基于优化SVMs多传感器融合的石化输油管网泄漏智能监测定位算法

Intelligent Monitoring and Localization Algorithm for Leakage in Petrochemi⁃cal Pipeline Network Based on Optimized SVMs and Multi-Sensor Fusion

为了实现对石化输油管网泄漏的智能监测与定位,文章提出了一种基于优化支持向量机(SVMs)的多传感器融合算法,用于石化输油管网泄漏的智能监测与定位。该算法融合了压力传感器、流量传感器和气体检测传感器等信息,通过提取奇异值流形特征(SVMF)来表征泄漏故障,并利用改进的果蝇优化算法选择SVMs模型的最优参数。同时,引入Dempster–Shafer证据理论实现多传感器信息的决策层融合,消除了诊断结论的冲突,提高了模型的鲁棒性和准确性。试验结果表明,该算法在石化输油管网泄漏监测与定位方面表现出色,具有广阔的应用前景。

In order to achieve intelligent monitoring and location of leaks in petrochemical oil pipeline networks,this paper proposes a multi-sensor fusion algorithm based on optimized Support Vector Machines(SVMs)for the intelligent monitoring and positioning of leaks in these networks.The algorithm integrates information from pressure sensors,flow sensors,and gas detection sensors,representing leakage faults by extracting Singular Value Manifold Features(SVMF).It employs an improved fruit fly opti-mization algorithm to select the optimal parameters for the SVMs model.Additionally,Dempster-Shafer evidence theory is intro-duced to achieve decision-level fusion of multi-sensor information,resolving conflicts in diagnostic conclusions and enhancing the model′s robustness and accuracy.Experimental results demonstrate that this algorithm performs excellently in the monitoring and location of leaks in petrochemical oil pipeline networks,presenting broad application prospects.