摘要
为保障船舶安全高效运行,提出4种船舶柴油机运行参数异常检测方法。该技术在空间向量统计的异常检测基础上,进一步提出新的方法监测运行参数时间序列的趋势异常。通过提取运行参数数据演化过程的趋势和规律,可以更早、更准确地识别异常趋势,为设备管理提供决策支持。实验结果表明,所提出的异常检测技术能够有效提高船舶柴油机异常检测的效率和准确性,及早发现安全隐患。
In order to ensure the safe and efficient operation of ships,four methods for detecting anomalies in the operating parameters of marine diesel engines were proposed.On the basis of the vector statistics-based anomaly detection,this technique could further introduce new methods to monitor the trend anomalies in the time series of operating parameters.By extracting the trends and patterns in the evolution process of operating parameter data,it is possible to identify abnormal trends earlier and more accurately,providing decision support for equipment management.Experimental results demonstrated that the proposed anomaly detection techniques can effectively improve the efficiency and accuracy of marine diesel engine anomaly detection,enabling the early detection of safety hazards.
作者
黄滔
陈冬梅
杨勇兵
HUANG Tao;CHEN Dong-mei;YANG Yong-bing(Shanghai Marine Diesel Engine Research Institute,Shanghai 201108,China)
出处
《船海工程》
北大核心
2024年第4期66-70,共5页
Ship & Ocean Engineering
基金
工信部项目(工信部重装函[2020]313号)。
关键词
时序数据
异常检测
趋势异常
变点检测
冲高回落异常
time series data
anomaly detection
trend anomaly
change point detection
abnormal sharp rise and fall