Fault Recognition Technology for Pipeline Systems Based on Multi-feature Fusion of Monitoring Data

Pipeline systems are important parts of modern infrastructure and complex industrial systems. Effective fault identification for pipeline systems is of great significance in ensuring safety and reliability. In this work, a method for pipeline fault identification based on data fusion analysis is proposed, which utilizes pipeline system condition monitoring data. First, monitoring data are analyzed to extract features that represent the operating state of the pipeline system, and a feature evaluation index based on sensitivity and volatility is proposed to optimize the extracted features. Second, multi-feature fusion analysis based on the Dempster–Shafer (DS) theory is performed to identify faults within the pipeline system. Meanwhile, to solve the problem of obtaining the basic probability assignment function (BPA) in DS theory, a BPA acquisition method is proposed, which considers both distance and correlation. Finally, this work is validated using case data of a residential heating test platform. The results show that the proposed method can directly use the monitoring data from the pipeline system for fault analysis and can effectively identify pipeline leakages, blockages, and other faults. The proposed method overcomes the shortcomings of traditional methods, which require a detailed mechanism analysis. It also conforms to emerging technology development trends, which utilize and apply Big Data analysis.