CLEAN

The joint project CLEAN was a Research & Development (R&D) action with its scientific programme accompanying a pilot Enhanced Gas Recovery (EGR) project designed by GDF SUEZ E&P Deutschland GmbH (GDF SUEZ) in cooperation with Vattenfall Europe. It was funded by Germanys Federal Ministry of Education and Research (BMBF) in the period from 1 July 2008 to 31 December 2011. Within the framework of this R&D project a total of 16 German scientific and economic institutions participated.

The project was set up as pilot to investigate the processes relevant to EGR by the injection of CO₂ into a subfield of the almost depleted Altmark natural gas field (Germany). Despite the setback that permission for active injection was not granted by the mining authority during the period of the project, important results fostering the understanding of processes linked with EGR were achieved.

The CLEAN results provide the technological, logistic and conceptual prerequisites for implementing a CO₂-based EGR project in the Altmark and provide a benchmark for similar projects in the world.

The almost depleted Altmark gas field was chosen by the owner and operator GDF SUEZ E&P Deutschland GmbH (GDF SUEZ) for an Enhanced Gas Recovery (EGR) project. GDF SUEZ took part in the joint research project CLEAN providing this site as a basis for the scientific work of the partners from academia and industry. In November 2007, GDF SUEZ filed the application for injection of up to 100,000 t of CO₂ with the State Office for Geology and Mining of Saxony Anhalt. The permitting process came to a halt towards the end of 2008, because the responsible mining authority considered a national CCS (Carbon Capture and Storage) law to be the only legal basis for approval. The national CCS law still has not been enacted in Germany. In January 2009, the erection of the interim CO₂ storage and conditioning unit in Maxdorf was completed. New flow lines between the Maxdorf and the potential injection wells were fully planned but never build. Corrosion resistant re-completion of the injection wells did not take place either. Within the funding period of the CLEAN project (2008–2011), there was no injection of CO₂.

The implementation of an underground CO₂ storage requires evidence that the storage is and will remain tight in the future. This refers to the cap rock and the wells penetrating it. Assessment and verification of well integrity of accessible wells is technically possible. The available methods, allowing a direct assessment, were evaluated and a measurement and testing strategy is proposed. Unlike accessible wells, already plugged ones require predictive methods for their assessment. These are based on well information and a comprehensive understanding of the coupled thermal, hydraulic, mechanical and chemical processes during well construction, operation and after abandonment. The methods have been applied to a well zone characterised by conditions typical for the subsurface in the area of interest and with regard to the potential injection site in the Altmark. The calculated safety margins emphasize that technical well integrity of the 12 examined boreholes is given for enhanced gas recovery injecting 100,000 t of CO₂ in the Altensalzwedel subfield without a need for any further intervention.

Self-healing of defects was investigated in full-scale experiments under in-situ conditions. In addition to the expected self-healing as a result of salt creep, healing was also observed resulting from the interaction of salt, cement and casing with dry or wet CO₂.

Methods and technologies for CO₂ well monitoring and intervention presented here are sufficient for the mining safety of CO₂ storage wells under high pressure. The system contains technologies proven under field conditions as well as procedures in which CO₂ was applied.

An innovative well abandonment concept was developed and tested in the field for the long-term containment of CO₂ in depleted Rotliegend gas reservoirs. It aims at amending the conventional standard well abandonment procedure, takes advantage of the natural creeping ability of the thick, homogeneous Zechstein salt formation located at around 3,000 m depth in the Altmark area and consists of four main sealing units: (1) a standard sealing element with cement from the reservoir to the impermeable cap rock, (2) a salt plug created in the formerly reamed section of casing within the plastic Zechstein (Upper Permian) rock salt formation, (3) two bridge plugs at the bottom and top of the salt plug and (4) standard cement sealing elements from the top bridge plug to the ground surface. Comprehensive numerical simulations conducted prior to and during the field test in 2010 and 2011 successfully predicted the evolution of the now proven convergence using downhole measurement data. This new long-term sealing concept has been successfully tested at the Altmark natural gas field.

