Geza Udvardi
Laszlo Gerecs
KFV Oil & Gas Co.
Hungary
Yasuo Ouchi
Fumio Nagakura
Mitsubishi Heavy Industries Ltd.
Tokyo
Edward A. Thoes
Consultant
Athens, Tex.
Charles B. Wallace
Shell Oil Co.
Houston
Use of a Shell Oil Co. gas-dehydration process in a Hungarian enhanced oil recovery (EOR) project has permitted maintenance of CO2 supercritical pressures and economic gas movement rates.
Since the early 1970s, carbon dioxide (CO2) produced in Hungary's Budafa field (Fig. 1) has been used in several of the country's EOR projects. For the early projects, minimal processing was required because CO2 was injected into oil reservoirs close to the source field.
In 1988, new facilities and a 33-km (20 mile) pipeline were installed to process and transport 35 MMscfd of CO2 in the supercritical state for injection into the Nagylengyel oil field.
The longer transport distance necessitated that the CO2 be dehydrated to avoid corrosion and hydrate problems in the pipeline and injection facilities.
Initially, conventional glycol dehydration processes were considered for Hungary. These, however, would have involved pressure reduction of the CO2 before dehydration to avoid excessive glycol losses. Recompression of the CO2 to pipeline pressure would also have been required.
Instead, Shell Oil Co.'s Glycerol Drying Process, which avoids these problems, was selected.
This process minimized energy consumption by operating at supercritical pressure. The facility in Hungary has operated reliably with acceptable desiccant losses at pressures in excess of 140 bar (2,000 psi) and at gas rates 130% of design.
CO2 FIELD TESTS
Created in 1978, the KFV Oil & Gas Co. is responsible for exploration, development drilling, and production of oil and gas in an area of Hungary generally west of the Danube River (Fig. 1).
Oil and natural gas production in Hungary commenced with the discovery of Budafa field in 1937. Other notable discoveries include the Lovaszi (1940) and Nagylengyel (1951) fields.
Hungary's domestic energy supplies are insufficient to meet national demands. Therefore, maximizing oil recovery from existing fields is imperative.
Secondary recovery techniques for pressure maintenance commenced at Budafa and Lovaszi fields during the early years of production. These projects utilized gas reinjection followed by water injection.
Beginning in 1953 and continuing into the 1960s, laboratory-scale tests were performed to evaluate the use Of CO2 as an agent for increasing oil recovery in Hungary.
These tests included PVT analyses and development of physical process models. Pilot field tests began at Lovaszi in 1969.
The injected CO2 source was initially from flue gas at the central boiler battery. Later a liquefied CO2 stream was transported from another location.
During the CO2 field tests, large gas reserves with a CO2 content exceeding 80% were discovered in a deep formation at Budafa. The composition of this gas is given in Table 1.
These wells each typically produce 8.8-17.6 MMscfd at flowing wellhead pressures of 155-165 bar (2,250-2,390 psi). Successful pilot tests with the Budafa high CO2 gas confirmed that it would provide a suitable resource for EOR projects at Budafa and Lovaszi fields.
FIELD PROJECTS
A large scale CO2-injection project began at Budafa in 1972. Based on success of this project, a pipeline was constructed to transport CO2 to nearby Lovaszi where injection began in 1975. At these fields, the injected gas is partially miscible at reservoir conditions.
Various CO2 injection and oil production methods have been utilized at the two fields. Wells are generally laid out in five-spot patterns with four injectors surrounding each producer. In certain reservoirs, CO2 and water were injected alternately in a cyclic operation. In other instances, CO2 and water were injected simultaneously on a continuous basis into opposite pairs of injection wells.
CO2 injection at Budafa and Lovaszi ended in 1988; currently only water is being injected. Ultimate oil recovery from the various formations has increased by 4-15%. The recovery increase is largely a function of residual reservoir oil saturation prior to CO2 injection.
Oil produced from the Nagylengyel field is 17-24 API gravity and highly viscous. During primary recovery, no gas was produced because the oil is undersaturated at reservoir conditions.
