摘要
针对舰船智能机舱多源数据处理,建立基于DBSCAN聚类算法的多源数据异常检测方法,包含聚类参数的自动获取及校正,并可根据试验数据进行机器学习和更新。介绍该检测方法的程序框架、子程序的程序流程,并基于机舱中典型的滑油系统进行数据检测测试。经过测试,该检测方法可有效对测试数据进行分析,自动得出并修正聚类参数,且对正常信号相差15%以上的异常数据的检测结果识别度达到100%。本文提出的异常数据检测方法可行有效,可以作为智能机舱数据处理系统的方案之一。
To provide effect method of intelligent engine room multi-source data processing for ship,the multi-source data detecting method based on DBSCAN cluster analysis algorithm was established,which could automatically get the cluster parameters,adjusted and updated based on machine learning.The program frame of the method and flow chart of subprogram were introduced.And a test for lubricating oil system data detection was implemented,which shows the test data could be effectively analyzed and the cluster parameters could be obtained and adjusted.The test indicated that the detecting method recognition rate achieves 100%against abnormal data,15%differed from normal data.The detecting method was feasible and effective,which could be one of the solutions for intelligent engine room multi-source data processing.
作者
陈砚桥
孙彤
张侨禹
CHEN Yan-qiao;SUN Tong;ZHANG Qiao-yu(College of Power Engineering,Naval University of Engineering,Wuhan 430033,China;China Ship Development and Design Center,Wuhan 430064,China)
出处
《舰船科学技术》
北大核心
2021年第9期156-160,共5页
Ship Science and Technology
关键词
智能机舱
多源数据处理
DBSCAN聚类算法
异常数据检测
intelligent engine room
multi-source data processing
DBSCAN cluster algorithm
abnormal data detection