Evolution of formation water chemistry and geochemical modelling of the CO2CRC Otway Site residual gas saturation test

摘要

The CO2CRC Otway Project Stage 2B field test was a residual CO2 saturation and dissolution experiment carried out on the Paaratte Formation in the Otway Basin of Australia. During the 87 day test a downhole U-tube sampling assembly was used to collect formation water samples at various time intervals. The waters were analysed for chemical composition, dissolved gas content and stable isotopic composition of H and O. Changes in the chemical and isotopic composition of the formation water were observed from the initial baseline samples collected through to the conclusion of the test. Systematic reoccurrences of relatively large linear changes in the alkalinity, Ca~(2+), Fe~(2+), Mg~(2+) and SiO_(2(aq)) were observed during different test sequences. Similar behaviour occurred for the δ~(18)O and δ~2H values suggesting a complex system controlled by at least two processes. The chemical evolution was consistent with CO2-water-rock interactions, dominated by carbonate dissolution with some silicate dissolution. The mineral dissolution driven changes in water composition are overlain by what was clearly mixing of the injected and in situ formation water. The isotopic composition of the water was used to confirm the role of mixing and the calculated residual gas saturation values were broadly consistent with those derived by other methods. The extent of reaction in the relatively short time frame of the test was used to evaluate the capabilities of kinetics based reaction path models. The models were generated using published reaction rate data with commonly used assumptions regarding upscaling of reactive surface area. These assumptions consist of reducing the calculated geometric surface area that accounts for grain shape and roughness by factors of 100 to 1000 depending on the mineral habit. The numerical models provided a very good fit to the field data indicating that the rate data and the associated assumptions are relatively robust.