Abstract
Early detection of subsea pipeline leaks is a very serious, and still continuing, issue for the oil and gas industry with limited successful cases reported. For example, aerial surveillance of pipeline can only be applied for relatively shallow and concentrated areas, and an advanced technology such as fiber optic cable can be considered at the significant expense of time and cost for installation and equipment. The objective of this study is to evaluate a software-based leak-detection technique through the complex multiphase flow mechanics. More specifically, this study investigates (i) how leak-detection problems can be formulated from fluid-mechanics viewpoint and (ii) how reliable such a technique can be under the conditions similar to deepwater Gulf of Mexico. Examining a wide range of scenarios, this study proves that software-based techniques have potential in playing a key role in the future. This study, first of all, defines a base case selected from theliterature review of deepwater GoM flowlines in terms of pressure and temperature conditions, fluid properties, reservoir properties, and flowline characteristicswhich allows a steady-state flow in pipeline to be determined with no leak present. Next, leaks with certain opening sizes (dleak) at different longitudinal locations(xD=x/L) are positioned, and new steady states in the presence of leaks are calculated. By comparing the two steady-state responses (with and without leak), finally the changes in two leak-detection indicators (i.e., change in upstream pressure (δPin) and change in downstream total flow rate (δqt out) can be calculated in a wide range of input parameters. This study presents the results in aform of contour plots for pressure and flow responses. The major finding of this study is that, theoretically, it is possible to estimate both size and longitudinal location of the leak, by using the two leak detection indicators in the software-based leak-detection method. The results from various subsea flowline conditions (such as different GORs and fluid types, water depths, pressures at the receiving facilities, inclination angles, pipe diameters, water cuts, and so on) show that the reliability of this technique is improved when the sink term (i.e., amount of leaking fluid) is more dominant, which in turn means, leaks positioned further upstream, with larger opening size, and occurring at higher pressure inside pipe are relatively easier to detect. In many of the scenarios considered, Δqtoutas a leak detection indicator shows more than10% change in the presence of a leak with dleak>1-inch, allowing relatively easier activation of leak-warning system, which demonstrates the robustness of this technique. Other scenarios where the indicators are less than a few percent changes, however, may be challenging - in those cases, additional responses from other methods (hardware-based or transient simulation) will be helpful.
Original language | English |
---|---|
Journal | Proceedings - SPE Annual Technical Conference and Exhibition |
Volume | 0 |
DOIs | |
State | Published - 2017 |
Externally published | Yes |
Event | SPE Annual Technical Conference and Exhibition 2017 - San Antonio, United States Duration: Oct 9 2017 → Oct 11 2017 |
Funding
This study is funded by the State of Louisiana Economic Development Assistantship. We are also thankful to Schlumberger for software donation, to LSU PERTT Lab staff (Jeannette Wooden, Colton Mall, Doug Hoy, and Randall Hughes), Renato Coutinho, LSU PETE Research Assistant, and George Ohrberg, LSU College of Engineering Computer Manager.
Funders | Funder number |
---|---|
State of Louisiana Economic Development Assistantship |