J.E. Johnson
Pritchard Corp.
Overland Park, Kan.
Stephen J. Tzap, Robert E. Kelley
Raytheon Engineers & Constructors
Denver
Lawrence P. Laczko
OXY USA Inc.
Midland Tex.
A design study for a new West Texas gas processing plant for a CO2 EOR project provides for installation of H2S removal processes to be delayed for 3 years after completion of the plant.
During this delay, a more precise produced gas composition will be obtained so that the process equipment for removing H2S can be properly selected and sized to handle the gas stream that at the peak will reach about 30 MMscfd from the initial 4 MMscfd.
Before the installation of the H2S removal processes, OXY USA Inc., the project operator, plans to reinject the sour gas into a specific area of the field.
The new plant's processing components include inlet separation, sulfur removal and recovery, compression, dehydration, and NGL recovery.
The existing Welch gas processing facility in Dawson County, Tex., will be replaced by the new plant, currently under construction and scheduled to be completed in the first quarter of 1994.
The new plant will be capable of processing CO2-contaminated associated gas, recovering valuable propane-plus NGLs, and producing a miscible CO2 for reinjection.
This first in a series of two articles details the process and configuration options. The concluding part will discuss in greater detail the sulfur recovery alternatives.
WELCH PLANT
The existing Welch plant was constructed to facilitate oil production from surrounding fields. The gas processing facility includes an inlet amine treating system and a 2 long ton/day sulfur recovery unit, followed by a 3 MMscfd straight refrigeration plant.
Westar Marketing Co. takes the residue gas stream while the liquid product goes to Mapco Gas Products Inc. Molten sulfur is stored in an underground sulfur storage pit and trucked out for sales. The facility processes gas from North Welch, South Welch, West Welch, Cedar Lake, and Terry County gas fields.
Welch area oil fields are amenable to CO2 flooding. The CO2 will extend the life of the fields and improve ultimate oil recovery.
During the CO2 flood, the initial volume of CO2-rich associated gas will be about 4 MMscfd and the flow will increase over several years to more than 30 MMscfd as CO2 breakthrough increases.
OXY required that a staged investment project be defined using parallel process systems wherever practical and economically attractive. The CO2/hydrocarbon recovery portion of the facility was to be modularized as much as possible to allow future expansion of processing capability, from an initial 15 MMscfd facility, to that required to process 30 + MMscfd.
It is expected that the primary expansion will occur in the fourth year after start-up with expansion of the CO2/hydrocarbon recovery unit.
The new CO2 facility also include compression of the low-pressure associated gas, ultimately to 1,700 psig for reinjection in the oil held. At appropriate intermediate compression stages the gas must be dried to avoid corrosion and freezing or hydrate problems downstream, and propane-plus NGLs are to be recovered for sale.
OXY awarded an engineering, procurement, and construction (EPC) contract for the project to Raytheon Engineers & Constructors Inc. (RE&C) in April 1992. The process selection study was completed in May 1992 and after completion of process and detail design, field construction began in June 1993. Start-up is scheduled for the first quarter of 1994.
PROCESS SELECTION
The process selection study defined the most cost effective, yet operationally sound process configuration and project execution plan, in conformance with OXY's best outline for anticipated plant expansion and eventual project phaseout as EOR operations cease.
RE&C established a recommendation for specific technology to be applied to the major required process steps:
- Inlet separation
- Sulfur removal and recovery
- Compression
- Dehydration
- NGL recovery.
After considering numerous options, the recommended overall plant process configuration combined the selected technologies in an optimum manner.
The Drizo process was selected for the gas dehydration process, primarily to ensure ability to achieve low water dew point and minimize benzene, toluene, ethyl benzene, and xylene (BTEX) emissions.
For propane-plus NGL recovery, a membrane/refrigerated stabilizer process was chosen because economics proved slightly better than the Ryan/Holmes process. Also, the membrane-based system offered greater flexibility in regard to staging the investment to more closely match expected changes in throughput and processing requirements.
The project has proceeded on the basis of initially installing only a refrigerated stabilizer. In about the fourth year of the project, an upstream membrane skid(s) will be installed to allow processing higher gas volumes containing more CO2.
Additional membrane units could be added if volumes increase, or if OXY processes third party gas.
Compared to the Redox-type process, two-stage selective amine/Claus sulfur recovery was found to provide better economics over a significant portion of the gas volume/sulfur content operating matrix. However, OXY has delayed final decision on the process and unit capacity until the EPC deadline to achieve sweet CO2-rich reinjectant approaches.
