Prevention of CO2 leakage from underground storage reservoirs
MetadataShow full item record
The risk of leakage from CO2 storage sites is recognised as one of the challenging aspects of large scale implementation of geologic sequestration of CO2. Uncertainties in characterizing a geologic reservoir and the current lack of a complete understanding of possible interactions between rock and fluids involved in CO2 storage have resulted in concerns over contingent leakages. The debate on the allowable rate of leakage has led to different perspectives among the CCS stakeholders; some believe that, by analogy to natural CO2 reservoirs, risk of having leakage of less than 1 %/year is dispensable and on the other side, some state that “Any non-zero leak-rate from a stored carbon system means that eventually the entire inventory will be released to the atmosphere”. There is also the issue of public acceptance which would be adversely affected by the non-zero potential of leakage of CO2 back to the surface. This negative impact on the members of the public has proved to be very powerful as it has resulted in the delay and even cancellation of some CCS (carbon capture and storage) projects. To the best of our knowledge, no practically viable techniques existed for prevention of CO2 leaks from unknown leakage paths. Our technique is based on in-situ precipitation of an appropriate solute dissolved in the stored super-critical CO2. Supercritical CO2 (SCCO2) has a distinct characteristic that its density changes from gaseous-like to liquid like monotonically and uniformly. This allows SCCO2 to act as manageable solvent for various solid solutes. Thus, once the solution of SCCO2 + solid solutes departs from the equilibrium conditions, the solute will appear in the form of crystallized particles. Based on this unique behaviour of the supercritical solutions, we have developed a novel technique for tackling contingent CO2 leakage from storage sites as a preventive method. The sealing process takes place in-situ at the exact location of the leak without the need for identifying the leak target area and the exact nature of the leak. In this study, an integrated research methodology was designed and employed to comprehend the physics behind our leakage prevention technique and also, to deliver the required package, i.e. suitable solutes and reliable simulator, for larger scale implementation of this technique. It was aimed firstly to demonstrate the performance of our proposed leakage prevention technique at different leakage scenarios and secondly, to put forward a number of solutes efficient in tackling contingent leakages. In order to identify the underlying mechanisms and the pertinent parameters controlling the efficacy of this technique, a good number of direct visualisation experiments were performed where the kinetics behind solute solidification and precipitation were visually investigated. Three different ranges of potential solid solutes were used in visualisation experiments to cover a wide spectrum of solute solubility in supercritical CO2, which would enable us to draw more general and consistent conclusions. The understandings acquired from the direct visualisations were employed to design efficiently a few yet adequate number of coreflood experiments in which the performance of our technique was studied in more realistic reservoir cores. Having attained the adequate information from the experimental part of this investigation, the findings was subsequently utilised to develop an in-house simulator to fundamentally model the kinetics of solid solute precipitation and consequently, the pertinent parameters of the semi-empirical equations were tuned to match and predict the coreflood experiments. In experimental part of this investigation, a series of visualisation experiments using transparent porous media (micromodel) to physically simulate CO2 leakage under conditions typical of geologic storage sites. In these experiments, degree of “supersaturation” was identified as an important parameter behind effectiveness of solute precipitation. In addition to evaluating the behaviour of different solutes, the impacts of resident water existing in storage reservoir and impurities in CO2 stream were taken into account in visualisation experiments. Utilising the findings from the visualisations, 6 coreflood experiments were carried out, which revealed that a strong and durable blockage was formed in the core and the flow (leakage) of CO2 was effectively sealed. Practically speaking, there should not be any premature precipitation as the solution travels inside the storage reservoir; therefore, apart from the performance of this technique in the vicinity of contingent leakages, the integrity of the solution (as it flows in the simulated storage reservoir) was also investigated in visualisation and coreflood experiments. From the findings revealed by the coreflood and micromodel experiments, it was identified that the solution made with solid-solute and SCCO2 may not be responsive in some scenarios. Therefore, the desire to better control the onset of blockage formation has triggered investigation of developing a complementary method to be able to adjust the response of the solution. It was rationalised that adding another solutes (co-solvent) to the solution would enable us to modify the response of the solution. Sandpack, micromodel visualisations, and coreflood experiments were performed to evaluate influence of co-solvent on the response of the solution to various leakage types. On the modelling the precipitation process in the leakage path, it was first demonstrated that conventional reservoir simulators could not adequately capture the physics leading to the blockage formation and the results of lab-scale coreflood experiments could not be correctly simulated. Therefore, there is a need for developing models, which can predict the performance of the LPT at different cases. Based on the experimental information, we have attempted to develop the relevant equations that describe the mechanisms behind particle formation due to pressure drops. After matching one coreflood experiment, the resultant model was used to predict another coreflood experiment performed at similar conditions, which demonstrated an encouraging performance for the developed mathematical model. The results and findings of this study have primarily verified that our leakage prevention technique, which is developed here through extensive experimental and modelling investigation, is well-capable of tackling various contingent leakages. A number of economically feasible solid solute has been found with positive responses to physically simulated leakage paths, which would be considered as the potential solutes for large scale implementation of our technique. Moreover, an in-house simulator was developed based on the finding observed in the different experiments. The simulator can successfully predict the results of coreflood experiments, which implies that it captures the underlying mechanisms adequately. Having developed the necessary equipment, i.e. appropriate solutes and reliable simulator, our proposed leakage prevention technique is ready to be incorporated in demonstration and pilot trials.