Carbon capture and sequestration (CCS) holds promise for use on an industrial scale. Total has been involved in CO2 injection and geological storage for more than 15 years, in Canada (Weyburn oil field) for enhanced oil recovery and in Norway (Sleipner and Snohvit fields) for aquifer storage. In 2006, the company decided to set up an experiment for CO2 capture, transportation, and injection in a depleted gas reservoir. The pilot is in the Lacq basin in southwestern France, 800 km from Paris, and has been on stream since January 2010.
Introduction
This CCS project entailed converting an existing air/fuel-gas combustion boiler into an oxygen/fuel-gas combustion (oxy-combustion) boiler using oxygen delivered by an air-separation unit (ASU) to obtain more-efficient combustion and a more-concentrated flue-gas-CO2 stream. The 30-MW oxy-combustion boiler can deliver up to 38 t/h of steam to the high-pressure steam network of the Lacq sour-gas-production and -treatment plant. After quenching the flue-gas stream, the rich CO2 stream is compressed to 27 barg, dried, and transported in gaseous phase by use of existing pipelines to a depleted gas field, 29 km away, where it is injected into the deep Rousse reservoir. Over a 3½-year period, up to 90 000 t of CO2 could be injected.
The main objectives in this experiment were:
- To demonstrate the technical feasibility and reliability of an integrated chain comprising steam production and CO2 capture, transportation, and injection into a depleted gas reservoir.
- To acquire operating experience and data to scale up the oxy-combustion technology from pilot scale (30 MW) to industrial scale (200 MW) while reducing costs.
- To develop on-site monitoring methods and technologies to serve future onshore storage-monitoring programs that will be larger in scale, longer in term, and economically and technically viable (i.e., microseismic and environmental monitoring).
Technical Description
The CCS pilot installation, shown in Fig. 1 above, consists of an ASU, an oxy-combustion boiler, a direct-cooling contactor, a CO2 compressor, a dryer system, a transportation pipeline, and an injection site (i.e., compressor, injection well, reservoir, and a subsurface-seismic network).
ASU. A dedicated ASU, the oxygen-production unit, was installed on the Lacq gas-treatment complex to produce 240 t/d of low-pressure oxygen (1.8 bara) at 95 to 99.5 vol% purity. Only 99.5% pure oxygen is used to feed the oxy-combustion boiler. The nitrogen rejected by the ASU is partially used for regenerating the dryer system.
Oxy-Combustion Boiler. At Lacq, one of the five air-combustion boilers built in 1957, which provides steam and electricity necessary for the industrial complex, was retrofitted as an oxy-combustion boiler with pure oxygen replacing air for the combustion of commercial gas. This technology is designed to manage heat transfer inside the combustion chamber, without requiring special construction materials, and to adapt the combustion characteristics to a conventional heat-exchanger design. The existing air-fired boiler was adapted to oxy-combustion, mainly by improving seals to limit air ingress, and by implementing a flue-gas recycle duct and fan, through which flue gas will be partially recycled to the inlet of the oxygen burners to dilute the oxygen flames and maintain the temperature at an acceptable level inside the combustion chamber. After this revamping, the boiler pressure was slightly higher than atmospheric pressure.
This 30-MW oxy-combustion boiler started operation in 2009 and can produce up to 40 t/h of high-pressure steam (60 bar and 450°C) to feed the high-pressure network of the Lacq industrial complex. The flue gas leaving the boiler contains 31–34 vol% CO2 and 62–65 vol% H2O along with other components such as nitrogen, argon, and oxygen.
Direct-Cooling Contactor. The flue gas leaves the oxy-combustion boiler at approximately 220°C and is cooled to 50°C in a direct-cooling contactor, removing most of the water by condensation. The dry-CO2 stream moves to the Lacq CO2 compressor, a three-stage reciprocating compressor with 1-MW power, designed to compress the CO2 stream from 1 to 27 barg.
Dryer System. The dryer system has two molecular sieves that remove the residual water from the CO2 stream. The system yields a maximum concentration of 30 ppm, which eliminates any condensation in the transportation pipelines.