PLANT CAPACITY
Some uncertainty remains as to the CO2-rich associated gas volumes and compositions produced over the project's life. OXY also desired the flexibility to process as yet unidentified third party gas.
From this somewhat uncertain data base, however, OXY provided a valid design basis (see box) that allowed RE&C to reach decisions regarding technology and overall configuration selections over a broad capacity/composition matrix.
OXY defined both low and high-volume cases to bracket the flow and CO2/NGL compositions expected by OXY's reservoir staff.
The initial tables provided by OXY were then modified by adding H2S and normalizing the resulting compositions for each case.
Three sulfur levels were established for each of the two volume scenarios: 9, 18, and 27 long tons/day.
The basic differences between the base (9 long tons/day) low and high-volume cases are Summarized in Table 1.
PROCESS AND CONFIGURATION
A staged design was used in evaluating the various processes applicable to individual systems, plus possible overall plant configuration options with these processes.
The analysis included consideration of each option or process technology with regard to delaying unnecessary or unreasonable equipment installations (i.e., capital expenditures) as allowed by predicted gas volumes and compositions.
Table 2 and Fig. 1 identify some of the primary process technologies and overall process options considered during the study.
Because this article is limited to sulfur recovery aspects, all options relating to dehydration and NGL recovery are not discussed. Included are only those that are important when considering future sulfur recovery.
RYAN/HOLMES
Fig. 1a illustrates Options 1, 2, and 3, all of which use the Ryan/Holmes technology for NGL recovery.
Option 1 includes operations for about the first 3 years of plant life. During this period, a sulfur recovery unit would not be installed, thus the entire process train would operate sour and the reinjected CO2 would also be sour.
Produced gas would be compressed in two stages to about 375 psig before being dehydrated and processed in a two-column Ryan/Holmes unit for NGL recovery. The produced high-CO2 stream would then be compressed in the final two-compression steps to 1,700 psig for reinjection. The recovered sour NGL would be treated and sold.
Option 2 has a subsequent sweet operation beginning in the fourth year. The choices for sulfur recovery are Redox, a selective amine/Claus, or Selectox-type. Sulfur recovery would be installed on the suction of the first compression stage (3090 psig). Dehydration and NGL recovery technologies would be unchanged, although now operating sweet. The recovered NGL would require less treating to meet sales specifications.
Option 3 is the same as Option 2 except for the assumption that the Redox or selective amine sulfur recovery would not be on the low-pressure produced gas stream. Instead, the various systems through NGL recovery would continue to operate sour.
The Ryan/Holmes process can be altered to also remove all H2S from the high-CO2 reinjectant stream, recovering the H2S as part of the produced NGL. The sweet reinjection stream would be compressed in the final two stages to 1,700 psig.
Option 3 includes a nonselective amine treater on the sour NGL to remove nearly all of the H2S and CO2. A final product treating system is still needed to remove other trace sulfur compounds. The acid gas (H2S + CO2) stream regenerated from the amine is fed to a Claus SRU for sulfur recovery.
The operating parameters and utility requirements for Option 3 are substantially higher than for Option 2. For future flexibility at reasonable cost, these differences must be considered in the initial plant design.
STABILIZER
For NGL recovery, Fig. 1b illustrates Options 4 and 5 that substitute a membrane/refrigerated stabilizer in place of the Ryan/Holmes process.
Both Options 4 and 5 allow initial operation with the stabilizer only (no membrane unit). As gas volume and CO2 increase, a recycle compressor and the membrane unit would be added.
Option 4, like Option 1, operates for the first 3 years in a sour mode. The low-pressure produced gas would be compressed in two stages to about 375 psig. Then after being dehydrated and compressed in a third stage to about 750 psig, the stream would enter the membrane system.
Most of the H2S and CO2 would permeate through the membranes into a lower pressure system at 350 psig. The nonpermeate stream would be processed in a refrigerated stabilizer for NGL recovery.
The stabilizer overhead would be combined with the 350 psig permeate. This combined stream, high in CO2, would be compressed to 1,700 psig and recycled to the injection wells.
The NGL recovered from the stabilizer would require treating prior to sales.
During the initial years of operation without the membrane unit, the stabilizer would be fed directly from the Drizo dehydration unit at about 350 psig. The CO2-rich overhead gas from the stabilizer would then proceed through the third and fourth stages of the compression system. The 1,700 psig stream would then be recycled to the injection wells.
Option 5 represents subsequent sweet operation, beginning in 1997. Again the choices for sulfur recovery are Redox selective amine/Claus, or Selectox-type installed on the first stage of compression (30-90 psig). Dehydration and NGL recovery technologies would be unchanged, though now operating sweet. NGL product treating would still be required to meet sales specifications.
