG W L, SHI H T, et al. Fault detection based on improved kernel principal component analysis[J]. Control Engineering of China, 2017, 24(2): 418-424.
[11] 张小龙. 基于改进核主成分分析的航空液压管路-卡箍故障诊断研究[D]. 鞍山: 辽宁科技大学, 2023.
[12] 高明. 基于主成分分析的综合评价研究及其应用[D]. 天津: 南开大学, 2006
[13] 李智. 基于主元分析的故障诊断方法研究及应用[D]. 沈阳: 东北大学, 2012.
[14] 董佳. 基于主元分析的多故障状态监测与诊断方法研究[D]. 沈阳: 东北大学, 2013.
[15] 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.
[16] 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.
[17] 李贺. 基于核主元分析的风电机组变桨距系统故障诊断研究[D]. 沈阳: 沈阳工业大学, 2017.
