Experimental data-driven multimodal approach for detection and localization of external disturbances in oil and gas pipelines using vibration characteristics

Monitoring pipeline infrastructure is essential for ensuring safety, efficiency, and environmental sustainability in energy transportation. External disturbances, such as knocking and drilling, are common challenges that can impact pipeline integrity. The proposed approach develops a multimodal framework for detecting and localizing external disturbances in oil and gas pipelines by integrating statistical methods, signal processing, and advanced diagnostics. The experimental study was conducted on pipelines under three conditions: healthy, knocking, and drilling to assess the capability of the framework to detect and localize external disturbances. The proposed approach effectively classified pipeline conditions. External disturbance localization was precise for knocking disturbances due to their distinct high-frequency characteristics. However, drilling localization was challenging with the current experimental setup. Drilling generated low-frequency vibrations with longer wavelengths, reducing energy loss and broader time responses across closely spaced nodes in the 2-meter pipe setup. This behavior provides insight into the spatial arrangement of the optimal node spacing for effective localization of low-frequency disturbances in long-range pipeline systems.