Experimental and numerical analysis of reservoir performance for geological CO2 storage in the Ordos Basin in China

Unique reservoir performance was observed in the Shenhua 100,000 t/year Carbon Capture and Storage (SHCCS) Demonstration Project. Suggested by the geological pre-assessments, hydraulic fracturing and a multi-layer injection procedure were employed to improve the injectivity and reduce the risk of an overpressure. However, in-situ data showed the total injection rate increased after the injection started, while the injection initiation pressure decreased with only a minor pressure build-up development. Additionally, the injectivity of the uppermost injection layer, which was not fractured, grew considerably over the years, making this layer potentially able to meet the target rate by itself. To clarify this unforeseen observation, the reservoir performance was investigated through numerical simulations and comparison against the 2.5-year historical data. The simulation results indicated that permeability heterogeneity of the injection layers might explain the observed reservoir performance. High CO2 injectivity in the uppermost injection layer could be attributed to its overall permeability being higher than that of other layers, and the considerable injectivity increase over the years could have been caused by the substantial permeability increase along the principal direction of CO2 migration in this layer. The injectivity improvement caused by hydraulic fracturing was significant in the early time of injection, but it dramatically reduced afterwards. The intermittent injection procedure could effectively reduce the pressure build-up in the reservoir and helped to maintain the injection at the target rate. Based on these assessments, the cumulative injected CO2 mass could reach 300,000 t in December 2015, but the yearly average injection rate would drop slightly. The predicted cumulative mass could be underestimated because the higher injectivity in 2014 was not accounted for in the calibration, and because the model size could have affected the reservoir performance, as shown by the sensitivity analysis. This research indicated that permeability heterogeneity and the injection procedure could significantly affect the reservoir performance, and should be given consideration in the performance assessment.