Application of Proxy Model and Reservoir Simulation for CO2 Injection Optimization in Multi-Layer Depleted Reservoirs: A Case Study of CCS Project in Gulf of Thailand

Abstract

PTTEP has committed to achieve net zero greenhouse gas emission in 2050 and the CCS project in the gulf of Thailand is the first CCS planned by PTTEP to achieve this goal. To evaluate the feasibility of the project, the reservoir simulation study was extensively performed from the geological, petrophysical, seismic inversion and reservoir engineering input. One of the keys in CCS project is the CO2 injection strategy and CO2 storage volume. Therefore, the objective of the study is to determine the optimized injection strategy and storage volume for CO2. Normally, the CO2 injector locations were targeted at the high permeability-thickness (KH) location based on reservoir simulation model. However, the optimization of injection with many layers in depleted reservoirs is cumbersome and required up to millions of simulation runs which is not possible to be performed. Therefore, this study proposes the new workflow to optimize the injection strategy by integrating reservoir simulation and proxy model or surrogate model simplified from actual simulation model results as pre-screening. First, we identify well location candidates based on KH map, pressure and shale thickness. Then, we construct the creaming curves scenario from reservoir simulation model and input them to train proxy model. With this workflow, instead of evaluating millions of injection scenarios by reservoir simulation, consisting of which well location and which reservoirs to be injected as separate target or commingle, we instead optimized the time required to obtain the injection strategy firstly from proxy model as pre-screening. Then, we run these possible optimized scenarios from proxy model to be run with reservoir simulation model. Finally, we obtain the most optimized injection strategy from simulation model including selected injector locations, number of injectors and sand targets for each injector. This can be obtained by evaluating the cumulative CO2 injection volume per number of injectors. To conclude, the optimized scenario of CO2 injection strategy and volume were determined and ready to be coupled with the integrated model including geomechanics, geochemistry and thermal simulation. With the proposed workflow, the study is the first CCS project in Thailand that adopts the injection strategy optimization based on the integration of dynamic reservoir simulation and proxy model. The proposed workflow can be used as the practical and robust injection optimization workflow used in any future CCS project in depleted reservoirs with multi-layers.