Controlled CO2 Injection into a Shallow Aquifer and Leakage Detection Monitoring by Two Different Leakage Events at the K-COSEM Site, Korea

It is a major concern about Carbon capture and storage (CCS) that the injected CO2 might be leaked through the injection pipe or fracture. These release events of CO2 can lead to a serious environmental hazard. Therefore, it is necessary to detect accidental CO2 leakage more accurately and to decrease the risk of CCS. Artificial carbon dioxide injection into a shallow aquifer system was carried out with two injection type imitating short- and long-term CO2 leakage events into a shallow aquifer. One is pulse type leakage of CO2 (6 hours) under a natural hydraulic gradient (0.02) which imitate the point leaking situation developed along the vertical preferential pathway such as fracture or a crack in an injection pipe and the other is long-term continuous injection (30 days) under a forced hydraulic gradient (0.2) which indicated situation that CO2 leakage can be happened at a wide-ranging areas. Injection and monitoring tests were performed at the K-COSEM site in Eumseong, Korea where a specially designed well field had been installed for artificial CO2 release tests. Total 45 monitoring wells which include 4 multi-depth monitoring well nests for depth-discrete gas monitoring at unsaturated zone and also 4 multi-depth monitoring well nests for groundwater monitoring at saturated zone were installed at the site aligned with the regional groundwater flow direction. CO2-infused and tracer gases dissolved groundwater was injected through a well below groundwater table and monitoring were conducted in both saturated and unsaturated zones. Real-time monitoring data on CO2 concentration in saturated zone and unsaturated zone using NDIR sensor and OL-AP (open loop-air purging system), temperature using TLS (thermal logging system) and hydraulic parameters (pH and EC) were collected. Also, periodical measurements of several gas tracers (He, Ar, Kr, SF6), TIC (total inorganic carbon), carbon isotope (δ13C) were obtained. The pulse type short-term injection test was carried out prior to the longterm injection test. Results of the short-term injection test, under natural hydraulic gradient, showed that CO2 plume migrated along the preferential pathway identified through hydraulic interference tests. On the other hand, results of the long-term injection test indicated the CO2 plume migration path was aligned to the forced hydraulic gradient. Compared to the short-term test, the long-term injection formed detectable CO2 concentration change in unsaturated wellbores. Results of process-based analysis based on concentration data of O2 and CO2 indicated that CO2 concentration in unsaturated zone before injection affected by both biological respiration and CO2 dissolution. However, the origin of CO2 is changed to by exogenous addition ever since CO2 injection carried out. As time passed after shut down the injection event, the CO2 concentration reverts to its original position. Recovery data of tracer gases made breakthrough curves. From the interpretation results of recovery data, the physical processes such as phase-partitioning and degassing process governing mass balance problem of leaking CO2 in shallow groundwater system was identified. The monitoring results indicated that detection of CO2 leakage into groundwater was more effectively performed by using a pumping and monitoring method in order to capture by-passing plume. With this concept, an effective real-time monitoring method was proposed.