Tests to create artificial gas caps were conducted in 1979 and 1980 with hydrocarbon and CO2 gases, respectively. CO2 was supplied from the Budafa source field by conversion of a 150-mm (6-in.) pipeline from oil transportation service. CO2 was injected in the Nagylengyel test at rates of 2.8-3.5 MMscfd.
The results of the tests through early 1989 indicated an increase in recovery factor from 42 to 49% of the original oil reserves in place.
Favorable results of the field tests at Nagylengyel provided economic justification for large-scale CO2 injection on a field-wide basis. This project began in late 1988 with artificial gas caps planned for seven large blocks in the field.
The planned injection rate of 49.5 MMscfd includes CO2 transported from Budafa as well as a small volume of produced associated gas which will be compressed and reinjected. CO2 injected into the initial blocks will be produced beginning in 1993 and then reinjected into succeeding blocks to create additional gas caps.
Reuse of the CO2 resource in this fashion should significantly improve project economics.
CO2 FACILITIES
The original processing facilities required to produce and inject CO2 at Budafa (Fig. 2) were quite simple. Produced CO2 flowed to a free-water knockout and thence to a distribution center where flow was regulated to individual injection wells.
The location of the CO2 source wells in the middle of the oil field greatly minimized problems in handling and injecting the sour, wet CO2 stream. The sufficiently high flowing pressures of the CO2 source wells permitted direct injection without compression.
The Lovaszi field is located about 11.3 km (7 miles) from Budafa. When compared to Budafa, additional water removal was necessary to minimize corrosion and hydrate problems in the transportation and injection systems for the Lovaszi project.
Processing facilities were constructed at Wells B-111 and B 500 (Fig. 3). To condense additional water, produced gas was cooled from 80 to 90 C. (176-194 F.) to 45-50 C. (113-122 F.) by a gas-gas heat exchanger and water cooler.
Water was separated from the cool gas which was then heated to 60-70 C. (140-158 F.) in the gas-gas exchanger. At full capacity these facilities could process 35 MMscfd of CO2.
The water-cooled process developed for Lovaszi eliminated hydrate problems in the transmission system except during periods when very cold weather coincided with low flow rates. In these instances, hydrates were successfully controlled by methanol injection.
Nagylengyel is located 34 km (21 miles) from Budafa. Transportation of CO2 over this greater distance requires a degree of water removal beyond the capability of the water-cooled process used at Lovaszi.
A temporary facility utilizing an expansion process (indicated by the dashed box in Fig. 3) was constructed near Well B 111 to furnish gas for the Nagylengyel field test. The process consisted of first cooling the gas in a watercooled heat exchanger.
To prevent hydrate formation, methanol was injected into the high-pressure cool gas stream which was then expanded across a valve from 120 to 130 bar (1,740-1,885 psi) to 40-45 bar (580-650 psi). The Joule-Thomson expansion reduced the temperature of the gas stream to about 4 C. (39 F.).
The relatively dry gas, now having a dew point of about + 4 C. (39 F.), was transported to the site of the pilot test at Nagylengyel where it was compressed and injected into the reservoir.
DEHYDRATION FACILITIES
An improved gas-treating process was desired for the full-scale operation at Nagylengyel to take advantage of the high source-well flowing pressures. Reduction in pressure to process the CO2 followed by recompression is thermodynamically inefficient and does not exploit the natural energy of the high-pressure CO2 reservoir.
A dehydration process using triethylene glycol (TEG) was first considered. TEG systems are very common for dehydrating natural gas, and a number of these units operate in CO2 service around the world.
However, most of the CO2 dehydrators operate at pressures well below the 74 bar (1,070 psi) critical pressure of CO2- A problem with glycol units in CO2 service at pressures near and exceeding critical is the high mutual solubility between CO2 and glycol which results at these conditions. Excessive glycol losses into the dry CO2 stream occur which results in high desiccant makeup costs.
The process selected for Nagylengyel utilizes CO2 dehydration technology developed and licensed by Shell Oil Co.
The patented Shell Glycerol Drying Process (SGDP) had been used previously in commercial units to dehydrate supercritical CO2 in Shell's operations at McElmo Dome in Colorado and Jackson Dome in Mississippi.
In contrast to glycol systems, SGDP is capable of operating at very high pressures with only minimal losses of the glycerol desiccant.