A holistic understanding of the physicochemical processes induced by CO₂ injection and storage in a reservoir is based on a geoscientific characterisation of the overall geological system consisting of reservoir rocks and cap rocks. It requires in a first step a comprehensive baseline characterisation (sedimentological, mineralogical, geochemical, mechanical, etc.) of pertinent parameters and conditions. To properly handle the large amount of different geoscientific information a Data Management System (DMS) was developed, which proved indispensable to conduct such a multi-disciplinary project. The DMS provides a tool for scientific process management, data analysis, integration and visualisation, data transfer and scheduling through specialised database systems and retrieval techniques, storage technology, and efficient data access.

Sedimentological (facies), mineralogical and petrophysical data classify the Altmark Rotliegend sandstones of high quality with best porosity and permeability in altered/bleached aeolian sandstones. Those properties are strongly controlled by sediment deposition (facies-type) and by early and late diagenesis, namely by early pore-filling cementation and late fluid-rock interaction. The fluids involved probably originated from Carboniferous formations and ascended along (re-) activated faults through the Rotliegend volcanic rocks during Triassic-Jurassic times. Major features controlling the extent of fluid-rock reactions are grain size and grain sorting, respectively pore size, pore shape and pore connectivity. Especially grain coating chlorite and pore-filling carbonate and anhydrite will most probably affect reactivity during CO₂ injection. Laboratory batch experiments with core samples reveal that CO₂ saturated brines cause dissolution of (pore-filling) minerals and alteration of fluid and flow properties of the reservoir rocks. This also was emphasised by a parameter study based on numerical modelling, which indicate a tendency of increasing anhydrite dissolution and calcite precipitation with increasing CO₂ partial pressure. Increasing temperature and salinity counteracts this effect. Changes of rock properties observed comprise an increase in porosity, permeability, water binding capacity, rock wettability and a slight decrease in residual gas saturation.

The recent stress field was determined by direct measurement of the effective primary principal stresses. The total primary stresses were calculated from the effective primary stresses, pore-pressure effectiveness and pore pressure. The orientations of the primary horizontal stresses coincide with the directions of the fault zones in the Salzwedel area. Stress ratios for the anhydrite and claystone strata, calculated from the obtained effective primary stresses, are expected to be modified with a pilot injection into the reservoir.

Methods and numerical tools were developed, which are dedicated to the numerical characterisation of the nearly depleted gas reservoir as well as to the simulation of the processes during CO₂ injection and migration storage. The only practical option for predicting the long-term behaviour of CO₂ in reservoirs is numerical analysis, supported by the understanding gained from the relatively short-term laboratory and field-scale experiments. Corresponding to the real site conditions, compositional gas flow has to be considered including mutual interactions with thermal, mechanical and geochemical (reactive) processes taking into account high pressures and temperatures. Process- as well as site-related benchmarks were jointly developed and implemented. It is observed that the high accuracy of complex equations of state (EoS) do not justify the higher computing costs compared to simple, cubic equations of state. The Peng-Robinson equation of state is the most suitable EoS with regard to pure CO₂. At no stage of the process there is any evidence of plastic deformation. Both reservoir and cap rock behave elastically. In general, CO₂-EGR operation might cause only local changes in the pore pressure of the reservoir, while CO₂ storage will increase the reservoir pressure dependent on the injection rate if no gas is produced at the same time. Simulations revealed that an initial tensile stress regime is the safest precondition while the compressive stress state is problematic with regard to the reservoir integrity. The calculated, virtual tracer test, applying krypton, indicates that concentrations at the monitoring wells would be too low to be detected. Hence a test under the considered constraints is not feasible.

One aim of the CLEAN project was to develop and test monitoring methods for the reservoir cap rock and the reservoir itself. It is shown here that advanced injection and production profile evaluation can be achieved using a combination of pressure, temperature and spinner flow meter data. Using distributed temperature sensing, temperature profiles in gas-filled wells can be acquired, as the sensor cable can be stationary during the measurement allowing for simultaneous thermal equilibration along the entire logged profile. A first field test of the developed hybrid wireline logging system was successfully performed under static conditions and the feasibility of warm-back monitoring was shown based on the results of numerical simulations for a possible CO₂ injection scenario. A combined approach using the developed hybrid system for enhanced production logging during injection followed by warm-back monitoring within a subsequent shut-in period would allow for accurate determination of the spatial extent and injectivity of individual CO₂ injection intervals.

Application of the PNG (pulsed neutron gamma) method for estimation of saturation changes is hampered under the considered conditions because of the low contrast between CO₂ and natural gas. However, considering the effect of drying-out of the bound water with associated salt precipitation, which is expected to occur close to injection wells, could be monitored with PNG logs depending on the volume of bound brine.

