The Kitan oil field consists of three subsea intelligent wells. The intelligent completions were modeled in detail using commercial dynamic-simulation software to establish a sound and safe operating procedure for the well cleanup and well test. The simulation results provided the petroleum engineer on the rig with key operational information, such as the time for the oil to arrive at surface and expected pressure at the downhole gauges (DHGs) and upstream of the choke manifold. This enhanced the ability of rigsite supervisors to anticipate well behavior, enabling a significant risk reduction.
Introduction
The Kitan oil field is approximately 200 km south of the East Timor coast and 500 km offshore of the northern coast of Australia (Fig. 1). The Kitan field development consists of three vertical producing wells, 6-in. production subsea flowlines and risers, 2-in. gas-injection lines, umbilical lines, and a floating production, storage, and offloading (FPSO) vessel. Each well has a designated flowline and riser tied back to the FPSO vessel (Fig. 2). Fig. 3 shows the intelligent completion for the Kitan development wells. The intelligent completion was designed to control flow from two separate zones, upper and lower. Flow from each zone is controlled by flow-control valves (FCVs) that have eight positions (fully opened, fully closed, and six intermediate choke positions).
Pressure and temperature for each zone are monitored by three DHGs. The FCVs have a fail-as-is design and stay in the last position should a communications problem be encountered.
Because of the remoteness of the location, the functionality of the intelligent completion and the well performance needed to be confirmed with a minimum degree of uncertainty. To establish a sound and safe operating procedure for the well cleanup and well test, commercial dynamic simulation software was used to model the intelligent completion and selected operations. A few different well-cleanup procedures were investigated to predict cleanup time, pressure and temperature at different points of interest, and flow rates on the rig. The simulation results provided valuable information, enabling the petroleum engineer on the rig to predict flow conditions of the well during the operations. Following the successful well-cleanup and well-test campaign, the well models were validated with actual well-cleanup and -test data.
Well-Cleanup and Well-Test Simulation
Operating Sequence. The high-level operating sequence for the well cleanup and well test was as follows:
- Place perforation fluid in the perforation interval.
- Perforate the upper and lower zone with tubing-conveyed perforation on drillpipe.
- Set lower packer with wireline.
- Run completion and stab into the lower packer.
- Pump the base oil to displace the brine with the upper FCV opened.
- Set upper packer.
- Conduct well cleanup with both FCVs opened (approximately 4 hours).
- Rupture shear bar in the gas lift orifice valve (GLOV) during the well cleanup.
- Close the lower FCV to build up the lower zone and flow from the upper zone (approximately 4 hours).
- Close the upper FCV to build up the upper zone.
- Open the lower FCV to flow from the lower zone (approximately 4 hours).
- Open the upper FCV to flow from the upper and lower zones (approximately 4 hours).
- Close the choke manifold.
- Close surface-controlled subsurface safety valve (SCSSV) for leakoff test.
- End of well cleanup and well test.
Cushion Fluids and Displacing Procedure. Before conducting the cleanup simulation, the advantages and disadvantages of using different cushion fluids and different displacing procedures were investigated, as follows:
- Case 1: Displace the base oil from the annulus to the tubing with the GLOV after setting the upper packer.
- Case 2: Displace nitrogen from the annulus to the tubing with the GLOV after setting the upper packer.
- Case 3: Displace the base oil from the tubing to the annulus with the upper FCV opened.
- Case 4: Displace the nitrogen from the tubing to the annulus with the upper FCV opened.
After conducting a series of simulations to compare and evaluate advantages and disadvantages, Case 3 was selected.
Predictive-Modeling Results. Well cleanup would be complete when the amount (mass%) of the muds (brine and base oil) in the produced reservoir oil became less than 1% at the surface in the simulation. This also ensured that the majority of the mud was removed from the wellbore. In reality, well cleanup is judged by many factors (e.g., basic sediment and water, pressures, and temperatures). The simulation results demonstrate that unloading time decreases with increasing oil rate, with reduced oil-rate sensitivity above 7,000 STB/D. Subsequently, 7,000 STB/D was considered to be the optimum oil rate from an unloading perspective. The maximum oil-flow rate of 7,000 STB/D was used for thermal radiation simulations and investigation of multiphase-flow velocity in the piping on the rig.
The choke was opened to 0.5 in. On the basis of the results, the brine was expected to arrive at the surface in approximately 1.5 hours. The pressure increased in relation to the displacement of the brine from the well. The reservoir oil arrived at the surface in approximately 2 hours. Once the reservoir oil arrived at the surface, the brine rate decreased and the oil rate increased.
The estimated oil rate at the end of the well cleanup was 6,700 STB/D. The oil rate decreased to 6,300 STB/D when the lower FCV was closed with the choke unchanged. Oil-flow rate was expected to return to 6,800 STB/D when the lower FCV was opened and the upper FCV was closed. Finally, the oil rate increased to 6,900 STB/D when both FCVs were opened. The estimated flowing wellhead pressure ranged from 1,200 to 1,400 psia.
Predictive-Modeling Summary. Pressure and temperature were predicted dynamically at different locations of interest, such as at DHGs and upstream of the choke manifold. The estimated pressures and temperatures were used as guidelines to assess well performance. These estimations assisted the petroleum engineer in providing relevant information to a testing service company far in advance, to enable them to prepare for the most-probable required actions.
