In order to prevent the failure of oil gathering and transportation pipelines,based on the pipeline failure records of an oil field from 2018 to 2021,the failure characteristics of various pipelines were analyzed by mathematical statistics method,and the types of accidents,common inducements and rules of association among them analyzed by word cloud image and Apriori algorithm;then a co-occurrence network was constructed by co-occurrence matrix,to identify the key inducement items and acceleration items of different inducement items;and at last the coupling relation of accident development was determined by multiple text mining techniques.The results show that the low-pressure pipeline segment of crude oil and the medium-pressure pipeline segment of natural gas located in the high-consequence class III region are the focus of risk management and control.Pipeline leakage,oil and gas accumulation,oil pool fire,flash fire are common types of accident,construction damage,soil corrosion,corrosion layer damage,pipeline pressure,containing corrosive substances common causes of accidents;pipeline leakage is in strong correlation with oil and gas accumulation.Under pipeline leakage,the probability of oil and gas accumulation is 85.21%,and the probability of their simultaneous occurrence 18.21%.Co-occurrence of construction damage,soil corrosion,failure of applied current protection and pipeline pressure occupying are the main factors behind pipeline failure.The main causes of soil corrosion and internal corrosion acceleration are related to a number of inducements,and co-occurrence of multiple inducements should be avoided as far as possible in the subsequent risk management and control.
关键词
集输管道/诱因/文本挖掘/词云图/Apriori算法/共现网络
Key words
gathering and transportation pipeline/inducement/text mining/word cloud map/Apriori algorithm/co-occurrence network
维普
万方数据