Duke Energy installs first Morphysorb unit at Kwoen gas plant
KWOEN GAS PLANT—Conclusion
The Morphysorb process, in its first commercial application at the Duke Energy Gas Transmission Corp.'s (DEGT) Kwoen plant, proved to have better performance than competing solvent processes.
Part 1 (OGJ, July 5, 2004, p. 00) provided background information about the Morphysorb process development and the opportunity for its first commercial demonstration.
Part 2 discusses Morphysorb process performance and operational issues associated with solvent performance since plant start-up in August 2002. Part 2 also highlights the technical advantages of Morphysorb compared to leading commercial physical solvent processes.
DEGT assessed Morphysorb's performance using five variables: acid-gas pickup, recycle-gas flow, total hydrocarbon loss in acid-gas stream, solvent losses, and foaming-related problems. Plant data gathered during 16 months of operations show that the Morphysorb solvent has performed well and exceeded its targets in all of these categories.
In addition to these specifications, the choice of a solvent in gas-treating applications depends upon practical attributes such as chemical stability, corrosivity, and cost-effectiveness. The first commercial application of Morphysorb confirmed these features of the solvent.
The solvent showed a higher relative absorption of acid-gas compounds compared to competing solvents, which results in lower circulation rates. It has a lower relative absorption of higher hydrocarbons, which also results in further reduced circulation rates, lower recycle-gas compression costs, and a higher heating value and less losses in the sales gas.
Morphysorb performance at Kwoen
The Kwoen plant was mechanically completed in early August 2002. Morphysorb solvent was delivered to the site, loaded to the process train, and brought up to operating pressure with sweet gas.
Initial testing with sour gas began in early September 2002. The plant achieved the initial planned flow rates of 90 MMscfd using available gas. The initial process data observations occurred in a single column that was running under capacity.
DEGT originally anticipated full flow in both contactors, but the availability of gas and bottlenecks downstream required that only one contactor run at full capacity. The test was to demonstrate that the Morphysorb process could achieve performance measured by several variables established between the process owners and DEGT.1 2
As part of an agreement, Gas Technology Institute (GTI) and Uhde GMBH conducted a 72-hr performance test at the Kwoen plant. Gas sampling and analytical equipment was set up in a laboratory trailer for the performance test. The parties, according to the agreement, established certain performance measures.
The measures included: acid-gas pickup per unit volume of liquid, recycle rate, methane losses in the reinjection gas, solvent losses, and absence of any foaming problems. A further provision of the agreement stipulated that GTI could adjust the specific values of the measures if the feed gas conditions (composition, flow, temperature, or pressure) were substantially different from the agreement.
The test data discussed in this article were collected during the performance test period Nov. 12-15, 2002.
GTI-Uhde used electronic and manual methods to collect the process parameter data. The electronically collected parameters were recorded using a computerized data-acquisition system at the Kwoen plant.
The process operating data, which was available on the data-acquisition system, was recorded every 5 min in a spreadsheet using macros.
The liquid circulation rate was a fixed value for the tests. Material balances were computed for the test periods that were sufficiently close to balance; the resulting values for the performance measures were then computed.
Table 1 shows the performance test data. Table 2 provides information on gas composition of individual streams.2 3
The test run was maintained for 53 hr, during which time the required process data were recorded every 5 min. Seven gas sampling points and two liquid sampling points were used during the performance test.
In all cases, the measurements satisfied the criteria that DEGT established. The Morphysorb solvent met or exceeded all performance targets; solvent losses were determined separately. The measurements were in-line with computer simulations.
Due to low coabsorption, only 0.95-1.2 mole % of the hydrocarbons were lost to the acid-gas stream.
The recycle-gas stream was 50% lower than predicted for the conventional physical solvent that DEGT originally considered.
DEGT formally accepted the Morphysorb process at the Kwoen plant in December 2002, and the process has operated continuously since then. In December 2002, the gas volumes at the Kwoen plant were available at design capacities.
Table 3 shows the typical operating data at design feed flow.
Although the Morphysorb plant can produce an acid-gas stream of 36 MMscfd or more at design capacities, the acid-gas removal capacities at Kwoen depend on the Pine River plant requirements and, in no case, should exceed 29 MMscfd until a regulatory approval for a new acid-gas injection well is obtained.
