基于模糊理论的输油管道泄漏状态监测研究

管道泄漏监测的主要目的是正确评价管道运行过程中的启停泵、开关阀等正常操作与管道发生泄漏事故的状态。为此,阐述了一种基于负压波波动理论与模糊C-均值聚类算法相结合的混合评估方法,其基本思想为:对预处理后的管道运行实时压力波动进行动态聚类,然后求解动态聚类中心的波动指数,波动指数越小,管道运行状态越好;反之,则认定管道运行状态差,应迅速采取相应措施。实验验证了混合算法的可行性与稳定性,评价结果与实际情况吻合,表明该混合算法对管道运行状态评价的有效性和实用性。

Study of oil pipeline leakage detection based on fuzzy methods

The main purpose of pipeline leakage detection system is to evaluate the running condition of the pipeline precisely, which aims to distinguish the conditions between starting/shutting down the pumps, opening/closing the valves from leakage accidents. However, one of the traditional ways to evaluate the condition is the utilization of fuzzy C-means clustering method that can cluster the original patterns correctly, but can not classify the running state of pipelines automatically. This paper presents a new scheme integrating with fluctuation exponent and fuzzy C-means clustering results to resolve the problems. The scheme is based on the classified patterns by fuzzy C-means method, then by calculating fluctuation exponent to evaluate. It also introduces a basic concept and structure of the intelligent leak detection system. Primarily an integrated leak-detection model is developed to carry out the remote monitoring and leakage detection. The application results show that the whole running status of pipeline can be monitored effectively, and a higher automation grade and an excellent leak positioning precision of the system can be obtained.