A Unified Modelling Approach using AI based Probabilistic Method to Address Structural Uncertainty in Carbonate Reservoirs

Résumé fourni par la source

Structural uncertainty is the degree of variability or error in the interpretation of subsurface geology, especially the geometry and position of the reservoir. It affects the estimation of hydrocarbon volume and the design of the optimal field development plan. In complex and heterogeneous carbonate reservoirs, structural uncertainty can be significant due to factors such as velocity variations, seismic resolution, multiple scenarios, and human bias. Therefore, it is crucial to have a robust and efficient workflow to quantify and reduce structural uncertainty using all available data sources. One of the challenges in conventional structural uncertainty workflows is the reliance on deterministic methods that often produce a single or few scenarios based on subjective assumptions and manual adjustments. These methods are time-consuming, inconsistent, and difficult to validate. Moreover, they do not account for the uncertainty in the seismic data quality and processing, which can propagate and affect the reservoir model. A more rigorous and objective approach is needed to capture the full range of possible structural outcomes and their probabilities. This study proposes an integrated stochastic structural workflow that leverages Artificial Intelligence (AI) techniques to generate multiple realizations of the reservoir structure and quantify their uncertainty. The workflow consists of four main steps: (1) generating an initial structural model based on seismic and well data, (2) applying an AI-based probabilistic method to estimate the combined uncertainty (horizon interpretation & velocity) and its impact on the depth conversion, (3) performing a stochastic simulation to sample the combined uncertainty distribution and generate multiple depth-converted structural models, and (4) evaluating the structural uncertainty ranges and their implications for the reservoir characterization, volumes {Gross Rock Volume (GRV) & Stock Tank Oil Initially In Place (STOIIP)}, and field development plan. The workflow is applied to a case study of heterogeneous carbonate reservoirs in the Middle East, and the results are compared with the conventional deterministic methods for capturing depth uncertainty.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.