%0 Journal Article
%T 基于序关系与聚类分析的气田站场关键放空排放源识别<br>Identification of Key Venting Emission Sources at Gas Field Stations Based on Ordinal Relationship Method and Cluster Analysis
%A 苏诗漫
%A 吴曼琪
%A 王琳玲
%A 李祎涵
%A 史春艳
%A 伍颖
%J Advances in Environmental Protection
%P 121-135
%@ 2164-5493
%D 2024
%I Hans Publishing
%R 10.12677/AEP.2024.141018
%X 放空排放是气田生产中重要环节之一，双碳背景下具体放空排放量的获取及关键排放源识别一直是实现气田生产放空减排的重要途径。鉴于气田站场放空数据的复杂性及其聚合性、不完整性及其数据量庞大等特征，本文提出了基于序关系法与聚类分析的关键排放源识别模型。首先依据气田站场工艺流程特征构建了4类30种生产中放空排放量化体系；其次采用序关系法并结合站场实际计算得到站场内具体放空排放源的排放量；再应用逆距离插值法对缺失值进行补充，采用三种聚类分析算法对关键排放源进行识别，对比分析结果并综合各算法特点，找到更加平衡的分类阈值并获取分类结果；最后将该方法应用于西南某气田，分析得出新的分类阈值为2.7756，并发现三种算法的关键排放源均主要分布在增压站中。研究表明，识别结果合理可靠，为关键放空排放源识别工作提供了新思路。<br />
Air release emission is one of the important links in gas field production, and the acquisition of specific air release emission and identification of key emission sources in the context of dual-carbon has been an important way to realize air release emission reduction in gas field production. In view of the complexity, aggregation, incompleteness and huge data volume of air release data from gas field stations, this paper proposes a key emission source identification method based on the ordinal relationship method and cluster analysis. First of all, based on the characteristics of gas field station process flow, we constructed a system of 4 categories and 30 types of air release emissions in production; secondly, we used the ordinal relationship method and combined with the actual calculation of the station to get the emissions of specific air release emission sources; then we applied the inverse distance interpolation method to supplement the missing values, and used three kinds of cluster analysis algorithms to identify the key emission sources, compared the results of the analysis and synthesized the characteristics of the algorithms, and then we found the balanced classification thresholds and obtained the classification results. Finally, the method is applied to an actual case in a gas field in Southwest China, and the new classification threshold is 2.775581, and it is found that the key emission sources of the three algorithms are mainly distrib-uted in the booster station. The study shows that the identification results are reasonable and re-liable, and provides new ideas for the identification of key venting emission sources.
%K 温室气体减排，气田站场，关键排放源，序关系法，聚类分析<br>Greenhouse Gas Emission Reduction
%K Gas Mine Production
%K Key Emission Sources
%K Ordinal Relationship Method
%K Cluster Analysis
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=81389