As of May 2004, about 135 bcf of sour gas was processed. Of this, about 12 bcf of acid gas containing mainly H2S and CO2 was injected back into the depleted reservoir and 123 bcf was sent to the Pine River plant for further processing.
DEGT's incentive for this project was to produce more gas and avoid adding sulfur-processing capacity at Pine River, which would have been prohibitively expensive.2 3
The process operation was uncomplicated and stable; no special attention to process control was needed once steady-state conditions were reached. No solvent-related process upsets were reported since the plant started up.
Because the compressors leaked hydraulic oil to the process, one addition of antifoam agent was temporarily required.
In 2003, GTI monitored solvent performance and observed no solvent degradation or undesired side reactions with gas constituents.
Solvent losses to the process gas due to evaporation were within the expected range. Because the solvent contains no water, carbon steel is generally used as the material of construction.
The Kwoen plant purified the gas going to Pine River in terms of acid-gas content so that more natural gas could be produced for sale without any additional sulfur production.
Fig. 1 shows the weekly averages of the feed, product, recycle, and acid-gas flows. The plant operated close to its full capacity of 300 MMscfd throughout 2003.
Fig. 2 shows the gas composition in the acid-gas injection stream.3
Solvent thermal, chemical stability
Solvent analyses performed at GTI indicate that the Morphysorb solvent is still stable after 16 months of operation. We have observed no solvent degradation, and the solvent's composition is similar to its start-up value. Total makeup solvent added to date is less than predicted initially.
The solvent-component ratio deviation compared to the initial solvent loading varied only ±4% from the value at initial filling.
The solvents N-formylmorpholine and N-acetylmorpholine are the products of the reaction of morpholine with formic and acetic acid, respectively.
The solvent samples were analyzed for thermal and chemical degradation products such as formate, acetate, sulfate, oxalate, and chloride. Frequent solvent sampling showed nearly the same composition as the fresh solvent. A total of 10 solvent samples were collected since plant start-up.
Table 4 lists some of the degradation products.
Sulfate concentration rose unexpectedly and then decreased with the later samples. This is most probably due to the co-elution of SO3 and SO4 during the analysis.
Unfortunately, we could not see a pure SO4 peak due to the dissolved H2S; instead, we saw a combination of two peaks.
The samples in which sulfate concentration measurement was a problem also had high levels of degassing H2S; therefore, the sulfate concentration values are incorrect and one should ideally measure them after degassing all H2S from the sample.
Chloride concentrations increased around June. This might have been due to Morphysorb addition to the system. There might have been some salt water trapped somewhere between the storage tanks and the system that was released during solution addition—this may have contributed to the increase in chloride concentration.
The concentration decreased in later samples.
Oxalate and formate concentrations remained nearly constant. This indicates that the Morphysorb is not degrading to oxalate or formate.
Corrosion
Corrosion monitoring during the first year of operation detected no measurable effect on the equipment. In the Kwoen plant, corrosion is monitored using an advanced field signature method (FSM) and a Microcor monitoring system manufactured by Corrpro Canada Inc.
FSM is a nonintrusive technique for corrosion monitoring of process equipment and is currently used in major oil and gas facilities. Any changes due to general corrosion, erosion, fatigue, or other cracking phenomena cause a change in the "fingerprint" or field signature.
FSM generates a graphical interpretation that shows the severity and location of pits, cracks, and corrosion.
In the Kwoen plant, these FSM probes are installed in three different locations: in a rich solvent stream downstream of the contactor, in a solvent stream between high-pressure and intermediate-pressure flash vessels, and in a solvent stream between the intermediate-pressure and medium-pressure flash vessels.
Fig. 3 shows the FSM graphical interpretation of corrosion in a location downstream of the contactor on four different dates.
Most of the metal loss at this location is due to erosion—it is located in an area of high-pressure, rich solvent loaded with acid gas and hydrocarbon components.
The erosion occurred immediately after plant start-up but leveled off during continued operation.
The Microcor corrosion monitoring technology was installed at the Kwoen plant to increase the speed of response vs. conventional monitoring techniques such as coupons and electrical resistance probes.
There are about 10 probes installed throughout the plant, mostly in liquid streams and a few in the gas phase where corrosion monitoring is needed.
Fig. 4 shows the corrosion data with Microcor probes.