Well-Models Validation With Well-Cleanup and Well-Test Data
The objective of this study was to validate three development-well models by conducting history matching to the corresponding actual well-cleanup and well-test data. Once the history matching was conducted successfully, the validated well models were integrated with flowline models to investigate the initial well-startup operations.
Simulation Workflow—History Matching. History matching was a trial-and-error process using different reservoir permeabilities, skin factors, and reservoir pressures to achieve pressure matching at key locations of interest and at desired flow rates.
Deviation From the Original Cleanup Study. Pressure/volume/temperature (PVT) data were updated using the newly available PVT data from the actual sample taken during the commingled-production period. The opening schedules of the choke and FCVs were updated on the basis of the actual schedule.
Post-Test Matching of Modeling Results. The brine- and oil-arrival times at surface predicted by the model were approximately 10 minutes faster than the actual time.
History matching was prioritized to the last commingled-flow period and shut-in period because the commingled sample was used to prepare the PVT table. The commingled-production period represents actual flow conditions until the lower FCV has been closed to shut off water production from the lower zone. Good history matching was achieved at all three DHGs with a maximum error of approximately 10 psi (less than 1% error) during cleanup.
Post-Test Matching Modeling Summary. In all three development wells, the well models were validated successfully against the actual well-cleanup and well-test data. Matching was achieved at all three DHGs, with a maximum error of less than 1%. Similarly, matching was achieved upstream of the choke manifold, with a maximum error of 1% during the commingled-flow period. The well models were validated and then integrated with flowline models for further application.
Initial-Well-Startup Simulation Using the Validated Models
The objective of this study was to estimate the optimum FCV positions to achieve a plateau rate of 40,000 STB/D without exceeding the system limitations.
Recommended FCV Positions. Well performance for steady-state conditions was investigated for different FCV positions using the validated well/flowline integrated model. The FPSO-vessel arrival pressure upstream of the topside choke was varied to predict flow rate at the FPSO vessel and pressures and temperatures at different locations of interest.
Operating Sequences. The following high-level sequence was established to optimize the startup operations:
- Cycle the upper FCV from fully closed to Position 1.
- Open the topside choke with minimum available step (approximately 8.5% stroke).
- Receive oil at the FPSO vessel.
- Cycle the lower FCV from fully closed to Position 1.
- Cycle the upper FCV from Position 1 to Position 2.
- Cycle the lower FCV from Position 1 to Position 2.
- Cycle the upper FCV from Position 2 to Position 3.
- Cycle the lower FCV from Position 2 to Position 3.
- Increase the topside choke to the next step (9% stroke).
- Cycle the upper FCV from Position 3 to Position 4.
- Cycle the lower FCV from Position 3 to Position 4.
- If target rate cannot be achieved, then increase topside choke.
- Monitor flowing pressure and flow rate.
- Increase topside choke until target rate is achieved.
Startup-Simulation Results. The reservoir oil was predicted to arrive at the FPSO vessel in approximately 8 hours. The simulation starts at 50 hours. The pressure upstream of the topside choke and the flow rate increased in relation to the FCV’s opening schedule. Once the FCVs were cycled to Position 3, the arrival pressure at the FPSO vessel was approximately 1,000 psia. At this stage, the topside choke was increased to 9% to reduce the arrival pressure. The temperature was predicted to continue increasing with producing time and to stabilize within 40 hours of opening the well. The flowing pressure and temperature upstream of the topside choke were predicted to be 830 psia and 100°C, respectively, when the flow rate was 14,000 STB/D (flow-rate target for Kitan 5).
Estimations were that a 60-psi drawdown produces 6,700 STB/D of oil from the lower zone and an 80-psi drawdown produces 7,200 STB/D of oil from the upper zone. The flowing bottomhole temperature was estimated to be 138°C.
Predictions were that flowing pressure at the subsea tree (SST) decreased when the oil reached the riser because the hydrostatic pressure in the riser was reduced with a constant arrival pressure at the FPSO vessel (at approximately Hour 57). This was observed during the actual operations and was a good indicator of flow behavior in the production system. It was predicted that the flowing pressure and flowing temperature at the SST was 1,250 psia and 122°C, respectively, when the flow rate was approximately 14,000 STB/D. The expected pressure and temperature values at the SST were less than the design operating pressure and temperature.
Startup-Simulation Summary. The actual startup operation generally followed the predictions of the simulation, with the exception of a few unplanned shutdowns caused by the process plant on the FPSO vessel.
The petroleum engineer on the FPSO vessel monitored flow parameters during the operations and was able to predict flow conditions in the production system, enabling him to instruct when FCVs and the topside choke should be adjusted with good confidence. This assisted in overcoming the flowline and topside-choke design limitations and in optimizing the startup operations.
This article, written by Editorial Manager Adam Wilson, contains highlights of paper SPE 156146, “Dynamic-Simulation Applications To Support Challenging Offshore Operations: A Kitan Oil Field Offshore East Timor Case Study,” Ryosuke Yokote, SPE, and Vanni Donagemma, Eni Australia, and Juan Carlos Mantecon, SPE, SPT Group, prepared for the 2012 SPE Annual Technical Conference and Exhibition, San Antonio, Texas, 8–10 October. The paper has not been peer reviewed.