Transportation Pipeline. The CO2 stream is sent to the Rousse injection compressor through a 29-km pipeline. Ten emergency-shutdown valves are installed along the pipeline to minimize consequences in the event of pipeline leakage. Depressurization devices also are available. The pipeline is made of carbon steel and was used for more than 30 years to transport natural gas produced at Rousse to Lacq. It was subjected to an in-depth inspection before being reused.
Injection Site. The injection site at Rousse is in an unpopulated rural area, 5 km south of Pau. The Rousse compressor is a one-stage reciprocating compressor, designed to compress the CO2 stream from 27 to 51 barg for injection into the reservoir. The CO2 is injected through an existing well, Rousse 1, which produced wet sour gas from 1972 to 2008. Well Rousse 1 was selected after cement and corrosion logging were undertaken in 2006. A workover in early 2009 converted the well to injection. It was equipped with four pressure sensors, four temperature sensors, and three microseismic sensors fitted along a fiber-optic cable at the bottom of the completion.
The Rousse field, now depleted, is a deep, isolated Jurassic horst reservoir at a depth of 4500 m below ground level. Cores were taken from 70 m of its 120-m thickness along with part of the Cretaceous caprock. The initial reservoir pressure was 485 barg. The average downhole temperature is 150°C. The initial gas in place contained 4.6 vol% CO2 and 0.8 vol% H2S.
A system is in place to identify any mechanical effects of the injection on the reservoir and on the caprock that might affect the integrity of the storage. The system comprises seven shallow wells (200 m deep), each with four microseismic sensors. Six of the wells are within a 2-km radius of the injector well, and the seventh is on the Rousse well pad. There are three microseismic sensors at the bottom of the Rousse injection well (4200–4400 m below ground level). In addition, a seismometer at the Rousse well pad, near the seventh well, detects and records natural Earth tremors.
Monitoring Program
The monitoring program was designed to provide key information and data from the injection well and the reservoir and from deep-subsurface, near-surface, and surface levels. The initial program planned for 3½ years of injection (until July 2013) followed by 3 years of post-injection observation (from July 2013 to July 2016). Then, because a new permitting procedure will be necessary to obtain authorization for permanent storage of the injected CO2, a long-term monitoring program will be designed on the basis of technical and economic lessons learned from the previous 6½ years and from research-and-development results.
The following parameters are monitored:
- CO2-stream composition, concentration, and flow (continuously)
- CO2 atmospheric concentrations at the injection-well pad (continuously)
- Well-annulus pressure (continuously)
- Pressure and temperature along the injection well (continuously)
- Injection-well reservoir pressure and temperature (continuously)
- Reservoir and caprock integrity—microseismic monitoring (continuously)
- Soil-gas concentration and fluxes (periodic campaigns)
- Groundwater quality (periodic campaigns)
- Surface-water quality (periodic campaigns)
- Ecosystem biodiversity—fauna and flora (annual inventory)
Operating Feedback
The total quantity of CO2 injected into the Rousse reservoir by the end of May 2012 was 43 000 t. After commissioning and fine tuning of each piece of equipment, the whole pilot plant, with its retrofitted oxy-combustion boiler, was fully operational in January 2010. Overall operation of the pilot plant has proved satisfactory. The choice of having the oxy-combustion boiler start up in air mode and switching to oxygen mode, with load variations up to full capacity, was found to be robust and in line with the predicted behavior. The flue-gas treatment is working as designed. The molecular-sieve dryers effectively lower the dewpoint of the CO2-rich stream to protect the carbon-steel transportation pipe from corrosion. No corrosion has been recorded along the pipeline. The transportation system works as per design.
The only piece of equipment in the CCS pilot plant that proved challenging was the Lacq CO2 compressor. The main parts of this three-stage reciprocating compressor are made of corrosion-resistant material, except for the cylinders, which are in molded cast iron. The suction chamber of the third-stage cylinder soon came under severe attack from acid corrosion.
CO2 Atmospheric Concentration at the Rousse Well Pad. Permanent catalytic CO2, CH4, and H2S sensors were placed around the injection wellhead on the Rousse injection pad to detect any abnormal concentration of these gases that might indicate a leakage. CO2 sensors are installed specifically for monitoring such an injection site. Since injection began, no abnormal detection has been recorded.