Data analysis from the FSM and Microcor probes indicate that corrosion is minimal (<0.24 mils/year=0.0068 mm/year), an advantage of this process.
Overall plant corrosion is minimal, which confirms earlier experience with N-formylmorpholine in aromatic-removal plants.
Iron, hydrocarbon accumulation
GTI also observed that the iron content in the solvent is increasing. Researchers expect this to level off in the near future.
The slow increase in iron concentration in the solvent from start-up to the present is a likely result of typical passivation at the carbon-steel surface.
The online corrosion-monitoring system confirmed that corrosion in the plant is low at less than 0.24 mils/year.
GTI suspects, and DEGT agrees, that the rise in iron concentration is due to iron sulfides in the incoming sour gas.
Table 5 gives details of hydrocarbon and iron content in the solvent samples.
GTI's analysis shows that the concentration of higher hydrocarbons in the solvent has also increased. This phenomenon is common in gas-treating plants that use physical solvents.
DEGT will continue to monitor the accumulation of "heavies" in the solvent and their effect on solvent performance.
Different procedures are available if the concentration of heavies increases beyond acceptable limits. DEGT has started skimming hydrocarbon layers from flash vessels to reduce the hydrocarbon content in the solvent.
Solvent foaming
As previously mentioned, about 135 bcf of gas had been processed as of May 2004 and antifoam agent has been used only once when leaking hydraulic oil from the recycle compressors to the process solvent stream caused foaming.
A small addition of antifoam agent was temporarily used successfully.
Total hydrocarbon losses
The acid-gas methane content is continuously monitored using an online analyzer.
The total methane losses are 0.95-1.2 mole %.
This is another advantage of this process. GTI-Uhde calculates that methane losses are 3-4 times lower than the competing physical solvent process.
Technical advantages
Figs. 5 and 6 show the increased acid-gas solubility for CO2 and H2S in the Morphysorb solvent compared to a conventional blend of dimethylether of polyethylene glycol (DMPEG). Because Morphysorb has a higher density than DMPEG, more H2S (up to 20%) and CO2 (up to 10%) can dissolve in the new solvent than in DMPEG per unit volume.
Fig. 7 shows the solubility of methane in the Morphysorb process and demonstrates the advantage of the new solvent in terms of lower hydrocarbon coabsorption.
The same behavior also applies for higher hydrocarbons. In physical solvent based, gas-treating plants in particular, it enables a lower recycle gas flow that dramatically reduces the recycle-gas compressor power consumption, leads to a smaller absorber column size for the same feed-gas flow, and reduces subsequently the required solvent circulation.
Morphysorb applications
Morphysorb can be tailored for either bulk or trace removal of acid-gas components. Morphysorb can be used for:
- Subquality natural gas upgrading to either pipeline or LNG specification.
- Selective removal of H2S from natural gas or syngas to generate acid-gas streams that are suitable for Claus plant feed for sulfur production.
- Selective removal of H2S, CO2, COS, CS2, mercaptans, and other components from coal or oil gasification syngas at integrated gasification combined-cycle facilities.
- Simultaneous removal of aromatic compounds such as benzene, toluene, ethylbenzene, and xylene.
- CO2 and H2S removal for CO2 sequestration and enhanced oil recovery schemes that require high pressure acid-gas recovery for acid-gas reinjection.
Acknowledgment
The authors thank Duke Energy Gas Transmission Corp. for providing help in gathering operating data for the Kwoen plant, and collecting and providing solvent samples for analysis. The authors also thank Glenn Kowalsky, for his efforts in bringing Morphysorb to commercialization, and Troy Wilfur, Neal McGarry, and other Kwoen plant staff.
References
1. Kowalsky, G., Leppin, D., Palla, R., and Jamal, A., "Large Scale Demonstration of Morphysorb Technology on Natural Gas Treating," presented to the GTI Natural Gas Technologies Conference, Orlando, Sept. 30-Oct. 2, 2002.
2. Kowalsky, G., Leppin, D., Palla, N., Jamal, A., and Hooper, M., "Performance of Morphysorb in a Commercial Acid Gas Treating Plant," presented to the 53rd Laurance Reid Gas Conditioning Conference, Norman, Okla., Feb. 23-26, 2003.
3. von Morstein, O., Menzel, J., Palla, N., and Leppin, D., "An Ace at Removal," Hydrocarbon Engineering, February 2004.
