可靠掌握电气设备的运行状态,是保证船舶安全航行的基础。因此,提出基于热红外图像的船舶电气设备状态异常检测方法。该方法依据红外成像技术获取船舶电气设备成像,获取其热红外图像结果,并计算电气设备温度概率密度函数,以此描述电气设备的温度分布特征。将该概率密度函数计算结果输入具备增量学习的宽度学习算法中,完成船舶电气设备不同异常状态检测。测试结果显示,将温度概率密度作为电气设备状态异常检测依据,能够更好地区分电气设备的正常放热以及故障升温;AUC的测试结果均在0.94以上,可确定电气设备运行过程中的不同程度异常状态。
Reliable understanding of the operating status of electrical equipment is the foundation for ensuring safe navigation of ships. Therefore, a method for detecting abnormal status of ship electrical equipment based on thermal infrared images is proposed. This method obtains imaging of ship electrical equipment based on infrared imaging technology and obtains its thermal infrared image results, and calculate the probability density function of electrical equipment temperature to describe the temperature distribution characteristics of electrical equipment; Input the calculated result of the probability density function into a width learning algorithm with incremental learning to complete the detection of different abnormal states of ship electrical equipment. The test results show that using temperature probability density as the basis for detecting abnormal electrical equipment status can better distinguish between normal heat release and fault heating of electrical equipment. The test results of AUC are all above 0.94, identify varying degrees of abnormal conditions during the operation of electrical equipment.
2024,46(3): 147-150 收稿日期:2023-09-05
DOI:10.3404/j.issn.1672-7649.2024.03.026
分类号:TP18
基金项目:河南省重点研发与推广专项(科技攻关)项目(222102210090)
作者简介:崔海花(1976-),女,副教授,研究方向为电气工程
参考文献:
[1] 赵颖祺, 陈玉虎, 张晰, 等. 一种基于箱线图的SAR图像船舶检测算法研究[J]. 中国海洋大学学报(自然科学版), 2021, 51(10): 130-140.
[2] 曾军, 王东杰, 范伟, 等. 基于红外热成像的电气设备组件识别研究[J]. 红外技术, 2021, 43(7): 679-687.
[3] 宋世静. 基于数据挖掘的电气设备状态自动检测方法[J]. 自动化技术与应用, 2023, 42(8): 133-136.
[4] 律方成, 牛雷雷, 王胜辉, 等. 基于优化YOLOv4的主要电气设备智能检测及调参策略[J]. 电工技术学报, 2021, 36(22): 4837-4848.
[5] 翟洪婷, 张庆锐, 卞若晨, 等. 基于图聚类的电力设备异常声音检测方法[J]. 南京理工大学学报, 2022, 46(3): 270-276.
[6] 党晓婧, 刘顺桂, 朱光南, 等. 基于特征提取的电气设备红外图像识别算法[J]. 沈阳工业大学学报, 2023, 45(3): 264-269.
[7] 霍成军, 史奕龙, 武晓磊, 等. 基于热成像技术的电气设备目标检测方法[J]. 激光与红外, 2021, 51(4): 530-536.