Well-Annulus Pressure. Annulus-pressure monitoring is common for tracking well integrity. A pressure increase may correspond to leakage through the tubing or casing. The well-annulus pressure is recorded continuously in the different annuli, but no increase has been recorded since injection began.
Reservoir- and Caprock-Integrity Monitoring. The seven shallow-well sensors have been running since March 2009, allowing 9 months of baseline survey. This permanent monitoring is sensitive enough to detect seismic events corresponding to a displacement of 0.05 mm for a 1-m-length fault (magnitude of −2) with a localization uncertainty of ±250 m. During the baseline survey, mainly natural seismic events caused by deep tectonic activity of the Pyrenees mountain range were recorded.
Since March 2010, four microseismic events have been detected around Rousse besides natural earthquakes related to activity in the Pyrenees. The four microseismic events may be linked to natural microseismic movements of the basement faults below Rousse linked to tectonic compressive stress of the Pyrenees, to the 30-year depletion effects of Rousse, or to injection.
The three microseismic sensors at the bottom of the well completion have been operating since March 2011. Interpretation of the record of microseismic events is ongoing. Since injection started, only low-magnitude (<0) microseismic events have been recorded, incapable of producing any effect on the caprock or reservoir integrity.
Soil-Gas Concentration and Fluxes. Soil-gas monitoring measures CO2 and CH4 concentrations 1 m below the ground surface, and CO2 and CH4 fluxes at the soil/atmosphere interface at 35 locations around the injection site. The gas fluxes are monitored by the accumulation-chamber method, using external recirculation, which is intermediate between static and dynamic principles.
Before CO2 injection in the Rousse reservoir, baseline monitoring was performed to define soil-gas behavior as a function of time. Quarterly data acquisitions were made from September 2008 to December 2009. The data collected during the baseline survey showed broad heterogeneity between the locations. However, seasonal variations were observed clearly, corresponding to the intensity of biological activity in the soil (high in summer and low in winter). The CO2 concentrations vary from 0.04–7.0% in winter to 12–16% in summer. The soil-gas concentrations and fluxes are measured during the injection period, in winter and autumn when biological activity is low. No deviation from the baseline has been detected.
Groundwater Quality. Four perched aquifers above the deep storage reservoir are monitored. Four parameters (pH, water conductivity, and carbonate and bicarbonate concentrations) are analyzed twice yearly at four natural springs in the vicinity of the injection site. The resulting indicators are compared with the baseline reference data from the surveys performed in 2009 (spring, summer, autumn, and winter) before injection began. So far, no change in water quality has been recorded, affirming the absence of CO2 leakage from the reservoir.
Surface-Water Quality. Surface-water monitoring consists of checking two standardized bioindicators every 6 months (in spring and in autumn). The indicators are the French standardized diatom index and the French standardized benthic invertebrate index. The water-chemistry and mineral-content parameters (i.e., pH, water conductivity, and carbonate and bicarbonate concentrations) are checked at three locations of the Arribeu brook that drains the Rousse area. Locations on two other brooks are used as distant references. The results of the biannual analyses of all the indicators are compared with the baseline reference data from the two surveys performed in spring and autumn 2009, in the preinjection period. To date, no change in water quality has been recorded, affirming no CO2 leakage from the reservoir.
Ecosystem Biodiversity. The biodiversity of the ecosystems around the Rousse injection site is checked every year. An annual inventory is drawn up at 33 places around the injection location for the flora of representative ecosystems and at 50 places around the injection location for several amphibian and insect species. So far, no change has been recorded, indicating no leakage of CO2 from the reservoir.
This article, written by Senior Technology Editor Dennis Denney, contains highlights of paper SPE 157157, “The Lacq CCS Pilot, a First,” by Jacques Monne, SPE, Total, prepared for the 2012 SPE/APPEA International Conference on Health, Safety, and Environment in Oil and Gas Exploration and Production, Perth, Australia, 11–13 September. The paper has not been peer reviewed.