The Hungarian facility, which began operation in late 1988, is designed to decrease the dew point of the produced CO2 to - 5 C. (23 F.) at pipeline pressure.
A general scheme of current CO2 facilities in Hungary is illustrated in Fig. 4.
Rated capacity of the Hungarian SGDP is 35 MMscfd. The original facility design provided for future expansion by the addition of a parallel unit of equal capacity. Utilities serving the plant include electrical supply, fuel gas, boiler plant, and instrument air systems.
Fig. 5 depicts a block diagram of the dehydration plant and related facilities. A simplified process flow drawing is illustrated as Fig. 6. CO2 is countercurrently contacted with concentrated glycerol in a bubble-tray contactor which operates at pressures of 144.2-152.4 bar (2,090-2,210 psi).
Rich glycerol solution is heated, flashed, further heated, and then reconcentrated in a direct-fired reboiler/still which operates near atmospheric pressure. The hot regenerated glycerol solution is cooled by heat exchange and then pumped and recirculated to the contactor.
CO2 lines, corrosion prevention Although the early projects at Budafa and LovAszi included only minimal CO2 processing, the transportation of the gas over the short distance was relatively troublefree.
CO2 was transported at subcritical pressure during the Nagylengyel field test because the converted oil pipeline was rated for only 64 bar (928 psi).
Detailed hydraulic and thermodynamic studies for the full-scale Nagylengyel operation indicated CO2 transport at supercritical pressure to be more economic.
A new, 300-mm (12-in.) line was installed to operate at pressures as high as 160 bar (2,320 psi).
In normal operation, the pipeline does not require heating.
An intermediate heating facility on the pipeline provided capability for heating during start-up.
The pipeline was constructed to allow for pigging and was provided with quick-operating isolation valves.
Corrosion prevention and monitoring are critical considering the operating conditions and the composition (i.e., very high pressure, wet, sour, high CO2) of the fluids handled.
Monitoring is achieved by corrosion coupons installed in pipeline systems, from iron counts of fluid samples taken at various points in the system, and from wall thickness measurements on flow lines and vessels.
Inhibitors to protect CO2-producing wells, flow lines, and production manifolds from corrosion are periodically injected downhole with squeeze techniques. These filming inhibitors have provided satisfactory protection when the system is operated within certain fluid velocity ranges.
The dehydration facility placed in operation in 1988 has essentially eliminated corrosion in the CO2 transportation and injection systems.
The component parts of the dehydration system which contact raw, wet CO2 gas are of corrosion-resistant materials.
ADMINISTRATION, FACILITY CONSTRUCTION
Process design of the dehydration unit was by Shell. Mitsubishi Heavy Industries (MHI), Tokyo, performed detailed engineering design under Shell's license. MHI also provided equipment procurement services.
Field construction totalling 50,000 man-hr commenced in May 1988 and concluded accident-free during October 1988.
Significant differences between the design conditions for the Hungarian project and Shell's prior U.S. operations included the much higher operating pressure and the presence of almost 20% non-CO2 constituents in the feed gas (which included 3,000 ppm H2S).
It was originally recognized that diluent gases could greatly reduce the water content of high-pressure CO2 streams as compared to pure CO2.
Unfortunately, water solubility data for supercritical CO2/hydrocarbon mixtures approximating the Hungarian gas composition were unavailable in the literature.
To confirm the Hungarian design basis, laboratory water content and dew-point measurements were obtained by MHI's research institution.
Calibration of the dew point analyzer computer was based on the lab data.
Physical and thermal properties of the gas mixture were estimated with the Peng-Robinson equation of state. Subsequent operating experience validated the accuracy of the design basis.
The materials utilized in the dehydration process were based on Shell's experience at Jackson Dome and other fields.
Potentially severe corrosion was possible in plant inlet facilities upstream of the dehydration process. This equipment would be exposed to high pressure, hot (90 C./194 F.), sour CO2 as well as low pH formation water containing chlorides.
Mitsubishi's research organization examined the compatibility of materials for these systems.
Major process equipment including pressure vessels, towers, reboiler, heat exchangers, filters, rotating equipment, etc. were fabricated in Japan, Europe, and the U.S. with inspection services and quality control provided by MHI.
