海上油气田热介质供热设备故障数字化识别技术  

Sea Oil and Gas Field Thermal Medium Heating Equipment Faulty Digital Identification Technology

作　　者：姜学华[1] 赵乾菊 石振赟 庞井蛟 王盼盼 JIANG Xuehua;ZHAO Qianju;SHI Zhenyun;PANG Jingjiao;WANG Panpan(Dagang Oilfield Foreign Cooperation Project Department,Tianjin 300280,China;Shanghai Digitwins Co.Ltd.,Shanghai 201702,China)

机构地区：[1]大港油田对外合作项目部,天津300280 [2]上海数映科技有限公司,上海201702

出　　处：《工业加热》2023年第6期35-39,共5页Industrial Heating

基　　金：国家重点实验室油田数字化建设技术项目(20XH19212)。

摘　　要：海上油气田热介质供热设备长期在高温易燃的环境中运行,其可靠性对油气田开发的安全稳定起到重要作用,有必要研究其故障识别方法。为此,提出海上油气田热介质供热设备故障数字化识别方法。采用线性谱聚类算法对红外设备图像展开超像素分割处理,利用基于最大相似度区域合并算法分割设备目标区域。通过基于梯度变化的补偿算法对目标区域展开盲元块补偿。在核函数估计的基础上提取红外设备图像的温度概率密度函数,采用K均值聚类算法根据温度概率密度划分目标区域,将目标区域分为故障区域和正常区域,实现海上油气田热介质供热设备故障的数字化识别。实验结果表明,所提方法可以准确地完成目标区域分割和过热区域定位、故障识别准确率最低为97.3%、识别时间在2s内。The heat supply equipment of offshore oil and gas fields operates in high-temperature and flammable environment for a long time,and its reliability plays an important role in the safety and stability of oil and gas field development,so it is necessary to study its fault identification method.Therefore,a digital fault identification method of heat medium heating equipment in offshore oil and gas fields is proposed.The linear spectral clustering algorithm is used to carry out superpixel segmentation on infrared device images,and the region merging algorithm based on maximum similarity is used to segment the target region of the device.Blind block compensation is carried out in the target area by the compensation algorithm based on gradient change.On the basis of kernel function estimation,the temperature probability density function of infrared equipment image is extracted,and the target area is divided into fault area and normal area by using K-means clustering algorithm,so as to realize the digital identification of heat medium heating equipment faults in offshore oil and gas fields.Experimental results show that the proposed method can accurately segment the target area and locate the overheated area,with the lowest fault identification accuracy of 97.3%and the identification time within 2 s.

关 键 词：海上油气田 热介质供热设备 线性谱聚类算法 温度概率密度函数 故障识别 

分 类 号：TN219[电子电信—物理电子学]