Impacts of probabilistic geological realizations in a geothermal reservoir using numerical and statistical investigations

Achieving a (?) reliable geological model is the foremost step in all underground resource assessments. However, regarding the sparsity of data and lack of knowledge, a spectrum of solutions makes more sense compared to a single deterministic model. It this study, a probabilistic geological modeler (Gempy) is used to understand the effect of existing uncertainty in the data representing subsurface layers and faults. A synthetic single fault model in which both the layers and fault are perturbed is designed. Random numbers are used for perturbation to prevent from any bias. In the first round of uncertainty analysis, thickness of reservoir layer in the footwall and location of the faults are perturbed. In the next round, dip and direction of fault are considered to be uncertain. In each of two rounds, 20 geological realization are resulted to act as a framework for later numerical simulations. After perturbing different elements of the synthetic geological setting and generating mesh (using GMSH) for each scenario (40 ones), TIGER code is exerted to simulate the tracer flow path. All the three packages are open source and availability of Gempy and GMSH in Python ecosystem facilitates the transfer from structural models to a high quality mesh. A doublet system (one injection and one production well) penetrating a geothermal reservoir is simulated in this study. In the base model, only the production well is passing through the fault but adding uncertainty to location of the fault resulted in having realization in which both wells penetrate the fault. Through simulating the tracer path for all geological realizations, sensitivity of results to the location of the fault is clearly observed. Statistical analyses revealed and numerically quantified the effect of structural uncertainty on the flow properties of a doublet system in a geothermal reservoir.