Intelligent monitoring of oil product leaks in case of damage of collapsible pipelines

Abstract

Modern pipeline systems for hydrocarbon transportation utilize a diverse array of technologies and equipment. The main oil product pipelines (MOPP) include modular (disassemblable) pipeline systems (MPS). MPS are mobile engineering complexes designed for the temporary transportation of crude oil, light petroleum products, and liquid hydrocarbons. These systems are utilized during the filling and emptying of MOPPs, as well as for scheduled maintenance and emergency response situations within the oil and gas sector.

During the operation of MPS, significant attention is focused on reliability and safety, in accordance with the directives from the President and Government of the Russian Federation aimed at improving the efficiency of oil and petroleum transport through advanced technologies. However, analyzing the experience of deploying the SRT indicate that, over distances of up to 150 km, losses can amount to 5, 5% (approximately 300,000 tons) of the total volume of transported oil products. Additionally, issues related to predictive monitoring and the timely detection of pipeline accident and damages remain largely unresolved.

One of the primary causes of product loss, as classified in MPS incidents, is the loss of pipeline integrity due to mechanical failures or operational accidents. To mitigate these losses, various monitoring systems and techniques are employed across MOPP facilities, relying on different operational principles and physical phenomena.

However, effective solutions for mobile modular pipeline systems are not available. Therefore, the development of advanced automated monitoring systems based on artificial intelligence is an urgent challenge. This paper presents a parameter-based method for detecting oil product leaks in modular pipeline systems and proposes a prototype and architecture of the oil product leakage monitoring system that incorporates a software layer powered by artificial intelligence.