In preparation to seismic MSP/VSP experiments numerical models were calculated, which helped to find boundary conditions for the survey design. In this framework it was possible to identify the receiver depths with largest amplitude changes (more than 5 %) depending on the offset, the dimension of density and velocity changes and the expansion radius of the CO₂ front. As expected, the amplitude changes during the replacement of the reservoir gas by CO₂ are very small. For a direct detection of the CO₂ front using the seismic wave field, large pressure and/or temperature changes would be required. The concentration of an active seismic monitoring in the Altmark must be focused on the appearance of possible leakages, which cause larger velocity changes. Tests with synthetic datasets showed that seismicity location with the diffraction summation method is possible. For this purpose four stations or rather four boreholes are necessary.

For CO₂ injection projects, stable isotopes and isotope mass balance calculations are new and promising tools to monitor migration of CO₂ plumes and to accurately quantify the amount of CO₂ dissolved in water respectively. The baseline monitoring of the Altmark site yielded a good overview of baseline variations of the DIC (dissolved inorganic carbon) and isotope values, which are clearly distinguishable from the CO₂ to be injected. Laboratory results revealed theoretical trends. They indicate that with high pCO₂, δ¹³C_DIC and δ¹⁸O\( _{{{{\rm{H}}_2}{\rm{O}}}} \) add up as valuable monitoring tools for different respective intervals of DIC concentrations. In these applications of isotope monitoring, fractionation plays a major role as they influence the isotope ratios resulting from dissolution and speciation processes. Thus, the determination of their temperature and pressure dependence as well as the influence of salinity are important to hone isotope techniques to serve as accurate tools for the assessment of CCS reservoirs before and after injection.

The microbial community of the 3.5 km deep nearly depleted gas reservoir was analysed by molecular genetic techniques. Comparative analysis of the fluid samples from three different wells, which differ in their temperature profiles and operative history are presented. DNA fingerprinting indicated the presence of microorganisms similar to previously identified microbes from thermophilic and anaerobic environments in deep hypersaline and hot reservoir environments. Phylogenetic analyses revealed a high microbial diversity including different H₂-oxidising bacteria (hydrogenophaga, acidovorax, ralstonia and pseudomonas), thiosulfate-oxidising bacteria (diaphorobacter), dissimilatory metal reducers (pantoea), aromatic-degrading deep sediment inhabitants, (sphingomonas), extremophiles (bacillus), and biocorrosive thermophilic microorganisms. Additionally several sequences of microorganisms similar to representatives from different saline, hot, anoxic, deep environments were detected, which had not been cultivated previously. Cell numbers of SYBR Green total cell counts were less than 102 cells/ml and the cells were usually attached to particles.

For enhanced gas recovery (EGR) using CO₂ as well as for CO₂ storage in depleted gas fields it needs to be shown that injection and storage is save and neither population nor environment is exposed to risks during operation or afterwards. This requires the development and application of methods to monitor groundwater, vadose zone and atmosphere. Therefore, extensive investigations of the near-surface aquifers were performed to characterize the geological structure and the geochemical and hydraulic conditions as part of a baseline-monitoring and to specify input parameters for model simulations. If CO₂ leakage should occur and CO₂ migrates upwards from the storage complex, shallow freshwater aquifers are the first protected good that might be affected. Based on the model simulations, parameters that would be affected by leakages were specified and parameter changes as well as spatial extension of the expected changes quantified. A comparison of the model results with measured natural variabilities show that especially pH and TIC (total inorganic carbon), but under certain conditions also electric conductivity and aqueous calcium concentration (Ca) are most suited parameters for the detection of CO₂ leakages based on observation wells in shallow aquifers. It was an important result that the temporal fluctuations of groundwater composition are generally small but spatial variations are large.

The simulation results demonstrate that CO₂ gas and dissolved inorganic carbon should be expected and measured in the upper part of a confined aquifer. They also show that CO₂ gas will follow the steepest gradient of the overlying aquitard and accumulate in anticlinal structures, while the dissolved inorganic carbon migrates in groundwater flow direction.