RECOMMENDED PROCESSES
A simplified process flow diagram for the Welch plant CO2 project, Fig. 2, illustrates the optimum overall CO2 project configuration.
Based on capital, operating, and maintenance cost estimates, sulfur recovery is most economical at the lowest available pressure and upstream of dehydration and NGL recovery. This is true for any of the sulfur recovery processes chosen, low or high gas volumes, and the entire range of sulfur quantity. Thus, sulfur recovery is placed upstream of the first compression stage, 30-90 psig.
Dehydration is most economical at the second stage discharge, about 375 psig, and is located upstream of the NGL recovery processes.
Primarily because of a delay in a significant portion of the total capital cost, membranes with a refrigerated stabilizer system are recommended for the NGL recovery process. The membranes are used for bulk CO2 removal with the final separation of CO2 and hydrocarbons occurring in the refrigerated stabilizer system.
The feed hydrocarbon-rich CO2 stream passes through a single-stage membrane unit at 750 psig. The permeate (CO2) stream leaves the membranes at 350 psig and mixes with the stabilizer overhead gas. The mixture forms the miscible CO2-rich stream which is then compressed to 1,700 psig for reinjection into the oil field.
The nonpermeate (hydrocarbon) stream leaves the membranes at 750 psig and is reduced in pressure to 350 psig before entering the refrigerated stabilizer system. NGL produced from the stabilizer is treated in a molecular sieve unit to remove residual levels of H2S, CO2, and other contaminants.
INLET SEPARATION
For inlet separation, a conventional stabilizer sweetens condensed hydrocarbon liquids from various compression steps and produces an 8-12 Rvp product.
CO2 REINJECTION
Reciprocating compressors best fit the Welch plant application. The basic pressure levels of the compression stages are reasonably fixed by the thermodynamics of the gas and general compressor mechanical limitations.
The approximate levels are as follows:
- Suction-80 psig (range 30-90 psig)
- First stage-175 psig
- Second stage-375 psig
- Third stage-750 psig
- Fourth stage-1,700 psig.
Based on the low and high gas volume cases over the plant life and on discussion with various compressor manufacturers, a staged investment plan was developed for both low and high volumes using combinations of 1,000, 1,500, or 2,000 hp skidded compressor units.
SULFUR RECOVERY
In the low-volume cases for the Welch plant, a two-stage, selective-amine gas treating process (Fig. 3) with Claus sulfur recovery shows optimum economics above about 8 long tons/day.
This process becomes significantly more economical as tons of sulfur (associated gas H2S concentration and volume) increases to more than 8 long tons/day.
In the high-volume cases, the Redox process shows optimum economics at less than about 15 long tons/day.
Although the economics favor Redox over much of the operating matrix, the Redox processes are relatively new and, in some cases, have suffered significant operating and maintenance problems.
The amine/Claus process is well proven and provides high reliability.
OXY chose to consider the amine/Claus process in both the low and high-volume cases but will re-evaluate this decision later in the project life when sulfur recovery is actually needed. This allows more industry experience to be gained for the Redox processes and possibly further technology refinements.
DEHYDRATION
The Drizo process offers significant advantages for the Welch plant, particularly in controlling BTEX emissions.
NGL RECOVERY
The integrated process of a membrane system followed by a refrigerated stabilizer is the most cost-effective design for handling the anticipated gas volume range and composition. This process allows staging of capital investment and maintaining reasonably high hydrocarbon recovery levels (Fig. 4).
A stand-alone refrigerated stabilizer system would be the least capital-intensive design, but it cannot achieve high product recoveries. Based on OXY's discounted cash flow analysis, it did offer the highest return-on-investment but would provide the lowest total dollar profit potential.
The competing Ryan/Holmes process would be more capital and operating intensive during the early years of the project when no sulfur would be recovered. With increased hydrocarbons or sulfur compounds in the feed gas stream, the cost of a Ryan/Holmes process is impacted to a greater degree than the membrane/refrigerated stabilization system. It does, however, offer the highest hydrocarbon recoveries of the three options.
Utility and product cost differences and a modified feed gas volume/composition profile over the life of a project could easily shift the decision in favor of the Ryan/Holmes process.
INITIAL PLANT COST
Table 3 summarizes the capital costs of the various process systems for the low and high volume design cases. NGL recovery costs reflect Option No. 4. Cost information for the various sulfur recovery options will be discussed in the concluding part of this series of two articles.
Capital costs have been developed to 25% accuracy or better. A contingency was applied to all capital cost to account for miscellaneous errors and omissions.
Copyright 1993 Oil & Gas Journal. All Rights Reserved.
