Energy Policy Act of 2005 (Ultra-deepwater and Unconventional Resources Program)
Flow Phenomena in Jumpers- Relation to Hydrate Plugging Risk
The objective of this project is to perform transient flow experiments in pipeline flow geometries typical of the low spots encountered in subsea flowline jumper systems. This project aims to improve the industry understanding of liquid displacement and flow patterns in jumper-like systems during production restart operations.
The University of Tulsa, Tulsa, OK 74104-3189
In deepwater and ultra-deepwater systems, hydrate formation and plugging is an important concern because of the difficulty and cost in remediating hydrate plugs and the associated deferral of production revenue. Design solutions such as flow line insulation and inhibitor injection (e.g., methanol) are the standard engineering methods deployed to avoid hydrate formation and plugging. Restart scenarios and flow profiles are evaluated using state-of-the-art transient flow models, yet despite very conservative standards and operating strategies, plug formation is still not completely avoided.
In deepwater systems, jumpers are used to connect subsea trees to manifolds or to connect flow lines together. The production jumper elements of a subsea production system seem to be at a higher risk of hydrate plugging during restart operations, in part because of their geometry, because of the difficulty in insulating such geometries, and also because of a probable misunderstanding of the complex flow patterns and phenomena taking place in the jumper during restart. Once a plug is formed in a jumper, current jumper designs make it difficult to remediate the plugs, leading to very large remediation costs.
This project will utilize the expertise and infrastructure available at the University of Tulsa Hydrate Research Project to improve the understanding of liquid displacement and flow pattern in jumper-like systems during restart operations. Previous research at the University of Tulsa has shown the importance of the presence of a free-water phase and its displacement on the plugging tendency of a system. This project will carry out transient flow experiments in geometries typical of the low spots encountered in jumper systems. The liquid displacement, and more importantly the water displacement, will be measured as a function of operating parameters such as water cut, liquid loading, restart rates, and oil viscosity. The measured water displacement, liquid carry-over from the low spot, and flow patterns taking place during the restart process will be compared to state-of-the-art transient simulation results. The effects of liquid loadings, water loadings, and restart rates on the displacement of the water phase will be studied.
Deliverables for this project will include a series of reports on the various tasks as they are completed and a final report integrating the results of the project and providing guidelines for its application elsewhere. Other deliverables will include: a) a database indicating liquid (oil and water) displacement, flow patterns and liquid carry-over from the low spot for different operating conditions, b) comparisons of the results with existing simulators to identify areas for improvement and the range of operating conditions where current models perform satisfactorily, c) an estimation of the risk of hydrate plug formation based on data and visual observations for different operating conditions, and d) potential new design criteria to be used in the development of safer restart strategies. The work will also result in the comprehensive training of a graduate student on the physical aspects of fluid flow and hydrate plug formation and dissociation.
The outcome of this project will be an improved confidence in the performance of transient flow simulators when applied to jumper geometries and/or the identification of their shortcomings. An accurate prediction of the water phase location and behavior during restart conditions is necessary to properly simulate hydrate formation and evaluate the risk of plug formation in jumper systems. Data collected from this project may lead to better prevention methods, such as better methods to displace water out of a non-inhibited jumper while avoiding plug formation. Information on inhibitor distribution and displacement could lead to better design of injection points in jumpers.
The research contract for this project was finalized in September, 2008 and work was carried out in five primary phases: Facility simulation and design (3 months), Facility construction and shake-down (6 months), Experimental studies, including both gas-dominated and liquid-dominated restarts (4 months), Data processing and model comparisons (3 months), and Reporting (2 months).
The design of the facility was completed and the facility was constructed. A jumper geometry model (incorporating two low spots with different length-to-height ratios) was constructed of 3 inch acrylic pipe and is now being used to study the displacement of water and oil into the system. At the beginning of each experiment, each low spot was charged with a known amount of oil and water, providing different bridging conditions of the pipe. After the fluids are allowed to segregate, a restart phase is applied at the section inlet, resulting in the displacement of the liquids accumulated in the low spot. The combination of restart rate, restart fluid properties (liquid or gas, liquid viscosity and density) and bridging conditions results in a variety of observed flow patterns and transitions, as well as different amounts of accumulated fluids being carried-out of the jumper section. The experiments provide a classification of these flow patterns as well as an estimate of the critical velocity at which most of the low spot fluids are swept out the low spot.
Fifty (50) transient gas-water tests in the jumper were conducted. Final liquid hold-up was determined to be a function of velocity. Simulations using OLGA (an industry standard multiphase flow simulator) were found to significantly over predict hold-up.
Fifty-four (54) gas restart experiments with single phase model oils were conducted. Gas restart data in a bridged configuration for water and two model oils were compared. As expected, the liquid hold-up is higher (14 to 33%) for the 220 cp model oil than water. For velocities less than 10 ft/sec the 19 cp oil behaved as expected, that is, exhibiting hold-ups between water and the 220 cp oil (17% more than water). However, at higher velocities an unexpected result was seen where the displacement was greater than that seen for water (6%). This trend was seen for all liquid loadings studied. The main variable affecting the water removal is the restart velocity; the higher the restart velocity the greater the amount of water removed. The amount of water remaining was also found to decrease with an increase in jumper volumes used. Another variable that was tested was the water cut, which shows higher water holdups after restart for higher initial water cuts. Even though the trends show that higher velocities and the displacement with additional jumper volumes at restart are more efficient, using the lowest velocity (~0.5 ft/s) for one jumper volume removed more than 90% of the water, which means the jumper will be non-bridged with less than 20% liquid in the low spot. Two main characteristics were also observed as the fluids were displaced during restart: 1) the liquids already in the jumper were “pushed out” by the incoming liquid, and 2) the water tends to accumulate at the lower riser elbows.
Seventy-four (74) gas restarts with oil-water mixtures were conducted. As with the single phase tests, OLGA simulations over predicted the liquid hold up. For the single phase experiments the liquid hold up was affected by velocity while for the two phase displacements the liquid hold up was affected by both velocity and water cut. The results from these tests also showed that not only the viscosity but also the density has an impact on the liquid carry over.
For the liquid restart experiments, the 19 cp oil was found to be more efficient than kerosene removing all the water from the jumper. Also, it was harder to remove all of the water in the jumper at higher water cuts. Even a jumper full of water flushed with one jumper volume using the lowest velocity tested (0.48 ft/s) removes at least 70% of the water from the jumper. Higher liquid restart rates were more effective pushing all the water out of the jumper, but lower restart rates promoted more mixing. Similarly, the liquid restart simulations over predict the water left. Simulations show no water left in the first low spot and less than 5% in the second low spot. By increasing the restart velocity or the number of jumper volumes displaced, the final water holdup was also decreased, so once again, the simulations matched the behavior observed in the experiments as with the gas restart.
This project is complete. The final report is available below under "Additional Information".
Project Start: September 22, 2008
Project End: January 21, 2010
DOE Contribution: $120,000
Performer Contribution: $30,797
RPSEA – Jim Chitwood (email@example.com or 713-372-2820)
NETL - Jay Jikich (Sinisha.Jikich@netl.doe.gov or 304-285-4320)
Performer Company – Dr. Michael Volk, Jr. (firstname.lastname@example.org or 918-631-5127)
Final Project Report [PDF] January, 2010