基于管道完整性数据的关联规则挖掘与应用
Mining and application of association rules based on pipeline integrity data
-
摘要: 在油气长输管道完整性管理领域引入数据挖掘中的关联规则技术, 发现完整性数据中潜在的关联关系, 充分发挥数据价值。通过对管道完整性数据关联规则挖掘流程进行研究, 对经典Apriori算法中频繁项集生成效率进行优化, 结合中国石油某管道开展完整性管理积累的外检测与内检测数据进行了关联规则挖掘, 并对挖掘结果进行分析和解释。挖掘结果表明: 通过关联规则技术可以发现管道本体缺陷与周边环境、本体属性数据之间潜在的关联关系。关联规则挖掘方法应用于管道完整性数据分析能够有效减少无兴趣规则的数量, 发现潜在的管理重点, 为长输管道完整性管理提供科学、准确的决策依据。Abstract: To identify potential correlation in pipeline integrity data and take full advantages of such data, this paper introduces the association rules in data mining to the scope of integrity management in long-distance oil and gas pipeline industry. By studying procedures related to association rules of pipeline integrity data, generation efficiency of frequent item sets in the classical Apriori algorithm is optimized. In addition to mining of the association rules in external and internal inspection data accumulated in integrity management over a pipeline of PetroChina, relevant results obtained through mining are analyzed and interpreted. Research results indicate that pipeline defects have potential correlation with environmental properties and pipeline attributes by using the association rules. With the technique for mining of association rules in pipeline integrity data, it is possible to minimize the number of uninterested rules and to highlight key points in management. Research results may provide accurate and scientific basis for integrity management of long-distance pipelines.