All equipment was shipped by the end of April 1988, 12 months after contract award.
INITIAL OPERATIONS
Early in the project's design phase, Hungarian and MHI engineers visited Shell's Jackson Dome field in Mississippi where supercritical CO2 is dehydrated with glycerol.
Shell's start-up consultant arrived at the plant site in Hungary in late September 1988 when mechanical construction was complete, and the contractor began installing insulation.
Three activities were involved in the plant start-up and initial operation: Equipment commissioning, a 30-day test run to measure glycerol losses, and a 72-hr test run to measure dehydrated CO2 water dew point and plant utilities consumption. The 30-day and 72-hr runs were required to satisfy contract guarantees.
A very early winter with -17 C./1 F. temperatures experienced on Nov. 10, 1988, created difficulties during the initial operations. At this time all thermal insulation work had not been completed and several problems resulted from level gauge bridles and controller float cages being either exposed to weather or inadequately heat traced.
Differences in temperature between the pressure vessels and related instruments resulted in false levels being detected in several instances.
The false signals in turn resulted in the separators being filled to extremely high levels or being completely emptied.
These problems were aggravated by the sensitivity of supercritical CO2 density to small changes in temperature.
During the initial operation performance of the glycerol reboiler was determined to be inadequate because of excess primary combustion air.
The problem was successfully resolved by the Italian supplier.
After three weeks of operation, failures resulting from corrosion occurred in the plungers of produced water reinjection pumps. Peroni, the pump supplier, provided replacement chromium-plated plungers.
There have been no further failures.
Excessive H2S and CO2 dissolved in water from the inlet separator flashed off at low pressure in a waste water-surge vessel (V-200). This problem was corrected by installation of a gas-recovery scrubber upstream of V-200 which operated at intermediate pressure.
Environmental regulations dictated that water stripped from the rich glycerol in the reboiler be piped through a closed system to a vent stack located 350 m (1,150 ft) away. This was due to the high H2S content of the vent stream.
Condensation in the vent line and stack increased the back pressure on the reboiler.
The higher operating pressure necessitated higher-than-design reboiler temperatures to strip adequately the glycerol.
With the more serious of these problems corrected, the plant throughput was then increased.
At a throughput of 42 MMscfd, or 120% of design, the plant was capable of dehydrating the CO2 to a dew point below the guarantee value, - 5 C. (23 F.). The test operations were successfully concluded on Dec. 28, 1988.
RECENT EXPERIENCES
Operation of the glycerol dehydration unit to date has been satisfactory. The required dew-point reduction is readily accomplished, and continuous and stable transportation Of CO2 to Nagylengyel has been provided.
Consumption of glycerol and utilities has been less than the process performance guarantees.
The excessive back pressure on the glycerol reboiler, mentioned previously, was corrected by the vapor to the atmosphere being vented directly above the reboiler. In order to reduce H2S content of the vent stream, it was necessary to reduce the operating pressure of the rich glycerol flash drum (V-310). This resulted in more H2S being flashed off ahead of the reboiler.
Level control bridles and level switches in supercritical CO2 service were temperature controlled at the process temperature with electric heat tracing.
This has eliminated the level problems experienced during start-up.
An alternative outlet for waste water was provided to enable transferring this stream to the nearest oil-gathering station whenever maintenance is performed on waste water-disposal pumps.
Glycerol pH has been adjusted twice by the addition of diethanolamine. Chloride content of the system has been maintained at an acceptably low level (70 ppm maximum).
During the summer of 1989, the inlet gas air coolers were overloaded due to high gas flow rates and to the gas being 6-8 C. (11-14 F.) warmer than design. During these periods the gas entering the dehydration process was 2-3 C. (4-5 F.) warmer than design.
Even so, no deterioration in water dew point was observed.
In September 1989, a 3-week capacity test was conducted at the plant. The process was capable of providing satisfactory dehydration (i.e., low dew point and glycerol losses) at a feed rate of 46 MMscfd which is 130% of rated capacity.
Following these modifications, the plant has operated at the planned capacity without shutdowns.
Copyright 1990 Oil & Gas Journal. All Rights Reserved.
