基于模糊理论的输油管道泄漏状态监测研究

Study of oil pipeline leakage detection based on fuzzy methods

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。

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.