Research subject: Multi-level leakage modeling in horizontal and inclined pipelines using real-time sensor data analysis
Research approach: The environmental severity of oil pipeline leakage depends on the leak volume, which is directly related to the time required for leak detection and containment. The use of real-time sensors installed at multiple locations along the pipeline facilitates effective leak detection through continuous data collection. Leak detection simulation for repair planning is a critical factor in pipeline design, implementation, and evaluation. This study proposes a framework for generating pipeline datasets using PipeSim software and Python’s random function. A horizontal–inclined pipeline configuration was initially analyzed, and pressure profiles under leak and non-leak conditions were evaluated. Subsequently, leak analysis was conducted on an inclined pipeline with a slope angle of 20 degrees. Because leakage alters the pressure profile, pressure variations were used as inputs to a machine learning algorithm, and the pressure distribution along the pipeline was simulated accordingly.
Main results: Before leakage, the local pressure drop at the leak point was 1270 psi in the horizontal–inclined pipeline and 792 psi in the inclined pipeline. After leakage, the minimum pressure dropped to 800 psi in the horizontal–inclined pipeline and 500 psi in the inclined pipeline. These results indicate a greater pressure drop reduction at the leak point in the horizontal–inclined pipeline (470 psi) compared to the inclined pipeline. This behavior is attributed to the higher flow velocity induced by gravity in the inclined pipeline, which mitigates severe pressure drops at the leak location. Consequently, leak severity decreases under conditions of high flow velocity and lower pressure reduction.