Automated Subsea Pipeline Leak Detection Using Real-Time Downhole Gauges

Abstract

Abstract In oil industry, there are several technologies used in detecting subsea pipeline leaks. These technologies are not costly only, but require extensive implementation fieldwide across hundreds of kilometers pipelines. Currently, there are few digital solutions or systems in the literature addressing the detection mechanism. In this study, an automated mechanism will be introduced to locate leaks in subsea pipeline network coupled with real-time monitoring. Real-time downhole pressure data from wells' permanent downhole monitoring systems (PDHMSs) are used to map the pressure drop across each reservoir in water injection subsea pipeline network. The data are collected from the gauges and are fed into the model with the incorporation of reservoir properties data. Monitoring of the possibility of leakages can be done by measuring the pressure of wells. A positive pressure differential in wells represent a maintained injection pressure across the reservoir. Wells with a negative pressure differential show low injectivity which might indicate a potential leak in the subsea water injection pipeline network. A dye injection technology can be used to detect leakages and confirm the model result. Pressure data from real-time PDHMS gauges are first validated with reservoir simulation models to ensure accuracy of the input parameters. The pressure data validation is achieved by verifying each reservoir data using its properties for reliable reservoir pressure identification. Then, the GAPTM network model can be used to validate the pressure drops. The model results can enable engineers' decision-making process easier to spot leak locations where pressure decreases drastically across the subsea water pipeline. The proposed model also shows a close match with the simulated values from GAPTM network model within an accuracy of 2%. Additionally, a dye liquid is injected into the sea near to the suspected pipelines with leaks. The dye introduces a new color and thus can confirm the leak locations based on calculated values from the automated model. Pipeline leakages have been a problem for many oil fields with water injection mechanism. Pipeline leak detection is an integral part of pipeline risk management. The model can assist the engineers to locate the leaks proactively and prevent incidents from occurring time to time. The leaked water could damage environment and cause injection rate loss.