Since the area that can be monitored by observation wells and also the extension of leakage plumes are limited, it is advised to additionally use further, e.g. geophysical monitoring techniques, which are able to cover wider areas. The tested methods ERT (electrical resistivity tomography) and especially SkyTEM (helicopter borne electromagnetic survey) can provide quite comprehensive information about the aquifer down to a depth of 300 m. This is to image the geological structure, which would have the strongest influence on the distribution of rising CO₂ gas and is therefore important for a site-specific monitoring concept. Repeated aero-electromagnetical measurements are so far the only method that could be used for an area-wide leakage monitoring on the relevant scale. All current laws and regulations demand an “adequate monitoring”, but it will be an iterative discussion in the future between operators, public, authorities and scientists to define what should be measured, how often and where.

In case of a large CO₂ leakage rate the gas penetrates the shallow aquifer structures and migrates into the vadose zone. Here the CO₂ can be measured before it discharges into the atmosphere. The CO₂ concentrations in the soil are temporarily variable due to microbial degradation of organic carbon. Soil gas monitoring can be a suitable and sensitive additional method to detect non-natural CO₂ concentrations in the vadose zone. The preconditions are the possibility to measure the CO₂ concentration significantly below microbiologically active soil zones but above varying ground or tail water levels. Both constraints have to be tested by preliminary site investigations. Because of strong local and seasonal variations individual baselines for at least 2 years are required for interpretation of the CO₂ concentration measurements.

CO₂ might leak through the geological formation into the atmosphere, but it can also be released due to accidents in the surface facilities. Only very few investigations exist how CO₂ spreads in the atmosphere and where hazardous concentrations are reached. Where CO₂ is decanted or handled the potential hazards of CO₂ release at the surface or around the facilities need to be considered. Calculations emphasize that the hazardous areas resulting from unwanted releases of CO₂ cannot simply be estimated. Specific hazards must be determined in an event- and location-specific manner. Nevertheless, in that way site-specific prognoses and thus the introduction of specific monitoring and intervention measures are possible. The simulations demonstrate that simple recommendations can improve safety, like turning the exhaust of the safety valves vertically upwards to enhance mixing in the atmosphere.

It is outlined here that many different aspects with regard to the monitoring of atmosphere, vadose zone and groundwater need to be taken into account with regard to EGR and CO₂ storage. In a future step a coherent monitoring concept needs to be defined now for a whole site on the demonstration scale. An intensive discussion between legal authorities and scientist will be required to harmonize what is technically feasible and what is socially and legally expected.

CLEAN was a scientific programme in support of a pilot Enhanced Gas Recovery (EGR) project and it was planned to inject nearly 100,000 t of carbon dioxide (CO₂) into the Altmark natural gas field. Due to delays in the permitting process, injection did not occur within the time frame of the project. Therefore a test case was studied of the theoretical injection of CO₂.

Modelling CO₂ injection in the Altensalzwedel segment of the Altmark depleted gas field shows that effective injection can be carried out at a tubing head pressure as low as 3.5 MPa and at tubing head temperature of 10 °C.

The history matching of the simulation model and the injection prediction were successfully conducted with a volume ratio of injected CO₂ to the gas in place of approximately 0.06 which did not show any enhancement in the gas recovery (no EGR effect). Within the pilot project period of 2 years there would be no CO₂ arrival at the production well located at about 1,600 m from the injector. Penetration of the CO₂ into the reservoir was maximum approximately 800 m and minimum around 250 m depending on the permeability of the formation layer. The gas mixing zone reaches the observation well which is located at about 660 m from the injector already before 1 year of injection.

Under the simulated conditions direct monitoring of the CO₂ front with active seismic measurements will not be feasible as the reservoir layers are too thin and the effects of CO₂ gas replacing reservoir gas are too small. Seismic monitoring has to concentrate on the detection of possible leakages in shallower aquifers where larger velocity and density changes would occur.

Critical public perception of capture and subsequent geological storage of CO₂ (CCS – Carbon Capture and Storage) may turn out to be a hindrance to the development of the CCS technology. Therefore, in addition to the scientific and technological development, acceptance by the public is a decisive prerequisite for the successful implementation of this technology. Notwithstanding the result of any decision, it is desirable that decisions on acceptance or rejection of a technology are based on factual arguments rather than on speculations. Public outreach within the CLEAN project therefore aimed to support opinion-forming based on scientific knowledge and transparent discussion of concerns, benefits and knowledge gaps. All this was addressed with a variety of media, tailored to the respective target audience. However, conducting effective public outreach does not necessarily guarantee the success of a project.